tag:blogger.com,1999:blog-80435804556762013642024-03-28T21:55:07.601-04:00NeoskepticsA blog devoted to the discussion of public policy based on data that's available to the general public.<br>
<i>"Reading maketh a full man, conference a ready man, writing an exact man" ... Sir Francis Bacon (1561-1626)</i>Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.comBlogger168125tag:blogger.com,1999:blog-8043580455676201364.post-69420118603831647302024-03-24T20:15:00.024-04:002024-03-27T21:35:38.611-04:00Nvidia's new chips ... Musk's Grok-1 ... Microsoft hires DeepMind founder .. GPT-5 ... Gemini on iPhones ... TL;DR + podcast 24Mar24<div style="text-align: right;"><span style="font-family: verdana; font-size: x-small;"><i>Last update: Sunday 3/24/24 </i></span></div><div style="text-align: left;"><span style="font-family: verdana;"><div class="separator" style="clear: both; text-align: center;"><div style="text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><span style="font-size: medium;"><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhjif6k8A_fWg29o4injUY3l9y4EQjKT7cpbytReHool_Ja6x8kXbK-8r4Fbkq-zGg0RTU8u_pt6I58K0OcMzR0tP3JIjzcCiDvD4nfDufMLcsJWKkKLBTmdXez2s_T89gefgYVRIxooslbenVdnrlt1PpFsbzKspDchQgQGnzY4nyyehj_fWY1ZxaVLrY/s750/NvidiaNewChip.jpg" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="563" data-original-width="750" height="240" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhjif6k8A_fWg29o4injUY3l9y4EQjKT7cpbytReHool_Ja6x8kXbK-8r4Fbkq-zGg0RTU8u_pt6I58K0OcMzR0tP3JIjzcCiDvD4nfDufMLcsJWKkKLBTmdXez2s_T89gefgYVRIxooslbenVdnrlt1PpFsbzKspDchQgQGnzY4nyyehj_fWY1ZxaVLrY/s320/NvidiaNewChip.jpg" vspace="10" width="320" /></a></div>Welcome to our 24Mar</span></span></span></span><span style="font-size: medium;"><span><span><span><span style="background-color: white; color: #1a222a;">24</span><span style="background-color: white; color: #1a222a;"> </span><span style="background-color: white; color: #1a222a;"><span style="color: #1a222a;">podcast + </span><span style="color: #1a222a;">TL;DR summary of the past</span> week's top 5 AI stories on our "</span><b><i><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html"><span style="color: #cc0000;">Useful AI News</span></a></i></b><span style="background-color: white; color: #1a222a;">" page </span></span><b style="background-color: white; color: #1a222a;">...</b><span style="background-color: white; color: #1a222a;"> </span></span></span><span style="background-color: white; color: #1a222a;">(1) More powerful Nvidia chips, (</span><span><span style="background-color: white; color: #1a222a;">2)<span style="color: black;"> Musk releases his chatbot's code</span>, 3) <span style="color: black;">Microsoft hires DeepMind co-founder</span>, (4) OpenAI's</span></span><span style="background-color: white; color: #1a222a;"> </span><span style="background-color: white; color: #1a222a;">GPT-5 "soon", and (5) Apple might use Google's Gemini for iPhones</span></span></div><div style="text-align: left;"><br /></div><div style="text-align: left;"><div class="separator" style="clear: both;"><div style="text-align: right;"><div style="color: #1a222a;"><span style="font-size: medium;"><span>Click link for podcast ... 9 minutes ... opens in new tab</span> </span></div><div><span style="font-size: medium;"><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; text-align: left; text-decoration-line: none;"><span style="color: #1a222a;">➡</span></a><span style="color: #1a222a;"> </span><i><span><b><a href="https://drive.google.com/file/d/1Qk_ktOUx0C7At71613gC0DVHWbdoc5nC/view?usp=sharing" target="_blank"><span style="color: #cc0000;">Nvidia + Musk + Microsoft + GPT-5 + Apple/Google</span></a><span style="color: #1a222a;"> </span></b></span></i></span></div></div></div><div class="separator" style="clear: both;"><div style="text-align: right;"><span style="font-size: medium;"><span style="color: black;"><span style="color: #1a222a;">Click the "start" button the when the podcast page is loaded<br />... </span></span><span style="color: #1a222a;"> If audio fails to start, or gets stuck, try reloading that page</span><br /><br /></span></div></div></div><div style="background-color: white; text-align: right;"><div class="separator" style="clear: both; font-family: Verdana, Geneva, sans-serif; text-align: left;"><span style="color: #1a222a;"><span style="font-size: medium;">TL;DR link<b> </b></span></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-family: verdana; font-size: large; text-decoration-line: none;"><span style="color: #1a222a;">➡</span></a><b><span style="font-size: medium;"><span style="color: #1a222a;"> </span><i><a href="https://neoskeptics.blogspot.com/2024/03/nvidias-new-chips-musks-grok-1.html#more"><span style="color: #cc0000;">HERE</span></a></i></span></b></div></div></div></span></div><span><span style="font-family: verdana;"><a name='more'></a></span></span><div style="text-align: left;"><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;">A. </b><b style="background-color: white; font-family: verdana; font-size: large;">TL;DR ... </b><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;">Top 5 stories in past week ...</b></div><div style="text-align: left;"><b><span style="font-family: verdana; font-size: medium;"><br /></span></b></div><div style="text-align: left;"><span style="font-family: verdana;"><div style="font-family: Times;"><span style="font-family: verdana;"><b>1) </b></span><span style="background-color: white; color: #1a222a; font-family: verdana;"><b>More powerful Nvidia chips</b></span></div><div><div>This week Nvidia announced the production of "Blackwell", its newest graphics processing unit (GPU). Blackwell chips are 30 times as powerful as Nvidia's H100 chips and use far less energy. </div><div><br /></div><div>In 2023, Big Tech's insatiable demand for Nvidia's H100 GPUs to support their generative AI services propelled Nvidia's market capitalization high enough to make it the third most valuable company in the world, behind Microsoft and Apple. Nvidia anticipates even greater demand for Blackwell. </div><div><br /></div><div>Nevertheless, we should remember two important exceptions: Google has been designing its own GPUs for a few years now, and Microsoft recently signed a contract with Intel to manufacture GPUs that Microsoft will design. The importance of these exceptions derives from their current positions as the leading providers of generative AI services, their vast corporate wealth, and their determination to spend whatever is required to maintain their leadership positions.</div></div></span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b><br /></b></span></div><div style="text-align: left;"><span style="font-family: verdana;"><b>2</b>) </span><span style="background-color: white; font-family: verdana;"><b>Musk releases code for Grok-1</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">On March 17, 2024, the xAI corporation posted a <b><i><a href="https://x.ai/blog/grok-os"><span style="color: #cc0000;">notice on its blog</span></a></i></b> that it was releasing the weights and architecture of its 314 billion parameter language model, the model underlying its chatbot Grok. In other words, Grok-1 was now an open source model.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">This move was consistent with <b><i><a href="https://neoskeptics.blogspot.com/2024/03/musk-sues-altmanopenai-tldr-and-podcast.html#more"><span style="color: #cc0000;">Musk's ongoing lawsuit</span></a></i></b> against Sam Altman and OpenAI because of OpenAi's failure to create open source models. Musk claims that this failure violated the commitment they made to him made when Musk invested the $44 million that launched OpenAI back in 2015.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">xAI's announcement did not provide the results of benchmark tests that could be used to compare its performance with the performance of other models. Both GPT-4 (Microsoft and OpenAI) and Gemini (Google) have over 1.5 trillion parameters, making each of them five times as large as Grok-1's 314 billion parameters.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">We will have to wait until someone else publishes benchmark tests for this new open source entry. If other developers find that it has sufficient power and other "interesting" features, it might attract widespread interest within the open source community, interest that would enable it to become a platform that others will work to improve.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><b><span style="font-family: verdana;">3) </span></b><span style="background-color: white; font-family: verdana;"><b>Microsoft hires DeepMind co-founder</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">Microsoft hired Mustafa Suleyman as CEO of its new Consumer AI division that will manage its Copilot, search, and browser applications. Suleyman is</span><span style="font-family: verdana;"> a co-founder of DeepMind and the current CEO of InflectionAI, a start-up whose biggest initial funding came from Microsoft. </span><span style="font-family: verdana;">Karén Simonyan, Inflection’s co-founder, will also join Microsoft as chief scientist for the new consumer AI group, together with other members of the Inflection team.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">It is highly unusual for a sponsor of a start-up to cannibalize the start-up's founding team in this </span></div><div style="text-align: left;"><span style="font-family: verdana;">manner. The well-connected publication</span><span style="background-color: white; font-family: verdana;"> </span><span style="background-color: white; font-family: verdana;"><a href="https://www.theinformation.com/articles/microsoft-backed-inflection-arranges-unusual-payout-for-startups-investors?rc=v6kkoz&shared=5173e98ebfdc2ce4" style="font-style: italic; font-weight: bold;"><span style="color: #cc0000;">The Information</span></a></span><span style="background-color: white; font-family: verdana;"> notes that:</span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="background-color: white; font-family: verdana;">"Inflection disclosed a licensing deal with Microsoft to make its models available for sale on Microsoft’s Azure, and said other cloud platforms would follow. It has also told investors they will fully recoup their investment, and more, as a result of the licensing agreement, according to a person briefed on the arrangement." <br /></span></li></ul><div><span style="background-color: white; font-family: verdana;">The Information conjectures that Microsoft is making this substantial payment to Inflection AI in order to (1) compensate Inflection AI for losses that might result from Microsoft's hiring so many of its key staff, and (2) to ward off review of the hiring by government antitrust regulators had the hiring resulted in losses for Inflection AI.</span></div></div><div style="text-align: center;"><span style="font-family: verdana;"><b style="background-color: white; font-size: large;"><i><span style="color: #cc0000;"><br /></span></i></b></span></div><div style="text-align: left;"><b><span style="font-family: verdana;">4) </span><span style="font-family: verdana;"><span style="background-color: white; color: #1a222a;">OpenAI's </span></span><span style="background-color: white; color: #1a222a; font-family: verdana;">mid-year release of </span><span style="background-color: white; color: #1a222a; font-family: verdana;">GPT-5</span></b></div><div style="text-align: left;"><span style="font-family: verdana;"><span>OpenAI will release GPT-5 </span>"in the coming months" or "soon" or "mid-year" or "likely during summer" according to <b><i><a href="https://www.blogger.com/blog/post/edit/8043580455676201364/6942011860383164730#"><span style="color: #cc0000;">Business Insider</span></a> </i></b>in an article cited by other tech publications. In other words, this story is a well founded rumor ... that has considerable credibility because </span><span style="font-family: verdana;">GPT-4's</span><span style="font-family: verdana;"> many shortcomings are well documented and have caused major enterprises to limit their purchases of the pricy services provided by GPT-4. </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">A section of last week's TL;DR ("<b><i><a href="https://neoskeptics.blogspot.com/2024/03/eu-ai-regs-google-blocks-gemini.html#GenAIPrifits"><span style="color: #cc0000;">Does generative AI really boost profits?</span></a></i></b>"<b style="color: #1a222a;">) </b><span style="color: #1a222a;">noted that the biggest providers of generative AI services are t</span></span><span style="font-family: verdana;"><span style="color: #1a222a;">empering expectations with their salespeople, saying the hype about the technology has gotten ahead of what it can actually do for customers at a reasonable price. </span></span></div><div style="text-align: left;"><span style="font-family: verdana;"><span style="color: #1a222a;"><br /></span></span></div><div style="text-align: left;"><span style="font-family: verdana;"><span style="color: #1a222a;">Indeed, if we are at the initial peak of a </span><b><i><a href="https://www.gartner.com/en/research/methodologies/gartner-hype-cycle"><span style="color: #cc0000;">Gartner hype cycle</span></a></i></b><span style="color: #1a222a;">, then a large scale collapse of demand for GenAI is inevitable ... unless ... unless demand can be stimulated by the release of a "new and improved GPT" ... Enter GPT-5, stage right ... applause signs flashing ... <b><i>:-)</i></b></span></span></div><div style="text-align: center;"><span style="font-family: verdana;"><b style="background-color: white; font-size: large;"><i><span style="color: #cc0000;"><br /></span></i></b></span></div><div style="text-align: left;"><b><span style="font-family: verdana;">5) </span><span style="background-color: white; color: #1a222a; font-family: verdana;">Apple might use Google's Gemini for iPhones</span></b></div><div style="text-align: left;"><span style="font-family: verdana;">According to <b><i><a href="https://www.bloomberg.com/news/articles/2024-03-18/apple-in-talks-to-license-google-gemini-for-iphone-ios-18-generative-ai-tools"><span style="color: #cc0000;">Bloomberg</span></a></i></b>:</span></div><div style="text-align: left;"><blockquote><span style="font-family: verdana;">"Apple Inc. is in talks to build Google’s Gemini artificial intelligence engine into the iPhone, according to people familiar with the situation, setting the stage for a blockbuster agreement that would shake up the AI industry.</span> </blockquote><blockquote><span style="font-family: verdana;">The two companies are in active negotiations to let Apple license Gemini, Google’s set of generative AI models, to power some new features coming to the iPhone software this year, said the people, who asked not to be identified because the deliberations are private.</span></blockquote><blockquote><span style="font-family: verdana;">Apple also recently held discussions with OpenAI and has considered using its model, according to the people."</span></blockquote></div><div style="text-align: left;"><span style="font-family: verdana;">Why would Apple prefer Google over OpenAI? Apple's commitment to privacy is more consistent with a genAI model that's small enough to run on an iPhone, rather a cloud-based model. Google has Gemini Nano, whereas OpenAI only has large cloud models, like GPT-4 and GPT-5 (forthcoming). Furthermore, Apple and Google already have a flourishing partnership via Google's search engine ... although federal regulators are challenging that partnership. </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><span style="font-family: verdana;">Taking a step back, we have to ask why Apple needs an AI partner? Why can't Apple go it alone? Perhaps (1) because Apple is coming very late to the genAI table and (2) because Apple may be feeling financial pressure from government regulators.</span></span></div><div style="text-align: left;"><span style="font-family: verdana;"><span style="font-family: verdana;"><br /></span></span></div><div style="text-align: left;"><span style="font-family: verdana;"><span style="font-family: verdana;">In a recent article </span></span><span style="font-family: verdana;">the </span><b style="font-family: verdana;"><i><a href="https://www.nytimes.com/2024/03/21/technology/apple-regulators-world.html?smid=nytcore-android-share&ugrp=c&pvid=6b9d2d93-2326-40a1-84b3-553920507eb4"><span style="color: #cc0000;">NY Times</span></a></i></b><span style="font-family: verdana;"> discussed the U.S. Department of Justice's new antitrust suit against Apple. The Times then moved on to provide a comprehensive overview of actions taken by government agencies all over the world that have contributed to Apple's slippage to second place behind Microsoft as the most valuable company in the world. </span><span style="font-family: verdana;"> As consequence, government pressure may also be moving Apple to jumpstart its late entry into generative AI via a partnership with Google</span><span style="font-family: verdana;">.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><div style="background-color: white;"><div><div id="Overview"><span style="font-size: medium;"><span style="color: black;"><span style="font-family: verdana;"><b>B. </b></span></span><span><span style="font-family: verdana;"><b>Top 5 stories in past week on "<i><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html"><span style="color: #cc0000;">Useful AI News</span></a></i>"</b></span></span></span></div></div><div><ol><li><b style="font-family: verdana; font-size: small; text-align: center;"><a href="#OpEds"><span style="color: #cc0000;">Misc</span></a></b><br style="font-family: verdana;" /><span style="font-family: verdana; font-size: medium;">"Nvidia's next-gen AI chips are way more powerful and use a lot less energy", Alex Perry, <b><i><a href="https://t.co/eZjNLpiMU5"><span style="color: #cc0000;">Mashable</span></a></i></b>, 3/19/24 <b><span style="color: #cc0000;">***</span></b> <br />-- This story also covered by <b><i><a href="https://t.co/GjNLWi1cH1"><span style="color: #cc0000;">Bloomberg</span></a></i></b>, <b><i><a href="https://www.cnbc.com/2024/03/19/nvidia-stock-dips-after-company-unveils-latest-ai-chips.html"><span style="color: #cc0000;">CNBC</span></a></i></b>, <b><i><a href="https://t.co/7CZyinR6qR"><span style="color: #cc0000;">VentureBeat</span></a></i></b>, <b><i><a href="https://www.fastcompany.com/91033514/nvidia-most-innovative-companies-2024"><span style="color: #cc0000;">FastCompany</span></a></i></b>, <b><i><a href="https://t.co/TrqwZUCDiH"><span style="color: #cc0000;">Reuters</span></a></i></b>,</span><br /><br /></li><li><span style="font-family: verdana;"><a href="#OtherLLMs" style="font-size: small; font-weight: 700; text-align: center;"><span style="color: #cc0000;">Other Models</span></a><br /><span style="font-size: medium;">"In Latest A.I. War Escalation, Elon Musk Releases Chatbot Code", Kate Conger and Cade Metz, <b><i><a href="https://www.nytimes.com/2024/03/17/technology/chatbot-xai-code-musk.html?smid=nytcore-android-share"><span style="color: #cc0000;">NY Times</span></a></i></b>, 3/17/24 <b><span style="color: #cc0000;">***</span></b> <br />-- This story also covered by <b><i><a href="https://techcrunch.com/2024/03/17/xai-open-sources-base-model-of-grok-but-without-any-training-code/"><span style="color: #cc0000;">TechCrunch</span></a></i></b>, <b><i><a href="https://www.theverge.com/2024/3/17/24097810/xai-open-source-grok-musk-generative-ai-llm"><span style="color: #cc0000;">The Verge</span></a></i></b>, <b><i><a href="https://www.wired.com/story/elon-musk-no-choice-open-chatbot-grok/"><span style="color: #cc0000;">Wired</span></a></i></b>, <b><i><a href="https://venturebeat.com/ai/musks-grok-ai-goes-open-source/"><span style="color: #cc0000;">VentureBeat</span></a></i></b>, <b><i><a href="https://mashable.com/article/grok-open-source-elon-musk-ai-openai"><span style="color: #cc0000;">Mashable</span></a></i></b>, ... and </span><i><span><span style="font-size: medium;"><b><a href="https://x.ai/blog/grok-os"><span style="color: #cc0000;">xAI</span></a><span style="color: #cc0000;"> </span></b><br /><br /></span></span></i></span></li><li style="font-size: medium;"><span style="font-size: medium;"><span style="font-family: verdana;"><a href="#Microsoft" style="font-size: small; text-align: center;"><b><span style="color: #cc0000;"><span>M</span>icrosoft</span></b></a><br />"Microsoft Hires DeepMind Co-Founder Suleyman to Run Consumer AI", Dina Bass, <b><i><a href="https://t.co/Yyn8lQQTSe"><span style="color: #cc0000;">Bloomberg</span></a></i></b>, 3/19/24 <b><span style="color: #cc0000;">***</span></b> <br />-- This story also covered by </span><span style="font-family: verdana;"><a href="https://t.co/9BXu7HPK7A" style="font-style: italic; font-weight: bold;"><span style="color: #cc0000;">The Verge</span></a><b style="font-style: italic;">, </b><a href="https://t.co/JuzP65trTU" style="font-style: italic; font-weight: bold;"><span style="color: #cc0000;">TechCrunch</span></a><b style="font-style: italic;">, </b><b><i><a href="https://www.nytimes.com/2024/03/19/technology/mustafa-suleyman-google-gemini.html?smid=nytcore-android-share"><span style="color: #cc0000;">NY Times</span></a>, </i></b><span><span><a href="https://www.forbes.com/sites/alexkonrad/2024/03/19/inflection-abandons-chatgpt-challenger-ceo-suleyman-joins-microsoft/?utm_source=newsletter&utm_medium=email&utm_campaign=dailydozen&cdlcid=623cf7856e1a1d1211df6bb6&section=tech&sh=3af46d2047fb" style="font-style: italic; font-weight: bold;"><span style="color: #cc0000;">Forbes</span></a><b> ... </b>and</span></span> <span><a href="https://www.theinformation.com/articles/microsoft-backed-inflection-arranges-unusual-payout-for-startups-investors?rc=v6kkoz&shared=5173e98ebfdc2ce4" style="font-style: italic; font-weight: bold;"><span style="color: #cc0000;">The Information</span></a></span> focuses on Microsoft's substantial payment to Inflection AI in order to (1) compensate Inflection AI for losses that might result from Microsoft's hiring some of its key staff, and (2) to ward off review of the hiring by government antitrust regulators if it resulted in losses for Inflection AI<br /><br /></span></span></li><li style="font-size: medium;"><span style="font-family: verdana;"><span style="font-family: verdana;"><span style="font-size: medium;"><b style="font-size: small; text-align: center;"><a href="#OpenAI"><span style="color: #cc0000;">OpenAI</span></a></b><br />"OpenAI is expected to release a 'materially better' GPT-5 for its chatbot mid-year, sources say", Kali Hays and Darius Rafieyan, <b><i><a href="https://www.businessinsider.com/openai-launch-better-gpt-5-chatbot-2024-3"><span style="color: #cc0000;">Business Insider</span></a></i></b>, 3/19/24 <b><span style="color: #cc0000;">***</span></b> <br />-- This story also covered by <b><i><a href="https://mashable.com/article/openai-gpt-5-release-this-summer?utm_campaign=mash-com-tw-main-link&utm_contnet=tech&utm_medium=twitter&utm_source=social"><span style="color: #cc0000;">Mashable</span></a></i></b>, <b><i><a href="https://www.zdnet.com/article/materially-better-gpt-5-could-come-to-chatgpt-as-early-as-this-summer/"><span style="color: #cc0000;">ZDNet</span></a></i></b>,<br /><br /></span></span></span></li><li style="font-size: medium;"><span style="font-family: verdana; font-size: medium;"><a href="#Google" style="font-size: small; font-weight: 700; text-align: center;"><span style="color: #cc0000;">Google</span></a><br />"Apple Is in Talks to Let Google Gemini Power iPhone AI Features", Mark Gurman. <b><i><a href="https://www.bloomberg.com/news/articles/2024-03-18/apple-in-talks-to-license-google-gemini-for-iphone-ios-18-generative-ai-tools"><span style="color: #cc0000;">Bloomberg</span></a></i></b>, 3/19/24 <b><span style="color: #cc0000;">***</span></b> <br />-- This story also covered by <b><i><a href="https://www.theinformation.com/articles/openai-and-apples-rude-wake-up-call?rc=v6kkoz&shared=68563311d2fc7fdb"><span style="color: #cc0000;">The Information</span></a></i></b>, <b><i><a href="https://www.reuters.com/technology/apple-talks-let-googles-gemini-power-iphone-ai-features-bloomberg-news-says-2024-03-18/"><span style="color: #cc0000;">Reuters</span></a></i></b></span></li></ol><div><span style="font-family: verdana;"><span style="font-size: medium;"><b><br /><span style="color: #1a222a;">C. Dozen Basic AI FAQs </span></b></span></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="background-color: transparent; color: #336699; font-family: verdana; font-size: large; text-decoration-line: none;"><span style="color: #1a222a;">➡</span></a><span style="font-size: large;"><span style="color: #1a222a;"> </span></span><b style="font-family: verdana; font-size: large;"><i><a href="https://neoskeptics.blogspot.com/2023/12/basic-ai-faqs.html"><span style="color: #cc0000;">HERE</span></a></i></b></div><div style="color: #1a222a; font-size: medium;"><span style="font-family: verdana;"><span style="font-size: medium;">This page contains links to responses by Google's Bard chatbot running Gemini Pro to 12 questions that should be asked more frequently, but aren't. As consequence, too many readily understood AI terms have become meaningless buzzwords in the media.</span></span></div></div></div></div>
Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-15717445154750975092024-03-18T20:43:00.020-04:002024-03-27T21:35:50.668-04:00EU AI regs ... Google blocks Gemini election queries ... more amazing SORA videos ... and is GenAI the new crypto ...TL;DR and podcast 18Mar24<div style="text-align: right;"><span style="font-family: verdana; font-size: x-small;"><i>Last update: Monday 3/18/24 </i></span></div><div style="text-align: left;"><span style="font-family: verdana;"><div class="separator" style="clear: both; text-align: center;"><div style="text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><span style="font-size: medium;"><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiA_ftfldG5cr_4dMulFmCmNjJVERwrhtwbRe5iOtlG1nvZn1EpKcOBzijPEFXb1ZENMqjzNaNfe9fbQkNQnpMmg_FPYZNVWqw1Hn9ThGYjh7nfZ2WOTG14DI1Zdd_DjG5jAHxsP4jVEnKT_gnG1XU-_kmP_LySljFrVfK1MYbMOS5OlLPuCEgo3yHMly8/s1800/GenAIShimmeringCrystalBall%20.jpg" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="1012" data-original-width="1800" height="180" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiA_ftfldG5cr_4dMulFmCmNjJVERwrhtwbRe5iOtlG1nvZn1EpKcOBzijPEFXb1ZENMqjzNaNfe9fbQkNQnpMmg_FPYZNVWqw1Hn9ThGYjh7nfZ2WOTG14DI1Zdd_DjG5jAHxsP4jVEnKT_gnG1XU-_kmP_LySljFrVfK1MYbMOS5OlLPuCEgo3yHMly8/s320/GenAIShimmeringCrystalBall%20.jpg" vspace="10" width="320" /></a></div>Welcome to our 18MAR</span></span></span></span><span style="font-size: medium;"><span><span style="background-color: white; color: #1a222a;">24</span><span style="background-color: white; color: #1a222a;"> </span><span style="background-color: white; color: #1a222a;"><span style="color: #1a222a;">podcast + </span><span style="color: #1a222a;">TL;DR summary of the past</span> week's top AI stories on our "</span><b style="color: #1a222a;"><i><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html">Useful AI News</a></i></b><span style="background-color: white; color: #1a222a;">" page </span></span><b style="background-color: white; color: #1a222a;">➡</b><span style="background-color: white; color: #1a222a;"> 1) <span style="font-size: medium;">EU approves AII regulations</span> (2) <span style="font-size: medium;">Google blocks Gemini responses to election queries everywhere </span>(3) <span style="font-size: medium;">OpenAI freaks out Wall Street Journal with Sora's videos </span>and (4) <span style="font-size: medium;">Little evidence so far that generative AI really boosts profits</span></span></span></div><div style="text-align: left;"><div class="separator" style="clear: both; font-size: 13.86px;"><div style="text-align: right;"><div style="color: #1a222a;"><span style="font-size: medium;">Click link for podcast (opens in new tab)</span> </div><div><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-size: large; text-align: left; text-decoration-line: none;"><span style="color: #1a222a; font-size: medium;">➡</span></a><span style="color: #1a222a;"> </span><i><span style="font-size: medium;"><b><a href="https://drive.google.com/file/d/1QG5hatFFbnmx6zw7SVTrxSndvjgN1gDQ/view?usp=sharing" target="_blank"><span style="color: #cc0000;">EU + Google + OpenAI + GenAI</span></a><span style="color: #1a222a;"> </span></b></span></i></div></div></div><div class="separator" style="clear: both;"><div style="text-align: right;"><span style="color: black; font-size: 13.86px;"><span style="color: #1a222a;">Click the "start" button the when the podcast page is loaded<br />... </span></span><span style="color: #1a222a; font-size: 13.86px;"> If audio fails to start, or gets stuck, try reloading that page</span><br /></div></div></div><div style="background-color: white; text-align: right;"><div class="separator" style="clear: both; text-align: left;"><div style="color: #1a222a; font-family: Verdana, Geneva, sans-serif;"><span style="font-family: verdana;"><div><br /></div></span></div></div><div class="separator" style="clear: both; color: #1a222a; font-family: Verdana, Geneva, sans-serif; text-align: left;"><span style="font-size: medium;">TL;DR link<b> </b></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-family: verdana; font-size: large; text-decoration-line: none;"><span style="color: #1a222a;">➡</span></a><b style="color: black;"><span style="font-size: medium;"><span style="color: #1a222a;"> </span><i><a href="https://neoskeptics.blogspot.com/2024/03/eu-ai-regs-google-blocks-gemini.html#more"><span style="color: #cc0000;">HERE</span></a></i></span></b></div></div></div></span></div><span><span style="font-family: verdana;"><a name='more'></a></span></span><div style="text-align: left;"><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;">A. </b><b style="background-color: white; font-family: verdana; font-size: large;">TL;DR ... </b><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;">Top 4 stories in past week ...</b></div><div style="text-align: left;"><b><span style="font-family: verdana; font-size: medium;"><br /></span></b></div><div style="text-align: left;"><b><span style="font-family: verdana;">1) </span></b><span style="background-color: white; color: #1a222a; font-family: verdana;"><b>EU approves AI regulations</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">As per the headline in Bloomberg's article: "World’s Most Extensive AI Rules Approved in EU Despite Criticism". More specifically, the EU parliament approved these new regulations; but the regs don't go into effect until all of the European Union's member states give their approval ... which is usually a formality, so their approval is expected soon. </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><b><i><a href="https://t.co/BLbzGUHTqI"><span style="color: #cc0000;">Bloomberg's article</span></a></i></b> asserts that:</span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">"The new law is intended to address worries about bias, privacy and other risks from the rapidly evolving technology. The legislation would ban the use of AI for detecting emotions in workplaces and schools, as well as limit how it can be used in high-stakes situations like sorting job applications. It would also place the first restrictions on generative AI tools, which captured the world’s attention last year with the popularity of ChatGPT ...<br /><br /></span></li><li><span style="font-family: verdana;">... "As talks reached the final stretch last year, the French and German governments pushed back against some of the strictest ideas for regulating generative AI, arguing that the rules will hurt European startups like France’s Mistral AI and Germany’s Aleph Alpha GmbH. Civil society groups like Corporate Europe Observatory (CEO) raised concerns about the influence that Big Tech and European companies had in shaping the final text."</span></li></ul></div><div style="text-align: left;"><span style="font-family: verdana;">So the resulting set of regulations is a compromise.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><b><span style="font-family: verdana;">2) </span></b><span style="background-color: white; color: #1a222a; font-family: verdana;"><b>Google blocks Gemini responses to election queries everywhere</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">Google announced its decision to prevent Gemini, its genAI chatbot, from responding to queries about elections in eveery country in which elections are being held this year. All election queries will be forwarded to its search engine which will merely return its usual lists of links to websites. Google's decision is understandable given (1) widespread fear that elections might be contaminated by AI-based misinformation, and (2) Google's recent embarrassment caused its "woke" chatbot returning racially diverse images when prompted for images of the Founding Fathers, Nazi soldiers, etc.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Nevertheless, Google's efforts may produce another backlash due to bias built into its search results. In other words, Google search has no "understanding" of a query; so it merely returns links to popular websites whose content is related to key words in the query. For example, when the editor of this blog submitted the politically controversial query "Is Biden too old to be president?" on Monday 3/18/24 at 4 pm, the top link returned by Google search was the most popular site related to "Biden" "too old" "president". <br /></span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">"Overwhelming majority of Americans think Biden is too old for another term", Meredith Deliso, <b><i><a href="https://abcnews.go.com/Politics/poll-americans-on-biden-age/story?id=107126589"><span style="color: #cc0000;">ABC</span></a></i></b>, 3/11/24</span></li></ul></div><div style="text-align: left;"><span style="font-family: verdana;">Who could doubt that these springtime polls are the voice of the people? They express the preferences of an overwhelming majority of the people, preferences that the majority will continue to express in their votes comes this November. Trump himself could not have said it better. Indeed, that's what Trump has been saying for four years. We can't let them steal it again, right? Hmmmmmmm</span></div><div style="text-align: left;"><span style="font-family: verdana;"><b><i><br /></i></b></span></div><div style="text-align: left;"><span style="font-family: verdana;"><b>3) </b></span><span style="background-color: white; color: #1a222a; font-family: verdana;"><b>OpenAI freaks out Wall Street Journal with Sora's videos</b><span style="font-size: large;"> </span></span></div><div style="text-align: left;"><span style="font-family: verdana;">The Wall Street Journal provided a few very simple prompts that Sora converted in another set of dazzling videos. If you haven't seen them yet, give yourself a treat by clicking on the links to the videos in the <b><i><a href="https://www.wsj.com/tech/personal-tech/openai-cto-sora-generative-video-interview-b66320bb?st=24jczt8gws17cc5&reflink=desktopwebshare_permalink"><span style="color: #cc0000;">Journal's article</span></a></i></b> ... Then remember DALL-E's images ... also dazzling ... but never quite what was requested if the requests had the most minimal levels of specificity. </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Readers of this blog may recall its editor's requests to DALL-E via ChatGPT<a href="https://neoskeptics.blogspot.com/2024/02/googles-gemma-intel-chips-designed-by.html#woke"><span style="color: #cc0000;"> <b><i>reported in a TL;DR two weeks ago</i></b></span></a><span style="color: #cc0000;"><b><i> </i></b></span>for a picture of five young white men in mid-twenties seated around a table. All should be seated and all should be wearing sandals. Nevertheless, DALL-E returned a picture of five seated white men, two of whom were clearly in their 40s, two were wearing sneakers, and three more men were standing behind the five seated at the table. </span></div><div style="text-align: center;"><span style="font-family: verdana;"><b style="background-color: white; font-size: large;"><i><span style="color: #cc0000;"><br /></span></i></b></span></div><div style="text-align: left;"><b><span style="font-family: verdana;">4) </span></b><span id="GenAIPrifits" style="background-color: white; color: #1a222a; font-family: verdana;"><b>Does generative AI really boost profits?</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">The well-connected publication <b><i><a href="https://www.theinformation.com/articles/generative-ai-providers-quietly-tamp-down-expectations?rc=v6kkoz&shared=255399288ea1bf3b"><span style="color: #cc0000;">The Information</span></a></i></b> provided another exclusive report based on interviews with well-connected insiders at the major Big Tech providers of GenAI. </span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">"In the past year, major technology firms have championed generative artificial intelligence as the next big thing, boosting the stock market to new highs. But behind the scenes, representatives of major cloud providers and other firms that sell the technology are tempering expectations with their salespeople, saying the hype about the technology has gotten ahead of what it can actually do for customers at a reasonable price.<br /><br />Several executives, product managers and salespeople at the major cloud providers, such as Microsoft, Amazon Web Services and Google, also privately said most of their customers are being cautious or “deliberate” about increasing spending on new AI services, given the high price of running the software, its shortcomings in terms of accuracy and the difficulty of determining how much value they’ll get out of it. ...<br /></span><br /><span style="font-family: verdana;">... Customers of such software are “struggling with [questions of,] is it providing value? How do you know if it’s giving good answers? How do you evaluate how it’s doing?” he said. “Generative AI is in the first inning.”</span></li></ul><div><span style="font-family: verdana;">In other words, generative AI may be at the initial peak of a <b><i><a href="https://www.gartner.com/en/research/methodologies/gartner-hype-cycle"><span style="color: #cc0000;">Gartner hype cycle </span></a></i></b>wherein the promise of a new technology greatly exceeds its demonstrated valua. The peak is usually followed by a dramatic collapse of demand.</span></div></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">That's the bad news. Fortunately there is one undeniable exception to the "peak hype" assessment. The resounding success of GitHub Copilot reflects its enthusiastic endorsement by the software development community. GitHub Copilot really helps professional developers produce better code in substantially less time.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><br /></div><div style="text-align: left;"><div style="background-color: white;"><div style="color: black;"><div id="Overview" style="font-size: medium;"><span><span style="font-family: verdana;"><b style="font-size: large;">B. </b></span></span><span><span style="font-family: verdana;"><b style="font-size: large;">Top 4 stories in past week ...</b></span></span> </div><div id="Overview"><ol><li style="font-size: medium;"><span style="font-family: verdana; font-size: medium;"><a href="#Policy" style="font-size: small; font-weight: 700; text-align: center;"><span style="color: #cc0000;">Public Policy</span></a><br />"World’s Most Extensive AI Rules Approved in EU Despite Criticism", Jillian Deutsch, <b><i><a href="https://t.co/BLbzGUHTqI"><span style="color: #cc0000;">Bloomberg</span></a></i></b>, 3/13/24 <b><span style="color: #cc0000;">***</span></b> <br />-- This story also covered by <b><i><a href="https://t.co/lNWrkuSUTA"><span style="color: #cc0000;">TechCrunch</span></a></i></b>, <b><i><a href="https://t.co/lsBG3wv1gR"><span style="color: #cc0000;">BBC</span></a></i></b>, <b><i><a href="https://t.co/h1Vn2e4DpY"><span style="color: #cc0000;">Wall StreetJournal</span></a></i></b>, <b><i><a href="https://t.co/0PIofGDRnA"><span style="color: #cc0000;">VentureBeat</span></a></i></b>, <b><i><a href="https://t.co/50oZ8srYyh"><span style="color: #cc0000;">The Information</span></a></i></b>, <b><i><a href="https://mashable.com/article/eu-ai-law"><span style="color: #cc0000;">Mashable</span></a></i></b> ... and <b><i><a href="https://www.europarl.europa.eu/news/en/press-room/20240308IPR19015/artificial-intelligence-act-meps-adopt-landmark-law"><span style="color: #cc0000;">European Parliament</span></a><br /><br /></i></b></span></li><li style="font-size: medium;"><span style="font-family: verdana;"><span style="font-size: medium;"><a href="#Google" style="font-size: small; font-weight: 700; text-align: center;"><span style="color: #cc0000;">Google</span></a><br />"Google won’t let you use its Gemini AI to answer questions about an upcoming election in your country", Jagmeet Singh, <b><i><a href="https://t.co/gTitVlI2g0"><span style="color: #cc0000;">TechCrunch</span></a></i></b>, 3/13/24 <b><span style="color: #cc0000;">***</span></b> <br />-- This story also covered by <b><i><a href="https://arstechnica.com/gadgets/2024/03/no-politics-google-disables-election-related-questions-for-gemini/"><span style="color: #cc0000;">Ars Technica,</span></a></i></b> <b><i><a href="https://www.bbc.com/news/business-68551206"><span style="color: #cc0000;">BBC</span></a></i></b>, <b><i><a href="https://www.reuters.com/technology/google-restricts-ai-chatbot-gemini-answering-queries-global-elections-2024-03-12/"><span style="color: #cc0000;">Reuters</span></a></i></b>, <b><i><a href="https://www.engadget.com/googles-gemini-will-steer-clear-of-election-talk-205135492.html"><span style="color: #cc0000;">Engadget</span></a></i></b>, <br /></span><br /></span></li><li style="font-size: medium;"><span style="font-family: verdana;"><span style="font-size: medium;"><b style="font-size: small; text-align: center;"><a href="#OpenAI"><span style="color: #cc0000;">OpenAI</span></a></b><br />"OpenAI Made AI Videos for Us. These Clips Are Good Enough to Freak Us Out.", Joanna Stern, <b><i><a href="https://www.wsj.com/tech/personal-tech/openai-cto-sora-generative-video-interview-b66320bb?st=24jczt8gws17cc5&reflink=desktopwebshare_permalink"><span style="color: #cc0000;">Wall Street Journal</span></a></i></b>, 3/13/24 <b><span style="color: #cc0000;">***</span></b> <br />-- See also "The 12 OpenAI Sora TikToks That Broke Our Brains", Maxwell Zeff. <b><i><a href="https://t.co/IwnViLIOX3"><span style="color: #cc0000;">Gizmodo</span></a></i></b>, 3/14/2</span>4<br /><br /></span></li><li><span style="font-family: verdana; font-size: medium;"><b style="font-size: small; text-align: center;"><a href="#OpEds"><span style="color: #cc0000;">Misc</span></a></b><br /><span>"Amazon, Google Quietly Tamp Down Generative AI Expectations", Aaron Holmes and Anissa Gardizy, </span><b><i><a href="https://www.theinformation.com/articles/generative-ai-providers-quietly-tamp-down-expectations?rc=v6kkoz&shared=255399288ea1bf3b"><span style="color: #cc0000;">The Information</span></a></i></b><span>, 3/12/24</span></span><span style="font-size: medium;"><b style="text-align: center;"><b><i style="font-weight: 400;"> </i></b></b><b style="font-family: verdana;"><span style="color: #cc0000;">***</span></b><span style="font-family: verdana;"> </span></span></li></ol></div></div><div style="font-size: medium;"><div><span style="color: #1a222a; font-family: verdana; font-size: 13.86px;"><span style="font-size: medium;"><b><br /></b></span></span></div><div><span style="color: #1a222a; font-family: verdana; font-size: 13.86px;"><span style="font-size: medium;"><b>C. </b></span></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="background-color: transparent; color: #336699; font-family: verdana; font-size: large; text-decoration-line: none;"><span style="color: #1a222a; font-size: medium;">➡</span></a><span style="color: #1a222a;"> </span><b style="font-size: 18px;"><i><a href="https://neoskeptics.blogspot.com/2023/12/basic-ai-faqs.html"><span style="color: #cc0000; font-family: verdana;">Dozen Basic AI FAQs</span></a></i></b></div><div style="color: #1a222a;"><span style="font-family: verdana;"><span style="font-size: medium;">This page contains links to responses by Google's Bard chatbot running Gemini Pro to 12 questions that should be asked more frequently, but aren't. As consequence, too many readily understood AI terms have become meaningless buzzwords in the media.</span></span></div></div></div></div>
Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-84719408773323068182024-03-06T19:33:00.056-05:002024-03-27T21:36:45.550-04:00Musk sues Altman/OpenAI ... TL;DR and podcast 3Mar24 <div style="text-align: right;"><span style="font-family: verdana; font-size: x-small;"><i>Last update: <span style="color: #cc0000;">Monday 3/11/24 </span></i></span></div><div style="text-align: left;"><span style="font-family: verdana;"><div class="separator" style="clear: both; text-align: center;"><div style="text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><span style="font-size: medium;"><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhQvktvbYUFPumu5F3ohw5Krklb0MUrW_BaQ2rvgwOLHfo_2muauZS7SXjZjFgF1AKJ31CZqUq86RCII2D9CGho_INVr8nRH211NG1vlY8gxF1uFxITKHzj_9UervCzw1r3_GxU7aE9hvbFMS1m3vrjIKQTeWci3tNPWjsLsFfH0Msn7g-uNZiLlfc_rqY/s980/ElonMusk.jpg" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="551" data-original-width="980" height="180" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhQvktvbYUFPumu5F3ohw5Krklb0MUrW_BaQ2rvgwOLHfo_2muauZS7SXjZjFgF1AKJ31CZqUq86RCII2D9CGho_INVr8nRH211NG1vlY8gxF1uFxITKHzj_9UervCzw1r3_GxU7aE9hvbFMS1m3vrjIKQTeWci3tNPWjsLsFfH0Msn7g-uNZiLlfc_rqY/s320/ElonMusk.jpg" vspace="10" width="320" /></a></div>Welcome to our 3Mar</span></span></span></span><span style="font-size: medium;"><span><span><span><span style="background-color: white; color: #1a222a;">24</span><span style="background-color: white; color: #1a222a;"> podcast + </span><span style="background-color: white; color: #1a222a;">TL;DR summary of the past week's top AI stories on our "</span><b><i><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html"><span style="color: #cc0000;">Useful AI News</span></a></i></b><span style="background-color: white; color: #1a222a;">" page. </span></span><span style="background-color: white; color: #1a222a;">Unfortunately the editor of this blog only had time to produce a TL;DR and podcast for only one top story </span></span><b style="background-color: white; color: #1a222a;">➡</b><span style="background-color: white; color: #1a222a;"> Elon Musk Sues OpenA</span></span><span style="background-color: white; color: #1a222a;">I and Sam Altman ... in two parts.</span></span></div><div style="text-align: left;"><span style="background-color: white; color: #1a222a; font-size: large;"><br /></span></div><div style="text-align: left;"><span style="background-color: white; color: #1a222a; font-size: medium;">The first part considers Musk's goal in funding the establishment of OpenAI -- the development of AI via open source code available to everyone; the failure of this strategy; and the likelihood that Musk will lose his lawsuit. </span><span style="background-color: white;"><span style="font-size: medium;"><span style="color: #1a222a;">The second part briefly sketches a high-powered revision of this strategy that replaces Musk with DARPA (</span><b><i><a href="https://www.blogger.com/blog/post/edit/8043580455676201364/8471940877332306818#"><span style="color: #cc0000;">Defense Advanced Research Projects Agency</span></a></i></b></span><span style="color: #1a222a;"><span style="font-size: medium;">) as the sponsor of a dual use initiative.</span><span style="font-size: large;"> </span></span></span></div><div style="text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><br /></span></span></span></div><div style="text-align: left;"><div class="separator" style="clear: both; font-size: 13.86px;"><div style="text-align: right;"><div style="color: #1a222a;"><span style="font-size: medium;">Click link for a podcast (opens in new tab)</span> </div><div><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-size: large; text-align: left; text-decoration-line: none;"><span style="color: #1a222a; font-size: medium;">➡</span></a><span style="color: #1a222a;"> </span><i><span style="font-size: medium;"><b><span style="color: #1a222a;"> 1. </span><a href="https://drive.google.com/file/d/1Q6J1_MBOr-NdB8gtOVEO90HYwDiNvObE/view?usp=sharing" target="_blank"><span style="color: #cc0000;">Musk's failed 2015 AI strategy and subsequent lawsuit</span></a><span style="color: #1a222a;"> </span><br /></b></span></i><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-size: large; text-align: left; text-decoration-line: none;"><span style="color: #1a222a; font-size: medium;">➡</span></a><i><span style="font-size: medium;"><b><span style="color: #1a222a;"> 2. </span><a href="https://drive.google.com/file/d/1Q8P9d6XT0FjXkM5u7WieP_-YyH6-6kGZ/view?usp=sharing" target="_blank"><span style="color: #cc0000;">High-powered 2024 version of Musk's AI strategy</span></a><span style="color: #e06666;"> </span><span style="color: #1a222a;"> </span></b></span></i></div></div></div><div class="separator" style="clear: both;"><div style="text-align: right;"><span style="color: black; font-size: 13.86px;"><span style="color: #1a222a;">Click the "start" button the when the podcast page is loaded<br />... </span></span><span style="color: #1a222a; font-size: 13.86px;"> If audio fails to start, or gets stuck, try reloading that page</span><br /></div></div></div><div style="background-color: white; text-align: right;"><div class="separator" style="clear: both; text-align: left;"><span face="Verdana, Geneva, sans-serif" style="color: #1a222a; font-size: medium;">TL;DR link<b> </b></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-size: large; text-decoration-line: none;"><span style="color: #1a222a;">➡</span></a><b style="font-family: Verdana, Geneva, sans-serif;"><span style="font-size: medium;"><span style="color: #1a222a;"> </span><i><a href="https://neoskeptics.blogspot.com/2024/03/musk-sues-altmanopenai-tldr-and-podcast.html#more"><span style="color: #cc0000;">HERE</span></a></i></span></b></div></div></div></span></div><span><span style="font-family: verdana;"><a name='more'></a></span></span><div style="text-align: left;"><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;">A. </b><b style="background-color: white; font-family: verdana; font-size: large;">TL;DR ...</b><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;"> story in past week ...</b></div><div style="text-align: left;"><br /></div><div style="text-align: left;"><span style="background-color: white; font-family: verdana; font-size: medium;"><b>Elon Musk Sues OpenAI and Sam Altman</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">Elon Musk, a co-founder of Open-AI, is suing OpenAI and Sam Altman because he claims that OpenAI's partnership with Microsoft violates OpenAI's "Founding Agreement". This agreement was made by three people -- Elon Musk, Sam Altman, and Greg Brockman -- when OpenAi was formed in 2015. Brockman became Chairman of OpenAi; Altman became its CEO; and Musk contributed $44 million dollars plus his time and valuable connections to the startup.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Although Musk's complaint has been reported by all of the major media, the editor of this blog believes that his readers will gain far greater insight into the immediate and longer term consequences of this lawsuit by reading a few paragraphs from a few primary sources . </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><b>Muks's complaint</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">The first source is the complaint that Musk filed with the Superior Court of California, County of San Francisco on 2/29/24 that requested a jury trial.</span><span style="font-family: verdana;"> A pdf copy can be found <a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-size: large; text-decoration-line: none;"><span style="color: #1a222a; font-size: medium;">➡</span></a> </span><i style="font-family: verdana; font-weight: bold;"><a href="https://www.courthousenews.com/wp-content/uploads/2024/02/musk-v-altman-openai-complaint-sf.pdf"><span style="color: #cc0000;">HERE</span></a>.</i></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><i style="font-family: verdana; font-weight: bold;"></i><span style="font-family: verdana;">The following quotes from this document convey the essence of Musk's complaint. Note that "the three" referred to in the first line of paragraph 24 refers to Musk, Altman, and Brockman.</span></div><blockquote><div style="text-align: left;"><ul style="text-align: left;"></ul><span style="font-family: verdana;">24. </span><span style="font-family: verdana;">In March 2023, "</span><span style="font-family: verdana;">Together with Mr. Brockman, the three agreed that this new lab: (a) would be a nonprofit developing AGI for the benefit of humanity, not for a for-profit company seeking to maximize </span><span style="font-family: verdana;">shareholder profits; and (b) would be open-source, balancing only countervailing safety </span><span style="font-family: verdana;">considerations, and would not keep its technology closed and secret for proprietary commercial reasons (The “Founding Agreement”). Reflecting the Founding Agreement, Mr. Musk named this new AI lab “OpenAI,” which would compete with, and serve as a vital counterbalance to, Google/DeepMind in the race for AGI, but would do so to benefit humanity, not the shareholders of a private, for-profit company (much less one of the largest technology companies in the world)"</span></div></blockquote><div style="text-align: left;"><span style="font-family: verdana;">Is GPT-4 really an AGI algorithm? Musk's complaint goes on to assert that Microsoft's own researchers have concluded that it is:</span></div><div style="text-align: left;"><blockquote><span style="font-family: verdana;">31. "Furthermore, on information and belief, GPT-4 is an AGI algorithm, and hence </span><span style="font-family: verdana;">expressly outside the scope of Microsoft’s September 2020 exclusive license with OpenAI. In this </span><span style="font-family: verdana;">regard, Microsoft’s own researchers have publicly stated that, “[g]iven the breadth and depth of </span><span style="font-family: verdana;">GPT-4’s capabilities, we believe that it could reasonably be viewed as an early (yet still incomplete) </span><span style="font-family: verdana;">version of an artificial general intelligence (AGI) system.</span><span style="font-family: verdana;">"</span></blockquote></div><div style="text-align: left;"><span style="font-family: verdana;"><b>Microsoft Research</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">Here's a link to a pdf copy of the Microsoft Research report to which Musk referred in his complaint:</span><span style="font-family: verdana;"><b><i></i></b></span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">"Sparks of Artificial General Intelligence", Sebastien Bubeck, Varun Chandrasekaran, et al., <b><i><a href="https://drive.google.com/file/d/1v_0T0m4TsMZUTqp3ToBxgHzRFLrTRc-A/view?usp=sharing"><span style="color: #cc0000;">Microsoft Research</span></a></i></b>, 4/13/23</span></li></ul></div><div style="text-align: left;"><span style="font-family: verdana;"><span>Readers whose time is strictly limited should merely consider the four examples in the document's "F</span>igure 1.1: Preliminary examples of GPT-4’s capabilities in language, vision, coding, and mathe</span><span style="font-family: verdana;">matics"</span><span style="font-family: verdana;"> </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Readers who have more tune should skim the rest of this long document for many more examples of stunning emergent cognitive skills that a pre-release version of GPT-4 displayed before the final guidelines imposed limits on the skills of the released final version. (Readers are reminded that "emergent" skills are capabilities that a model displays that it was not programmed to display.)</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><b>Probable court decision</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">Musk will probably lose this civil case. </span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">First, his assertion that GPT-4 was an AGI was flatly contradicted by the same experts in the same report that Musk citied to support his allegation. </span><span style="font-family: verdana;">The "Sparks" in the title of the report from Microsoft Research reflects the authors' concerns that the pre-release version of GPT-4 did not display all of the cognitive skills that might be expected in an AGI system, i.e., that it was only "</span><span style="font-family: verdana;">an early (yet still incomplete) </span><span style="font-family: verdana;">version of an artificial general intelligence (AGI) system."<br /><br /></span></li><li><span style="font-family: verdana;">Second, his filing does not include a copy of the</span><span style="font-family: verdana;"> “Founding Agreement” or a link to such a formal document. If Altman claims that Musk misunderstood whatever agreement had been made, the court would be facing a "He said, he said" situation.</span></li></ul></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">Third, does Musk have what lawyers call "standing" that gives him the right to file suit against Altman and OpenAI, i.e., how was Musk harmed by OpenAI's partnership with Microsoft?<br /><br /></span></li><li><span style="font-family: verdana;">Fourth, given that Microsoft had to invest $13 billion into OpenAI within three years to reap the success of GPT-4, it seems likely that Musk's original $44 million investment was an order of magnitude too small to support a comparably successful outcome, even if it was just a down-payment with more payments to come within three or four years.</span></li></ul></div><div style="text-align: left;"><span style="font-family: verdana;">But even if, somehow, the court decides in Musk's favor, it might compensate him by requiring OpenAI to refund his $44 million dollar investment plus a hefty penalty. Let's think really big here; let's suppose the court decides that Open AI and Altman together must provide Musk with one billion dollars. </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">After going through time consuming motions of objections and appeals, OpenAI and Altman would probably pay the billion dollars, then get on with their lucrative multi-billion partnership with <b>Microsoft</b>. Where would Altman and OpenAI find one billion dollars to protect their lucrative partnership with <b>Microsoft</b>? ... Hmmmmmm ... Can you think of a lucrative three trillion dollar corporation that might be willing to fund a billion dollar payment that would preserve OpenAI's lucrative partnership with <b>Microsoft</b>? ... Hmmmmmm</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><b>Longer term consequences</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">The longer term consequences of Musk's lawsuit, no matter whether he wins or loses, are far more consequential because he filed his suit within the context of massive shifts in two underlying conceptual frameworks: </span></div><div style="text-align: left;"><ol style="text-align: left;"><li><span style="font-family: verdana;">The increasing appetite for substantial regulation of Big Tech in the US, in the UK, and in the EU ... and <br /><br /></span></li><li><span style="font-family: verdana;">The growing consensus among genAI (generative AI) experts that AGI is likely to emerge much sooner than previously predicted, i.e., not twenty years from now, but within the next five to ten years. </span></li></ol></div><div style="text-align: left;"><span style="font-family: verdana;">The unexpected emergence of AGI much sooner than previously anticipated means that the speed with which we are developing AGi now greatly exceeds the speed with which we can develop effective policies for regulating genAI technology</span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">Do any readers of this blog over 12 years of age really believe that the U.S. elected officials can formugate effective regulations for any highly impactful new technology, such as genAI, in less than five years? The EU might be able to develop effective policies faster than the U.S. or the U.K., but EU policies are unlikely to be adopted quickly by the (disunited) U.S. (Congress) or the (Brexit) UK.</span></li></ul></div><div style="text-align: left;"><span style="font-family: verdana;">Nevertheless, extensive public discussions of Musk's lawsuit in the media during the coming weeks or months will provide substantial opportunities for our policymakers to consider the implications of his primary concern that for-profit companies cannot be trusted to develop AGI in the public interest. </span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">For-profit Big Tech companies are likely to exaggerate the potential value of their current AI technology, but minimize or even hide their tech's vulnerabilities because honest disclosures would depress future sales. <br /><br /></span></li><li><span style="font-family: verdana;">For-profit Big Tech companies are also unlikely to recommend strong regulations when consulted by legislators ... unless their recommendations provide such high bars of entry into the market that they substantially inhibit the participation of smaller potential competitors ... and unless their recommendations don't impose substantial inhibitions on Big Tech's current AI technology.</span></li></ul></div><div style="text-align: left;"><span style="font-family: verdana;">Back in 2015, Musk didn't trust Google, Big Tech's AI leader at the time. Being a conservative, he didn't give serious consideration to the imposition of regulations. Instead, he funded the establishment of a non-profit developer of AI that would compete with Google and thereby act as a check on Google's efforts to dominate the evolution of this powerful technology. His efforts failed. Ironically, OpenAI was not captured by Google, but by Microsoft, a Big Tech giant that was running a distant third to Google back in 2015, behind Amazon.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Musk's strategy failed, but had it succeeded, OpenAI would have become a highly prized source of insightful recommendations for our legislators. Could a stable, well funded, non-profit competitor to Big Tech be created today? </span><span style="font-family: verdana;">This paragraph marks the end of this week's TL;DR. </span><span style="font-family: verdana;">The next section of this blog note provides a brief sketch of a more effective new competitor, but one that requires another major shift in our underlying conceptual framework.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><b style="font-size: large;">End of TL;DR</b></span></div><div style="text-align: center;"><b style="font-family: verdana; font-size: large; text-align: left;">... TL;DR Post Script ...</b></div><div style="text-align: center;"><i style="color: #1a222a; font-family: verdana; font-size: 13.86px; text-align: right;"><span style="font-size: medium;"><b>High-powered 2024 revision of Musk's AI strategy</b></span></i></div><div style="text-align: left;"><span style="font-family: verdana;"><b><br />The weaponization of genAI</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">Demis Hassabis, CEO of Google DeepMind, dedicated his career to transforming AI into tools that scientists could use to address mankind's biggest challenges. Kevin Scott, Microsoft's CTO, sees genAI chatbots as great levelers that enable ordinary people who have no AI expertise to be able to use AI as a tool to increase their workplace productivity and/or to identify ways to live more satisfying lives. </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">For some reason, the biggest cloud over both ambitions has been determined to be the low probability event that AGI, the most advanced form of genAI, might spontaneously evolve into an entity that was no longer a tool for humans, but an existential threat to all mankind. </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">But before genAI can generate AGI, there is an undeniably high probability, nay, a certainty, that the most powerful genAI will be weaponized, i.e., converted into weapons that nation states or non-state entities can use to demolish other nation states or non-state entities. Why? Because that's what we humans usually do: we turn powerful technologies into powerful weapons that we turn on our adversaries.</span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">Do any readers of this blog seriously doubt that North Korea, China, and Russia are already laying out roadmaps for developing genAI weapons that can be used against the U.S.? And does anybody seriously doubt that the Pentagon is quietly framing the development of genAI-based weapons that could be directed against North Korea, China, and Russia? <br /><br />Skeptics are referred to a recent article in the Washington Post:<br />-- "Pentagon explores military uses of large language models", Eva Dou, Nitasha Tiku and Gerrit De Vynck, <b><i><a href="https://www.washingtonpost.com/technology/2024/02/20/pentagon-ai-llm-conference/"><span style="color: #cc0000;">Washington Post</span></a></i></b>, 2/20/24<br /><br /></span></li><li><span style="font-family: verdana;">Some readers may recall the "fire sale" that was the focal event of the 2007 Bruce Willis "Die Hard 4" movie. What's a "fire sale"? <b><i><a href="https://www.youtube.com/watch?v=GpUkpPoVP2g&pp=ygUVZGllIGhhcmQgNCBmaXJlIHNhbGUg"><span style="color: #cc0000;">Please watch this brief YouTube clip to find out</span></a></i></b>. How could a "fire sale" be triggered? Back then, via lots of pseudo-tech movie double talk. But today? GPTs, baby, high powered G ...P ... Ts.</span></li></ul></div><div style="text-align: left;"><span style="font-family: verdana;">Unfortunately, legislators responsible for funding the development of new high powered weapons that will destroy our adversaries will receive the same inappropriate self-serving responses from Big Tech as were the legislators seeking to formulate regulations that would prevent the development of an AGI that might destroy all of humanity. That's why another big shift in our underlying conceptual framework is required.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><b>Dual use technologies</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">Many decades ago, DARPA -- the <b><i><a href="https://www.darpa.mil/"><span style="color: #cc0000;">Defense Advanced Research Projects Agency</span></a></i></b> in the U.S. Department of Defense -- recognized that our defense contractors were sometimes far less cost-effective in developing new applications of new technologies than were the companies in our economy's civilian sector, even when the new technologies had obvious applications in the defense sector. This insight was reinforced by DARPA's identification of many examples of things like "$800 toilet seats" that had been designed by defense contractors. </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">The first correction in purchasing strategy was straight-forward: Buy cost-effective products from the civilian sector, then "<b><i><a href="https://www.merriam-webster.com/dictionary/ruggedize"><span style="color: #cc0000;">ruggedize</span></a></i></b>" them and "<b><i><a href="https://www.merriam-webster.com/dictionary/militarize"><span style="color: #cc0000;">militarize</span></a></i></b>" them to make them useful in the defense sector. But that's the low hanging fruit</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Now what about new technologies that had high potential value in the civilian sector and in the defense sector, but the venture capitalists (VCs) in the civilian sector regarded them as too risky, so the VCs declined to invest invest in their development. No low hanging fruit. If the defense contractors were tasked with developing these risky new technologies, they would probably produce more "$800 toilet seats". So what to do? What to do?</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Being one of the cleverest agencies in the U.S. government, no, being one of the cleverest agencies in the galaxy, DARPA came up with a clever new strategy. It awarded substantial contracts to private sector firms to develop new technologies in their earliest phases, when the risks of failure were highest. It continued to fund the best private developers until the risks of failure in the next phase were low enough to attract investments from VCs. Then DARPA stepped back and let the VCs compete for the chance to fund the developers' final phases. When the products based on the new technology went on the civilian market, the Department of Defense bought the inexpensive new low hanging fruits, then ruggedized them and militarized them to make them useful tools for our troops.</span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">Readers unfamiliar with DARPA's achievements may not know that DARPA funded the pre-competitive, high risk development of the Internet, robots, self-driving cars, high powered workstations, GPS, high definition flat panel displays, parallel processing, computer mice, high performance computing (a/k/a supercomputers), etc, etc, etc</span></li></ul></div><div style="text-align: left;"><span style="font-family: verdana;">So much for context. The final paragraphs of this note sketch an updated and, hopefully, more effective version of Musk's failed strategy.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><b><br /></b></span></div><div style="text-align: left;"><span style="font-family: verdana;"><b>The Open Source Generative AI Development Consortium (OSGAIDC)</b></span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;"><b>Members</b><br />DARPA will convene a consortium of leading U.S. universities whose faculty include the most esteemed AI experts in the U.S.<br /><br /></span></li><li><span style="font-family: verdana;"><b>Mission</b><br />The mission of the consortium will be the development of generative AI via o<span style="background-color: white; color: #5f6368;">pen</span><span style="background-color: white; color: #4d5156;">-</span><span style="background-color: white; color: #5f6368;">source</span><span style="background-color: white; color: #4d5156;"> software that is published and made available to the public. None of the consortium's activities or outputs will be classified<br /><br /></span></span></li><li><span style="font-family: verdana;"><b>Duration</b></span><br /><span style="font-family: verdana;">Five years.<br /><br /></span></li><li><span style="font-family: verdana;"><b>Funding</b></span><br /><span style="font-family: verdana;">-- DARPA will provide $20 billion per year to the consortium for five years. <br />-- Member universities whose faculty take full or partial leave from their current teaching responsibilities will receive funds from the consortium to hire additional faculty to assume the teaching responsibilities of their on-leave faculty members<br /><br /></span></li><li><span style="font-family: verdana;"><b>Faculty, support personnel, and other resources<br /></b>-- The consortium will pay the on-leave faculty members no more than four times the hourly rate they are paid by their universities ... This "pay raise" will provide incentives for their participation in the consortium's activities, rather than serving as highly paid AI consultants for Big Tech firms <br />-- The consortium's full-time non-faculty techs and other support personnel will be paid competitive market rates for their services<br /><br />-- Reference: "Silicon Valley is pricing academics out of AI research", Naomi Nix, Cat Zakrzewski and Gerrit De Vynck, <b><i><a href="https://www.washingtonpost.com/technology/2024/03/10/big-tech-companies-ai-research/?utm_campaign=wp_todays_headlines&utm_medium=email&utm_source=newsletter&wpisrc=nl_headlines&carta-url=https%3A%2F%2Fs2.washingtonpost.com%2Fcar-ln-tr%2F3d085ab%2F65eed626d347a964a2d552d1%2F596bf8beade4e24119bf6e07%2F32%2F55%2F65eed626d347a964a2d552d1"><span style="color: #cc0000;">Washington Post</span></a></i></b>, 3/10/24<br /><br /></span></li><li><span style="font-family: verdana;"><b>Advisory responsibilities</b><br />From time to time DARPA may require participating faculty members to share their expertise with legislators and other key federal decision makers.<br /><br /></span></li><li><span style="font-family: verdana;"><b>Acquisition of scarce non-personnel resources via the Defense Production Act</b><br />Given DARPA's goal to facilitate the development of powerful open source genAI software that has dual use, i.e. civilian and defense applications, the consortium's success will have a profound positive impact on our national security. DARPA will therefore advise the administration to invoke the <b><a href="https://www.cfr.org/in-brief/what-defense-production-act"><i><span style="color: #cc0000;">Defense Production Act</span></i></a> </b>to secure scarce resources for the consortium at (negotiated) below market rates, e.g., computing time to run models on expensive chips in the cloud.</span></li></ul></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><div style="background-color: white;"><div><div id="Overview"><span><span style="font-family: verdana;"><b><span style="color: black; font-size: medium;">B. </span><span style="color: black; font-family: Times; font-size: medium; font-weight: 400;"><span style="font-family: verdana;"><b style="font-size: large;">Top story in past week ...</b></span></span><span style="color: black; font-family: Times; font-size: medium; font-weight: 400;"> </span><div id="Overview" style="font-family: Times;"><ol><div style="color: black; font-size: medium; font-weight: 400;"><span style="font-family: verdana;"></span></div><li><span style="font-family: verdana;"><span style="color: black; font-weight: 700; text-align: center;"><a href="#Policy"><span style="color: #cc0000; font-size: x-small;">Public Policy</span></a></span><span style="color: black; font-size: small; font-weight: 700; text-align: center;"> </span><br /><span style="font-size: medium;"><span style="color: black; font-weight: 400;">"Elon Musk Sues OpenAI and Sam Altman for Violating the Company’s Principles", Adam Satariano, Cade Metz and Tripp Mickle, </span><b style="color: black;"><i><a href="https://www.nytimes.com/2024/03/01/technology/elon-musk-openai-sam-altman-lawsuit.html"><span style="color: #cc0000;">NY Times</span></a></i></b><span style="color: black; font-weight: 400;">, 3/1/24 </span><span style="color: #cc0000; font-weight: 400;">***</span><span style="color: black; font-weight: 400;"> </span><br /><span style="color: black; font-weight: 400;">-- A copy of Musk's electronically filed complaint can be found </span><i><a href="https://www.courthousenews.com/wp-content/uploads/2024/02/musk-v-altman-openai-complaint-sf.pdf"><span style="color: #cc0000;">HERE</span></a></i><br /><span style="color: black; font-weight: 400;">-- This story also covered by </span><b style="color: black;"><i><a href="https://on.ft.com/3T24QgG"><span style="color: #cc0000;">Financial<span style="font-weight: 400;"> </span>Times</span></a></i></b><span style="color: black; font-weight: 400;">, </span><b style="color: black;"><i><a href="https://t.co/w9yBQ38gLP"><span style="color: #cc0000;">Wired</span></a></i></b><span style="color: black; font-weight: 400;">,</span><span style="color: black;"> </span><b style="color: black;"><i><a href="https://t.co/YB7TYHpcdO"><span style="color: #cc0000;">Bloomberg</span></a></i></b><span style="color: black;">, </span><b style="color: black;"><i><a href="https://t.co/wwWvYJPZU3"><span style="color: #cc0000;">Engadget</span></a></i></b><span style="color: black;">, </span><b style="color: black;"><i><a href="https://t.co/XbgMbjBmwD"><span style="color: #cc0000;">Washington Post</span></a></i></b><span style="color: black;">, </span><b style="color: black;"><i><a href="https://t.co/uItdNK4aVG"><span style="color: #cc0000;">The Verge</span></a></i></b><span style="color: black;">, </span><b style="color: black;"><i><a href="https://t.co/HBwo5TtnCW"><span style="color: #cc0000;">Wall Street Journal</span>,</a></i></b><span style="color: black;"> </span><b style="color: black;"><i><a href="https://t.co/ebQGKQJkYT"><span style="color: #cc0000;">Reuters</span></a></i></b><span style="color: black;">, </span><b style="color: black;"><i><a href="https://t.co/3qqoVmJtK3"><span style="color: #cc0000;">BBC</span></a></i></b><span style="color: black;">, </span><b style="color: black;"><i><a href="https://t.co/Tv54aSGr2A"><span style="color: #cc0000;">TechCrunch</span></a></i></b><span style="color: black;">, </span><b style="color: black;"><i><a href="https://t.co/2pZETQpPVR"><span style="color: #cc0000;">VentureBeat</span></a></i></b><span style="color: black;">, </span><b style="color: black; font-size: medium;"><i><a href="https://www.nytimes.com/2024/03/02/technology/elon-musk-openai-lawsuit-paradox.html?smid=nytcore-android-share"><span style="color: #cc0000;">NY Times #2</span></a></i></b><span style="color: black; font-weight: 400;"> </span><br /><br /></span></span></li></ol></div><br /></b></span></span></div></div><div style="font-size: medium;"><br /></div></div></div>
Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-60655680217301548092024-02-26T16:30:00.046-05:002024-03-27T21:36:57.316-04:00Google's Gemma, Intel chips designed by Microsoft, and Gemini's "diverse" hallucinations ... TL;DR and podcast 25Feb24<div style="text-align: right;"><span style="font-family: verdana; font-size: x-small;"><i>Last update: Sunday 2/25/24 </i></span></div><div style="text-align: left;"><span style="font-family: verdana;"><div class="separator" style="clear: both; text-align: center;"><div style="text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><span style="font-size: medium;"><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjU9jL9m7Pj6D_LxSaCbdV6Ft9VCdh9Pr6jeBEG7z9RVUhZBWcTVnzSOVvBK_QY6wySNjpQEJvcjKwt5bWruLcB5E_gsjYrGThiJVrc3gzpW2XbPRg1U2k7ymhOzdcryTrBZhJFTtHhdoYesERSBh44E3_X-ZHv-i_Yr7NWjWigLfm6ttVhDITxNnBeN7s/s920/GemmaFreeOpenSourceSLMGoogle2.jpeg" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="518" data-original-width="920" height="180" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjU9jL9m7Pj6D_LxSaCbdV6Ft9VCdh9Pr6jeBEG7z9RVUhZBWcTVnzSOVvBK_QY6wySNjpQEJvcjKwt5bWruLcB5E_gsjYrGThiJVrc3gzpW2XbPRg1U2k7ymhOzdcryTrBZhJFTtHhdoYesERSBh44E3_X-ZHv-i_Yr7NWjWigLfm6ttVhDITxNnBeN7s/s320/GemmaFreeOpenSourceSLMGoogle2.jpeg" vspace="10" width="320" /></a></div>Welcome to our 25Feb</span></span></span></span><span style="font-size: medium;"><span><span style="background-color: white; color: #1a222a;">24</span><span style="background-color: white; color: #1a222a;"> </span><span style="background-color: white; color: #1a222a;">TL;DR summary + podcast about the past week's top 3 AI stories on our "</span><b style="color: #1a222a;"><i><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html">Useful AI News</a></i></b><span style="background-color: white; color: #1a222a;">" page </span></span><b style="background-color: white; color: #1a222a;">➡</b><span style="background-color: white; color: #1a222a;"> 1) Google'e new open source model Gemma, (2) Microsoft's chip partnership with Intel, and (3) Gemini's "woke" hallucinations ... Plus links to other notable new reports.</span></span></div><div style="text-align: left;"><br /></div><div style="text-align: left;"><div class="separator" style="clear: both; font-size: 13.86px;"><div style="text-align: right;"><div style="color: #1a222a;"><span style="font-size: medium;">Click link for podcast (opens in new tab)</span> </div><div><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-size: large; text-align: left; text-decoration-line: none;"><span style="color: #1a222a; font-size: medium;">➡</span></a><b><i><a href="https://drive.google.com/file/d/1tKc0S8twy1MMbRRVJxvVbeO0F8hKy5Rl/view?usp=sharing"><span style="color: #cc0000;"> </span></a></i></b><i><span style="font-size: medium;"><b><a href="https://drive.google.com/file/d/1PzWAfq7-iN-vZpig_pKsLRtqBPdI7hav/view?usp=sharing" target="_blank"><span style="color: #cc0000;">Google + Microsoft + Google</span></a><span style="color: #1a222a;"> </span></b></span></i></div></div></div><div class="separator" style="clear: both;"><div style="text-align: right;"><span style="color: black; font-size: 13.86px;"><span style="color: #1a222a;">Click the "start" button the when the podcast page is loaded<br />... </span></span><span style="color: #1a222a; font-size: 13.86px;"> If audio fails to start, or gets stuck, try reloading that page</span><br /></div></div></div><div style="background-color: white; text-align: right;"><div class="separator" style="clear: both; text-align: left;"><div style="font-family: Times;"><span style="font-family: verdana;"><div class="separator" style="clear: both; text-align: center;"><div style="text-align: right;"><div class="separator" style="clear: both; text-align: left;"><span face="Verdana, Geneva, sans-serif" style="color: #1a222a; font-size: medium;">TL;DR link<b> </b></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-size: large; text-decoration-line: none;"><span style="color: #1a222a;">➡</span></a><b style="font-family: Verdana, Geneva, sans-serif;"><span style="font-size: medium;"><span style="color: #1a222a;"> </span><i><a href="https://neoskeptics.blogspot.com/2024/02/googles-gemma-intel-chips-designed-by.html#more"><span style="color: #cc0000;">HERE</span></a></i></span></b></div></div></div></span></div><span style="font-family: Times;"><span style="font-family: verdana;"></span></span></div></div></div></span></div><span><span style="font-family: verdana;"><a name='more'></a></span></span><div style="text-align: left;"><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;">A. </b><b style="background-color: white; font-family: verdana; font-size: large;">TL;DR ... </b><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;">Top 3 stories in past week ... TL;DR</b></div><div style="text-align: left;"><b><span style="font-family: verdana; font-size: medium;"><br /></span></b></div><div style="text-align: left;"><b><span id="gemma" style="font-family: verdana; font-size: medium;">1) Gemma: </span></b><span style="background-color: white; color: #1a222a; font-family: verdana;"><b>Google's new open source model</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">This blog has repeatedly noted that Google has been striving mightily to catch up to the Microsoft/OpenAI partnership ever since the partners released their GPT-4 large language model in March 2023. No one seriously doubts that Google is capable of closing this gap. Nevertheless, the impact of Google's gains have been repeatedly undermined by the fact that the partner have posed a moving target. </span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">For example, Google announced Gemini Nano, a small language model, in late December 2023. </span><span style="font-family: verdana;">Then Microsoft announced a far more <b><i><a href="https://www.microsoft.com/en-us/research/blog/phi-2-the-surprising-power-of-small-language-models/"><span style="color: #cc0000;">powerful small model, Phi-2</span></a></i></b>, (2.7 billion parameters) in January 2024. Microsoft's announcement included the results of benchmark tests that showed that Phi-2 was far more powerful than Google's Nano and other small models, like Llama2 and Mistral. Microsoft's announcement also declared that it would make <span style="box-sizing: border-box;">Phi-2</span><span class="sr-only" style="border: 0px; box-sizing: border-box; clip: rect(1px, 1px, 1px, 1px); height: 1px; margin: -1px; overflow: hidden; padding: 0px; position: absolute; text-wrap: nowrap; width: 1px;">(opens in new tab)</span><span face=""Segoe UI", SegoeUI, "Helvetica Neue", Helvetica, Arial, sans-serif" style="background-color: white; box-sizing: border-box; font-size: 16px; font-weight: bolder;"> </span><span face=""Segoe UI", SegoeUI, "Helvetica Neue", Helvetica, Arial, sans-serif" style="background-color: white; font-size: 16px;">available in the Azure AI Studio model catalog to foster research and development on language models.</span><br /></span><br /></li><li><span style="font-family: verdana;">Indeed, Microsoft determined that small models could have such surprising power that CEO Nadella subsequently announced that he had established a crash project that will deploy Microsoft's best AI experts to <b><i><a href="https://www.theinformation.com/articles/microsoft-forms-team-to-make-cheaper-generative-ai?rc=geicgp"><span style="color: #cc0000;">develop small language models</span></a></i></b> that will be more cost-effective than the large language models that Microsoft previously paid OpenAI billions of dollars to develop. This team will report directly to Kevin Scott, Microsoft's Chief Technology Officer</span></li></ul></div><div style="text-align: left;"><span style="font-family: verdana;">Last week, Google announced </span><span style="font-family: verdana;">Gemma, a new family of small language models. </span><span style="font-family: verdana;">According to </span><b style="font-family: verdana;"><i><a href="https://www.blogger.com/blog/post/edit/8043580455676201364/6065568021730154809#"><span style="color: #cc0000;">Google's blog</span></a></i></b><span style="font-family: verdana;">: </span></div><div style="text-align: left;"><blockquote><span style="font-family: verdana;">"Gemma is a family of lightweight, state-of-the-art open models built from the same research and technology used to create the Gemini models ... We’re releasing model weights in two sizes: Gemma 2B and Gemma 7B ... Gemma models share technical and infrastructure components with Gemini, our largest and most capable AI model widely available today. This enables Gemma 2B and 7B to achieve best-in-class performance for their sizes compared to other open models."</span></blockquote><p><span style="font-family: verdana;">Was Gemma Google's response to Microsoft's efforts to develop small but powerful models that could replace OpenAI's expensive GPT family of large models? Not really. </span></p><p></p><ul style="text-align: left;"><li><span style="font-family: verdana;">We assume that Gemini Nano is more powerful than Gemma. The benchmarks cited by Microsoft in its announcement of Phi-2 show that Phi-2 is more powerful than Nano by a wide margin; so we infer that Phi-2 is also far more powerful than Gemma.<br /><br /></span></li><li><span style="font-family: verdana;">Indeed, the benchmarks cited by Google in its announcement of Gemma showed that Gemma was only marginally more powerful than Llama and Mistral. </span></li></ul><div><span style="font-family: verdana;">So why did Google announce such a weak small language model? The editor of this blog is inclined to believe that the announcement reflects Dr. Hassabis' life-long commitment to the scientific method, the same scientific method he employed so successfully in DeepMind's development of AlphaGo, AlphaFold, and other breakthroughs. (See links to articles describing a few of DeepMind's achievements in section D of this note, below, before it was merged with "Google Brain" to become "Google DeepMind".) </span></div><div><ul style="text-align: left;"><li><span style="font-family: verdana;">"A.I. Could Solve Some of Humanity's Hardest Problems. It Already Has.". Guest = Demis Hassabis, <a href="https://www.nytimes.com/2023/07/11/opinion/ezra-klein-podcast-demis-hassabis.html"><b><i><span style="color: #cc0000;">The Ezra Klein Show</span></i></b></a></span><span style="font-family: verdana;"> (podcast with transcript), 7/11/23 </span></li></ul></div><div><span style="font-family: verdana;">In traditional science, even marginal achievements are shared with the other members of one's research community. But this would not justify Google's full blast, </span><span style="font-family: verdana;">sales-pitch</span><span style="font-family: verdana;"> assertion that the Gemma models can "</span><span style="font-family: verdana;">achieve best-in-class performance for their sizes" ... No, they can't. They place a very distant second to Phi-2.<br /></span></div><div style="text-align: center;"><span style="font-family: verdana;"><b style="background-color: white; font-size: large;"><i><span style="color: #cc0000;"><br /></span></i></b></span></div></div><div style="text-align: left;"><b><span style="font-family: verdana; font-size: medium;">2) </span></b><span id="intel" style="font-family: verdana; font-size: medium;"><b>Intel chips designed by Microsoft</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">The opening paragraph of the article in <b><i><a href="https://t.co/khNH6PHP5C"><span style="color: #cc0000;">The Verge</span></a></i></b> provides a succinct summary of this story:</span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">"Microsoft and Intel strike a custom chip deal that could be worth billions. With the Microsoft deal, Intel notches a big partnership as it seeks to regain its former position at the top of chip manufacturing</span>."</li></ul></div><div style="text-align: left;"><span style="font-family: verdana;">Indeed, the article goes on to say Intel expects $15 billion from this order. Now let's provide some context for this partnership:</span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">Older readers will remember a time when Intel made most of the chips in computers smaller than mainframes and designed all of the chips that it produced. However, in recent years Intel has ceded its leadership position to other producers.<br /><br /></span></li><li><span style="font-family: verdana;">Nowadays the leading chip makers, called "foundries", produce chips that were designed by their customers. Last year's global surge of interest in AI propelled the Taiwan Semiconductor Manufacturing Company (TSMC) to the top of the chip heap because TSMC's fabrication facilities ("fabs") manufacture the chips designed by Nvidia.<br /><br /></span></li><li><span style="font-family: verdana;">The large language models that everyone wants to develop nowadays must be run on graphics processing units (GPUs) that enable multiple instructions to be executed simultaneously. Nvidia makes the most powerful GPUs, so the demand for Nvidia's chips is running far higher than the limited supply that TSMC can provide. This gap has driven the price of Nvidia's chips and the market value of Nvidia's shares so high that Nvidia is now the third most valuable company in the world, behind Apple and Microsoft.</span></li></ul></div><div style="text-align: left;"><span style="font-family: verdana;">To escape this price inflation, Microsoft is now joining Amazon, Google, and Apple, all or whom have been designing their own chips in recent years. But Microsoft will not only design its own chips; it will not have them produced by TSMC, a company whose production lines are already operating at maximum capacity, but by Intel, an underutilized producer that desperately wants to resume its position at the top of the chips heap.</span></div><div style="text-align: center;"><span style="font-family: verdana;"><b style="background-color: white; font-size: large;"><i><span style="color: #cc0000;"><br /></span></i></b></span></div><div style="text-align: left;"><span style="font-size: medium;"><b><span style="font-family: verdana;">3) </span></b><span id="woke" style="background-color: white; color: #1a222a; font-family: verdana;"><b>Gemini's "woke" hallucinations</b></span></span></div><div style="text-align: left;"><span style="font-family: verdana;">The headline from Business Insider captures the essence of this story:</span></div><div style="text-align: left;"><blockquote><span style="font-family: verdana;">"Google suspends Gemini from making AI images of people after a backlash complaining it was 'woke'"</span></blockquote></div><div style="text-align: left;"><span style="font-family: verdana;">Yes, when asked to produce images of the Founding Fathers, Gemini produced pictures that included Black men. It seems likely that Google has instructed Gemini to produce images that are racially diverse whenever possible. Unfortunately, it also seems that Gemini continued to produce "diverse" groups, even when corrected to do otherwise. Of course the anti-"woke" crowd cried out in anger and anguish. But Google quickly suspended Gemini's capacities to produce any images of people until this defect is corrected. </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">If AI's generation of images was anywhere near as good as its generation of text or code, Google's temporary suspension and quick resumption would end this story. But as many of us know, AI generation of images by any chatbot is still in early phases of development, even for ChatGPT plus DALL-E, a flawed combo that's been accessible to the public for almost a year. </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">As hard as it tries to play catch-up to Microsoft/OpenAI, Google will probably be playing this kind of whack-a-mole game for the foreseeable future, with unexpected glitches popping up here and there and all over the place. However, no one should doubt that, given its determination, its depth of AI expertise, and its vast finnncial resources, that Google will, indeed, catch up, sooner or later. </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">So the real question is: catch up to what? Whenever Google stumbles, we should immediately ask: How well has Microsoft/OpenAi addressed this issue? With regards to AI generated images, the honest answer today must be: not very well, not yet. To be sure, the animators and other visual creatives are correct in their perception that AI generated images represent an imminent threat to their livelihoods. But the threat is still imminent, not a current reality ... yet. </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">The editor of this blog decided to test his pessimism by asking ChatGPT Plus to generate a group image. Here's his initial prompt, followed by two correction prompts, then the final images:</span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;"><b>Original prompt:</b><br />"Draw a picture of five white American males in their mid-twenties, sitting around a table, with big smiles on their faces. They are casually dressed, wearing jeans, sandals, and colorful sports shirts. There is a large birthday cake with lighted candles in the center of the table because they are celebrating the birthday of the youngest member of the group"<br /><br /></span></li><li><span style="font-family: verdana;"><b>Correction prompt #1</b><br />"Please correct the image to show five white males wearing sandals and colorful shirts"<br /><br /></span></li><li><span style="font-family: verdana;"><b>Correction prompt #2<br /></b>"One more correction. They all should be seated at the table; no one should be standing</span>"</li></ul><span style="font-family: verdana; font-weight: bold;">Final images returned<br /><br /></span><div class="separator" style="clear: both; font-weight: bold; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhgTYJZqM6O5lcN0jwQ5pkBVXc44IUBzP5GNz9MBD-V6oYtxrk-phBTcV-ecVJabvk3yHWvIQ7bswKZJO59618hOzKG-ZOfElYratrGGiZ0x8KsAo5zmg7IJgXTMRvBlp4tZ-vumfIDBa5C7jn80UBS-nHNH0dLI6NAT7bd6tXfOJp6KZCMPvcEYl9G5X8/s1024/testingDALL-EMyPromptsAndResponsesRight.jpeg" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1024" data-original-width="1024" height="320" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhgTYJZqM6O5lcN0jwQ5pkBVXc44IUBzP5GNz9MBD-V6oYtxrk-phBTcV-ecVJabvk3yHWvIQ7bswKZJO59618hOzKG-ZOfElYratrGGiZ0x8KsAo5zmg7IJgXTMRvBlp4tZ-vumfIDBa5C7jn80UBS-nHNH0dLI6NAT7bd6tXfOJp6KZCMPvcEYl9G5X8/s320/testingDALL-EMyPromptsAndResponsesRight.jpeg" width="320" /></a></div><br /><div class="separator" style="clear: both; font-weight: bold; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhrWLq5COuKTx1Fay6PV3oQ3UYu-mGtAaDZ1-LykK29bzlnC-FEtULaChnVdfmDHd6kcs-rr04VeNHF-AaE0xe9k1KZ_f4ADm_bQN_n4QUPGsbLwFAH1cTqoccECPkgYsiBmnAzsSilNSBaS3wHTucRfbgBAnxL1m77_Efz37NOA9KokuG2r83PBUsOMK0/s1024/testingDALL-EMyPromptsAndResponsesLeft.jpeg" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1024" data-original-width="1024" height="320" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhrWLq5COuKTx1Fay6PV3oQ3UYu-mGtAaDZ1-LykK29bzlnC-FEtULaChnVdfmDHd6kcs-rr04VeNHF-AaE0xe9k1KZ_f4ADm_bQN_n4QUPGsbLwFAH1cTqoccECPkgYsiBmnAzsSilNSBaS3wHTucRfbgBAnxL1m77_Efz37NOA9KokuG2r83PBUsOMK0/s320/testingDALL-EMyPromptsAndResponsesLeft.jpeg" width="320" /></a></div><div class="separator" style="clear: both; font-weight: bold; text-align: center;"><br /></div><span style="font-family: verdana;">As the reader can see, ChatGPT returned two images:</span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">All of the men in both pictures are white ... so ChatGPT is not "woke" ... or is it?</span></li><li><span style="font-family: verdana;">There are eight men in the top picture, seven in the bottom ... despite reminders that there should only be five men.</span></li><li><span style="font-family: verdana;">Two men in the bottom picture are wearing sneakers/trainers, instead of sandals</span></li><li><span style="font-family: verdana;">Three men are standing in both pictures, whereas the prompts requested that all be seated</span></li><li><span style="font-family: verdana;">And the top picture is definitely "WOKE!!!" because it shows diversity in the men's ages. Two of the men are at least 40 -- middle aged men with faces that are "youthful", but not young, with white hair and white beards -- whereas they should all be young men in their mid-twenties</span> </li></ul></div><div style="text-align: left;"><div style="background-color: white;"><div style="color: black;"><div id="Overview"><span style="font-family: verdana;">So, ChatGPT and DALL-E did better than Gemini, but they did not meet the prompts' specifications. </span></div><div id="Overview"><span style="font-family: verdana;"><br /></span></div><div id="Overview" style="font-size: medium;"><span style="font-family: verdana;"><br /></span></div><div id="Overview" style="font-size: medium;"><span><span style="font-family: verdana;"><b style="font-size: large;">B. </b></span></span><span><span style="font-family: verdana;"><b style="font-size: large;">Top 3 stories in past week ...</b></span></span> </div><div id="Overview" style="font-size: medium;"><ol><li><span style="font-family: verdana;"><span style="font-size: x-small;"><span style="font-weight: 700; text-align: center;"><a href="#NonLLMNews"><span><span><span style="color: #cc0000;">SLM News</span></span></span></a></span><span style="font-weight: 700; text-align: center;"><span style="color: #cc0000;"> </span></span></span><br /><span style="font-size: medium;">"Google goes “open AI”" with Gemma, a free, open-weights chatbot family", Benj Edwards, </span><b style="font-size: large;"><i><a href="https://arstechnica.com/information-technology/2024/02/google-goes-open-ai-with-gemma-a-free-open-weights-chatbot-family/"><span style="color: #cc0000;">Ars Technica</span></a></i></b><span style="font-size: medium;">, 2/21/24 </span><b style="font-size: large;"><span style="color: #cc0000;">***</span></b><span style="font-size: medium;"> </span><br /><span style="font-size: medium;">-- This story also covered by </span><b style="font-size: large;"><i><a href="https://www.reuters.com/technology/google-releases-open-ai-models-after-meta-2024-02-21/"><span style="color: #cc0000;">Reuters</span></a></i></b><span style="font-size: medium;">, </span><b style="font-size: large;"><i><a href="https://www.google.com/search?q=Google+Gemma&sca_esv=1e0a15d19e51fa54&rlz=1C5CHFA_enUS1026US1026&sxsrf=ACQVn09v-fZZe6j4oqfVeFi5A8fZeALFMA%3A1708640295053&ei=J8jXZZfjAuat5NoP3MaOiA8&ved=0ahUKEwjXkteO_b-EAxXmFlkFHVyjA_EQ4dUDCBA&uact=5&oq=Google+Gemma&gs_lp=Egxnd3Mtd2l6LXNlcnAiDEdvb2dsZSBHZW1tYTIQEAAYgAQYChixAxiDARixAzINEAAYgAQYChixAxiDATIGEAAYAxgKMhAQABiABBiKBRgKGLEDGIMBMgYQABgDGAoyDRAAGIAEGAoYsQMYgwEyEBAAGIAEGIoFGAoYsQMYgwEyBhAAGAMYCjIQEAAYgAQYigUYChixAxiDATINEAAYgAQYChixAxiDAUj0mQFQ0pABWNKQAXACeAGQAQCYAVSgAVSqAQExuAEDyAEA-AEBwgIKEAAYRxjWBBiwA8ICDRAAGIAEGIoFGEMYsAOIBgGQBgk&sclient=gws-wiz-serp#:~:text=Fast%20Company,1%20day%20ago"><span style="color: #cc0000;">Fast Company</span></a></i></b><span style="font-size: medium;">, </span><b style="font-size: large;"><i><a href="https://huggingface.co/blog/gemma"><span style="color: #cc0000;">Hugging Face</span></a></i></b><span style="font-size: medium;">, <b><i><a href="https://fortune.com/2024/02/21/google-new-family-open-source-ai-models-gemma/"><span style="color: #cc0000;">Fortune</span></a>, <a href="https://www.nytimes.com/2024/02/21/technology/google-open-source-ai.html?smid=nytcore-android-share"><span style="color: #cc0000;">NY Times</span></a></i></b> ... and </span><a href="https://blog.google/technology/developers/gemma-open-models/" style="font-size: large; font-style: italic; font-weight: bold;"><span style="color: #cc0000;">Google</span></a><br /><br /></span></li><li><span style="font-size: medium;"> <a href="#Microsoft" style="font-family: verdana; font-size: small; text-align: center;"><b><span style="color: #cc0000;"><span>M</span>icrosoft</span></b></a><br /></span><span style="font-family: verdana; font-size: medium;">"Microsoft and Intel strike a custom chip deal that could be worth billions", Wes Davis, <b><i><a href="https://t.co/khNH6PHP5C"><span style="color: #cc0000;">The Verge</span></a></i></b>, 2/23/24 <b><span style="color: #cc0000;">***</span></b> <br />-- This story also covered by <b><i><a href="https://t.co/tyc3yWt4JA"><span style="color: #cc0000;">Reuters</span></a></i></b>, <b><i><a href="https://www.bloomberg.com/news/articles/2024-02-21/microsoft-will-use-intel-to-manufacture-home-grown-processor"><span style="color: #cc0000;">Bloomberg</span></a></i></b>, <b><i><a href="https://www.wsj.com/livecoverage/stock-market-today-dow-jones-earnings-02-21-2024/card/intel-unveils-microsoft-as-customer-gives-new-detail-about-contract-chip-making-arm-ZPojb7KKgnfkhk4KNIBa"><span style="color: #cc0000;">Wall Street Journal</span></a></i></b>, <b><i><a href="https://www.engadget.com/microsofts-upcoming-custom-chip-will-be-made-by-intel-063323035.html"><span style="color: #cc0000;">Engadget</span></a></i></b>,<br /><br /></span></li><li><span style="font-family: verdana; font-size: medium;"><a href="#Google" style="font-size: small; font-weight: 700; text-align: center;"><span style="color: #cc0000;">Google</span></a><br />"Google suspends Gemini from making AI images of people after a backlash complaining it was 'woke'", Joshua Zitser, <b style="font-size: medium;"><i><a href="https://www.businessinsider.com/google-gemini-ai-pause-image-generation-people-woke-complaints-2024-2"><span style="color: #cc0000;">Business Insider</span></a></i></b><span style="font-size: medium;">, </span><span style="font-size: medium;">2/22/24 </span><b style="font-size: medium;"><span style="color: #cc0000;">***</span></b><span style="font-size: medium;"> </span><br style="font-size: medium;" /><span style="font-size: medium;">-- This story also covered by </span><b style="font-size: medium;"><i><a href="https://t.co/JxG038vCtS"><span style="color: #cc0000;">The Verge</span></a>, </i></b><i style="font-size: medium; font-weight: bold;"><a href="https://t.co/yUAPEz2gsR"><span style="color: #cc0000;">Gizmodo</span></a>, <a href="https://www.wsj.com/tech/ai/google-restricts-ai-images-amid-outcry-over-chatbots-treatment-of-race-34954fb6"><span style="color: #cc0000;">Wall Street Journal</span></a></i><span style="font-size: medium;">, </span><b style="font-size: medium;"><i><a href="https://www.fastcompany.com/91034892/google-suspends-gemini-ai-image-generation-inaccuracies"><span style="color: #cc0000;">Fast Company</span></a></i></b><span style="font-size: medium;">, </span><b style="font-size: medium;"><i><a href="https://techcrunch.com/2024/02/22/google-gemini-image-pause-people/"><span style="color: #cc0000;">TechCrunch</span></a></i></b><span style="font-size: medium;">,</span><b style="font-size: medium;"><i><a href="https://apnews.com/article/google-gemini-ai-chatbot-image-generation-1bd45f1e67dfe0f88e5419a6efe3e06f"><span style="color: #cc0000;"> Associated Press</span></a></i></b><span style="font-size: medium;">, </span><b style="font-size: medium;"><i><a href="https://www.nytimes.com/2024/02/22/technology/google-gemini-german-uniforms.html?smid=nytcore-android-share"><span style="color: #cc0000;">NY Times</span></a></i></b></span><b style="font-family: verdana; font-style: italic;"> </b><span style="font-family: verdana;">... <span style="font-size: medium;">and </span></span><span style="font-size: medium;"><b style="font-family: verdana; font-style: italic;"><a href="https://twitter.com/Google_Comms/status/1760354549481546035/photo/1"><span style="color: #cc0000;">Google</span></a> </b><span style="font-family: verdana; font-style: italic;">(on X/Twitter)</span></span></li></ol></div></div><div><div style="font-size: medium;"><span style="color: #1a222a; font-family: verdana; font-size: 13.86px;"><span style="font-size: medium;"><b><br />C. </b></span></span><b style="font-size: 18px;"><i><a href="https://neoskeptics.blogspot.com/2023/12/basic-ai-faqs.html"><span style="color: #cc0000; font-family: verdana;">Dozen Basic AI FAQs</span></a></i></b></div><div style="color: #1a222a; font-size: medium;"><span style="font-family: verdana;"><span style="font-size: medium;">This page contains links to responses by Google's Bard chatbot running Gemini Pro to 12 questions that should be asked more frequently, but aren't. As consequence, too many readily understood AI terms have become meaningless buzzwords in the media.</span></span></div><div style="color: #1a222a; font-size: medium;"><span style="font-family: verdana;"><span style="font-size: medium;"><br /></span></span></div><div style="color: #1a222a; font-size: medium;"><span style="font-family: verdana;"><span style="font-size: medium;"><br /></span></span></div><div style="color: #1a222a;"><span style="font-family: verdana;"><span><b style="font-size: medium;">D.</b> <b style="font-size: medium;">A few of DeepMind's scientific achievements using machine learning</b><br /><i>Note: Publication dates usually lag far behind the end dates of research</i></span></span></div><div style="font-size: medium;"><span style="font-family: verdana;"><span style="font-size: medium;"><ul style="font-family: Times; font-size: medium;"><li><span style="font-family: verdana;"><span style="color: black;"><span style="color: #1a222a;">"</span></span>One of the Biggest Problems in Biology Has Finally Been Solved", Tanya Lewis, <span style="color: #cc0000;"><b><i><a href="https://www.scientificamerican.com/article/one-of-the-biggest-problems-in-biology-has-finally-been-solved/">Scientific American</a></i></b>,</span> 10/31/22 ... protein foldimg problem ... <i style="color: black;">DeepMind</i><br />-- This story also covered by <b style="color: black;"><i><a href="https://venturebeat.com/ai/a-year-ago-deepminds-alphafold-ai-changed-the-shape-of-science-but-there-is-more-work-to-do/">VentureBeat</a></i></b>, <b style="color: black;"><i><a href="https://www.cnet.com/science/biology/googles-deepmind-ai-predicts-3d-structure-of-nearly-every-protein-known-to-science/">CNET</a></i></b>, <b style="color: black;"><i><a href="https://www.science.org/content/article/google-s-deepmind-aces-protein-folding">Science</a></i></b>, <br /><br /></span></li><li style="color: black;"><span style="font-family: verdana;"><span style="color: #1a222a;">"Courtesy of AI: Weather forecasts for the hour, the week and the century", Devin Coldewey, <b><i><a href="https://t.co/7szMYJTtMu" style="color: #336699; text-decoration-line: none;">TechCrunch</a></i></b>, 11/14/23 ... <i style="color: black;">DeepMind</i><br />-- This story also covered by <b><i><a href="https://t.co/ZVUWEP6aUh" style="color: #336699; text-decoration-line: none;">Gizmodo</a></i></b>, <b><i><a href="https://www.ft.com/content/ca5d655f-d684-4dec-8daa-1c58b0674be1" style="color: #336699; text-decoration-line: none;">Financial Times</a></i></b>, <b><i><a href="https://www.technologyreview.com/2023/11/14/1083366/google-deepminds-weather-ai-can-forecast-extreme-weather-quicker-and-more-accurately/" style="color: #336699; text-decoration-line: none;">MIT Tech Review</a></i></b>, <b><i><a href="https://arstechnica.com/science/2023/11/ai-outperforms-conventional-weather-forecasting-for-the-first-time-google-study/" style="color: #336699; text-decoration-line: none;">Ars Technica</a></i></b>, <b><i><a href="https://www.wired.com/story/google-deepmind-ai-weather-forecast/" style="color: #336699; text-decoration-line: none;">Wired</a></i></b>, <b><i><a href="https://t.co/2P7gFHA3Ut" style="color: #336699; text-decoration-line: none;">ZDNet</a></i></b>, <br />-- These articles describe research published in </span><b style="color: #1a222a;"><i><a href="https://www.nature.com/articles/d41586-023-03552-y" style="color: #336699; text-decoration-line: none;">Nature</a><b> </b><span style="font-style: normal; font-weight: 400;">and</span><b> <a href="https://www.science.org/doi/10.1126/science.adi2336" style="color: #336699; text-decoration-line: none;">Science</a><br /><br /></b></i></b></span></li><li><span style="font-family: verdana;"><span><span><span><span>"AI Tool Pinpoints Genetic Mutations That Cause Disease", Even Callaway & Nature Magazine, <a href="https://www.scientificamerican.com/article/ai-tool-pinpoints-genetic-mutations-that-cause-disease/"><b><i><span style="color: black;">Scientific American</span></i></b></a></span><b style="color: black;"><i><a href="https://www.scientificamerican.com/article/ai-tool-pinpoints-genetic-mutations-that-cause-disease/"><span style="color: #cc0000;">,</span></a></i></b><span style="color: #1a222a;"> 11/21/23 ... <i style="color: black;">DeepMind</i><br /></span>-- This story also covered by<span style="color: black;"> </span><span style="color: black;"><a href="https://www.wsj.com/podcasts/tech-news-briefing/how-googles-deepmind-is-using-ai-to-uncover-the-causes-of-diseases/a0a6e750-b1f0-4265-a675-d00129059e7a#:~:text=Trying%20to%20determine%20who%20might,mutations%20that%20could%20cause%20illness" style="font-style: italic; font-weight: bold;"><span>Wall Street Journal</span></a><b style="font-style: italic;">, </b><a href="https://www.wired.com/story/deepmind-ai-alphamissense-genetics-rare-diseases/"><b><i>Wired</i></b></a><b style="font-style: italic;">, </b></span><span style="color: black;"><br /><br /></span></span></span></span></span></li><li style="color: black;"><span style="font-family: verdana;"><span><span style="color: #1a222a;"><span>"Google’s DeepMind finds 2.2M crystal structures in materials science win", Michael Peel, </span><b><i><a href="https://arstechnica.com/ai/2023/11/googles-deepmind-finds-2-2m-crystal-structures-in-materials-science-win/"><span>Ars Technica</span></a></i></b><span>, 11/29/23 ... "The trove of theoretically stable but experimentally unrealized combinations identified using an AI tool known as GNoME is more than 45 times larger than the number of such substances unearthed in the history of science, according to a paper published in Nature on Wednesday."</span><br /><span>-- This story also covered by </span><b><i><a href="https://www.wired.com/story/an-ai-dreamed-up-380000-new-materials-the-next-challenge-is-making-them/"><span>Wired</span></a></i></b><span>, </span><b><i><a href="https://venturebeat.com/ai/google-deepminds-materials-ai-has-already-discovered-2-2-million-new-crystals/"><span>VentureBeat</span></a></i></b><span>, </span><b><i><a href="https://www.nature.com/articles/d41586-023-03745-5"><span>Nature</span></a></i></b><span>, </span><b><i><a href="https://www.science.org/content/article/materials-predicting-ai-deepmind-could-revolutionize-electronics-batteries-and-solar"><span>Science</span></a></i></b><span><span>, </span><b><i><a href="https://www.technologyreview.com/2023/11/29/1084061/deepmind-ai-tool-for-new-materials-discovery"><span>MIT Tech Review</span></a></i></b><span>,<br /></span></span></span></span></span></li></ul></span></span></div></div></div></div>
Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-57644061012426996242024-02-19T21:14:00.019-05:002024-03-27T21:37:10.372-04:00Google's new Gemini 1.5 ... plus OpenAI's disruptions ... TL;DR and podcast 18Feb24<div style="text-align: right;"><span style="font-family: verdana; font-size: x-small;"><i>Last update: Monday 2/19/24 </i></span></div><div style="text-align: left;"><span style="font-family: verdana;"><div class="separator" style="clear: both; text-align: center;"><div style="text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><span style="font-size: medium;"><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgnZskCH0r8cNUlWjjyaYu6tt14L8v5Pj8bBdp997eszz1gA_5T2yNczRyIv2KPEp8d-dtoEzM3ht65GvYoFGpR34ecXjtSG2eNZxmebFaG8EJWVlngrcywHBhisju6blwruL-98fpWPrY_bj6XoTWghyeo6cyw5vCt4xCv0DJAsHlCwE8MAL-RsvVekGo/s750/WoolyMamoths.jpg" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="422" data-original-width="750" height="180" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgnZskCH0r8cNUlWjjyaYu6tt14L8v5Pj8bBdp997eszz1gA_5T2yNczRyIv2KPEp8d-dtoEzM3ht65GvYoFGpR34ecXjtSG2eNZxmebFaG8EJWVlngrcywHBhisju6blwruL-98fpWPrY_bj6XoTWghyeo6cyw5vCt4xCv0DJAsHlCwE8MAL-RsvVekGo/s320/WoolyMamoths.jpg" vspace="10" width="320" /></a></div>Welcome to our </span></span></span></span><span style="font-size: medium;"><span><span style="background-color: white; color: #1a222a;">18Feb24</span><span style="background-color: white; color: #1a222a;"> </span><span style="background-color: white; color: #1a222a;">TL;DR summary + podcast about the past week's top AI stories on our "</span><b><i><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html"><span style="color: #cc0000;">Useful AI News</span></a></i></b><span style="background-color: white; color: #1a222a;">" page </span></span><b style="background-color: white; color: #1a222a;">➡</b><span style="background-color: white; color: #1a222a;"> 1) Google's new Gemini 1.5, (2) ChatGPT's personal memory, (3) OpenAI's Sora video generator, and 4) OpenAI's Internet search </span></span></div><div style="text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><br /></span></span></span></div><div style="text-align: left;"><br /></div><div style="text-align: left;"><div class="separator" style="clear: both; font-size: 13.86px;"><div style="text-align: right;"><div style="color: #1a222a;"><span style="font-size: medium;">Click link for podcast (opens in new tab)</span> </div><div><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-size: large; text-align: left; text-decoration-line: none;"><span style="color: #1a222a; font-size: medium;">➡</span></a><span style="color: #1a222a;"> </span><a href="https://drive.google.com/file/d/1Pd9rUdx5X6OYK4zPu2cdlP5hMYAI2uIb/view?usp=sharing" target="_blank">
<span style="color: #cc0000;"><i><span style="font-size: medium;"><b>Google + OpenAI + </b></span></i><i><span style="font-size: medium;"><b>OpenAI</b></span></i><i><span style="font-size: medium;"><b> + </b></span></i><i><span style="font-size: medium;"><b>OpenAI</b></span></i></span></a><i style="color: #1a222a;"><span style="font-size: medium;"><b> </b></span></i></div></div></div><div class="separator" style="clear: both;"><div style="text-align: right;"><span style="color: black; font-size: 13.86px;"><span style="color: #1a222a;">Click the "start" button the when the podcast page is loaded<br />... </span></span><span style="color: #1a222a; font-size: 13.86px;"> If audio fails to start, or gets stuck, try reloading that page</span><br /></div></div></div><div style="background-color: white; text-align: right;"><div class="separator" style="clear: both; text-align: left;"><div style="color: #1a222a; font-family: Verdana, Geneva, sans-serif;"><span style="font-family: verdana;"><span face="Verdana, Geneva, sans-serif" style="font-size: medium;">TL;DR link<b> </b></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-size: large; text-decoration-line: none;"><span style="color: #1a222a;">➡</span></a><b style="color: black; font-family: Verdana, Geneva, sans-serif;"><span style="font-size: medium;"><span style="color: #1a222a;"> </span><i><a href="https://neoskeptics.blogspot.com/2024/02/googles-new-gemini-15-plus-openais.html#more"><span style="color: #cc0000;">HERE</span></a></i></span></b></span></div></div></div></div></span></div><span><span style="font-family: verdana;"> <a name='more'></a></span></span><div style="text-align: left;"><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;">A. </b><b style="background-color: white; font-family: verdana; font-size: large;">TL;DR ... </b><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;">Top 4 stories in past week ...</b></div><div style="text-align: left;"><span style="background-color: white; font-family: verdana;"><span style="color: #1a222a;">Most readers of this blog know that Google has been striving mightily to catch up to the Microsoft/OpenAi partnership ever since the partners released their GPT-4 language model in March 2023. Google made game changing progress in </span><b><i><a href="First came the AI Big Bang; now comes the aftershocks and tremors... TL;DR and podcast 10Dec23"><span style="color: #cc0000;">December 2023</span></a></i></b><span style="color: #1a222a;"> when it released its new Gemini family of language models. It made additional progress in early </span><b><i><a href="https://neoskeptics.blogspot.com/2024/02/big-techs-investor-conference-calls.html"><span style="color: #cc0000;">February 2024</span></a></i></b><i style="font-weight: bold;"><span style="color: #cc0000;"> </span></i><span style="color: #1a222a;">by offering subscriptions to users that provided them with access to Gemini enhancements to the applications in Google's Workspace, enhancements that were </span></span><span style="background-color: white; color: #1a222a; font-family: verdana;">akin to Microsoft's Copilots</span></div><div style="text-align: left;"><span style="background-color: white; color: #1a222a; font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="background-color: white; font-family: verdana;"><span style="color: #1a222a;">However, Google surprised the Ai community this week by releasing Gemini 1.5, a powerful upgrade to its initial model. Unfortunately for Google, this substantial effort to close the remaining gap with the Microsoft/OpenAI partnership was offset by Open/AI's simultaneous announcements of comparably powerful upgrades to GPT-4. Indeed, these disruptive enhancements are the most significant upgrades to the partners' language model since they moved from GPT-3.5 to GPT-4.</span></span></div><div style="text-align: left;"><b><span style="font-family: verdana; font-size: medium;"><br /></span></b></div><div style="text-align: left;"><b><span style="font-family: verdana; font-size: medium;">1) </span></b><span style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;"> </span><span style="background-color: white; color: #1a222a; font-family: verdana; font-size: medium;"><b>Google's new Gemini 1.5</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">According to <b><i><a href="https://blog.google/technology/ai/google-gemini-next-generation-model-february-2024/"><span style="color: #cc0000;">Demis Hassabis, CEO of Google DeepMind</span></a></i></b>, </span></div><div style="text-align: left;"><blockquote><span style="font-family: verdana;">"</span><span style="font-family: verdana;">The first Gemini 1.5 model we’re releasing for early testing is Gemini 1.5 Pro. It’s a mid-size multimodal model, optimized for scaling across a wide-range of tasks, and performs at a similar level to 1.0 Ultra, our largest model to date. It also introduces a breakthrough experimental feature in long-context understanding.</span></blockquote><blockquote><span style="font-family: verdana;">Gemini 1.5 Pro comes with a standard 128,000 token context window. But starting today, a limited group of developers and enterprise customers can try it with a context window of up to 1 million tokens via AI Studio and Vertex AI in private preview </span></blockquote><blockquote><span style="font-family: verdana;">... 1.5 Pro can seamlessly analyze, classify and summarize large amounts of content within a given prompt. </span><span style="font-family: verdana;">For example, when given the 402-page transcripts from Apollo 11’s mission to the moon, it can reason about conversations, events and details found across the document </span></blockquote><blockquote><span style="font-family: verdana;"><span>... </span><span face="Roboto, sans-serif" style="background-color: white; color: #202124;">1.5 Pro can perform more relevant problem-solving tasks across longer blocks of code. When given a prompt with more than 100,000 lines of code, it can better reason across examples, suggest helpful modifications and give explanations about how different parts of the code works...</span> </span></blockquote><blockquote><span style="font-family: verdana;"><span face="Roboto, sans-serif" style="background-color: white; color: #202124;">... </span><span face="Roboto, sans-serif" style="color: #202124;">1.5 Pro can perform highly-sophisticated understanding and reasoning tasks for different modalities, including video. For instance, when given a 44-minute silent Buster Keaton movie, the model can accurately analyze various plot points and events, and even reason about small details in the movie that could easily be missed.</span><span>"</span></span></blockquote></div><div style="text-align: left;"><span style="font-family: verdana;">Impressive, very impressive.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><b><span style="font-family: verdana; font-size: medium;">2) </span></b><span style="font-size: medium;"><span style="color: #1a222a; font-family: verdana; text-align: right;"><span><b>ChatGPT's </b></span></span><span style="background-color: white; color: #1a222a; font-family: verdana;"><b>personal memory</b></span></span></div><div style="text-align: left;"><span style="font-family: verdana;">According to <b><i><a href="https://openai.com/blog/memory-and-new-controls-for-chatgpt"><span style="color: #cc0000;">OpenAI</span></a></i></b>:</span></div><blockquote><div style="text-align: left;"><span style="font-family: verdana;">"We’re testing memory with ChatGPT. Remembering things you discuss across all chats saves you from having to repeat information and makes future conversations more helpful ... </span> </div></blockquote><blockquote><span style="font-family: verdana;">... As you chat with ChatGPT, you can ask it to remember something specific or let it pick up details itself. ChatGPT’s memory will get better the more you use it and you'll start to notice the improvements over time ... </span> </blockquote><blockquote><div style="text-align: left;"><span style="font-family: verdana;">... </span><span style="font-family: verdana;">You can turn off memory at any time (Settings > Personalization > Memory). While memory is off, you won't create or use memories.</span><span style="font-family: verdana;">"</span></div></blockquote><div style="text-align: left;"><span style="font-family: verdana;">When the editor of this blog read the first report in tech media about this enhancement (see our fourth top event this week, below), he dismissed it as "interesting", "useful", and possibly "intrusive" ... then he encountered reports that OpenAi was developing a "search" function that would challenge Google's dominance in search and it hit him: Wow!!! Telling ChatGPT what you like or dislike, or merely letting ChatGPT infer your preferences from your queries might provide an OpenAI search engine with a substantial advantage over Google's search engine with regards to issues or items that were important enough for you to ask ChatGPT to explain them to you.<br /><br /></span></div><div style="text-align: left;"><b><span style="font-family: verdana; font-size: medium;">3) </span></b><b style="color: #1a222a; font-family: verdana; font-size: large; text-align: right;">OpenAI's Sora video generator</b></div><div style="text-align: left;"><span style="font-family: verdana;"><span style="text-align: right;"><span style="color: #1a222a;">Here's a quote from </span><b><i><a href="https://www.blogger.com/blog/post/edit/8043580455676201364/5764406101242699624"><span style="color: #cc0000;">OpenAI's introduction</span></a></i></b><span style="color: #1a222a;"> to Sora:</span></span><b style="color: #1a222a; text-align: right;"> </b></span></div><div style="text-align: left;"><blockquote><span style="color: #1a222a; text-align: right;"><span style="font-family: verdana;">"Introducing Sora, our text-to-video model. Sora can generate videos up to a minute long while maintaining visual quality and adherence to the user’s prompt."</span></span></blockquote><p><span style="font-family: verdana;">Given the many absurdly inaccurate still images that DALL-E has produced, the photo-realism of Sora's videos is astounding. Here are links to a few examples of the prompts and resulting videos:</span></p></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="text-align: right;"><span style="font-family: verdana;"><a href="https://media.wired.com/clips/65cd6097640589f91cb00713/360p/pass/monster.mp4" style="font-style: italic; font-weight: bold;"><span style="color: #cc0000;">a short fluffy monster</span></a><b style="font-style: italic;"> </b>... Wired<br />Prompt = "animated scene features a close-up of a short fluffy monster kneeling beside a melting red candle. the art style is 3d and realistic, with a focus on lighting and texture. the mood of the painting is one of wonder and curiosity, as the monster gazes at the flame with wide eyes and open mouth. its pose and expression convey a sense of innocence and playfulness, as if it is exploring the world around it for the first time. the use of warm colors and dramatic lighting further enhances the cozy atmosphere of the image.”<br /><br /></span></span></li><li><span style="text-align: right;"><span style="font-family: verdana;"><a href="https://media.wired.com/clips/65cd609bb4d2e54e7c66f6b7/360p/pass/mammoth.mp4" style="font-style: italic; font-weight: bold;"><span style="color: #cc0000;">giant wooly mammoths</span></a><b style="font-style: italic;"> </b>... Wired<br />Prompt = "several giant wooly mammoths approach treading through a snowy meadow, their long wooly fur lightly blows in the wind as they walk, snow covered trees and dramatic snow capped mountains in the distance, mid afternoon light with wispy clouds and a sun high in the distance creates a warm glow, the low camera view is stunning capturing the large furry mammal with beautiful photography, depth of field.”<br /><br /></span></span></li><li><b style="text-align: right;"><i><a href="https://media.wired.com/clips/65cd609a1b47a15ce1b4001e/360p/pass/tokyo.mp4"><span style="color: #cc0000; font-family: verdana;">Tokyo with snowflakes and cherry blossoms </span></a></i></b><span style="text-align: right;"><span style="font-family: verdana;">... Wired<br /></span><span style="font-family: verdana;">Prompt = "Beautiful, snowy Tokyo city is bustling. The camera moves through the bustling city street, following several people enjoying the beautiful snowy weather and shopping at nearby stalls. Gorgeous sakura petals are flying through the wind along with snowflakes."</span></span></li></ul></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><b><span style="font-family: verdana; font-size: medium;">4) </span></b><b style="color: #1a222a; font-family: verdana; font-size: large; text-align: right;">OpenAI's Internet search</b></div><div style="text-align: left;"><span style="font-family: verdana;">According to an exclusive report in <b><i><a href="https://www.theinformation.com/articles/openai-develops-web-search-product-in-challenge-to-google?rc=v6kkoz&shared=8b3a659a73e4c6e3"><span style="color: #cc0000;">The Information:</span></a></i></b></span></div><div style="text-align: left;"><blockquote><span style="font-family: verdana;">"</span><span style="font-family: verdana;">OpenAI has been developing a web search product that would bring the Microsoft-backed startup into more direct competition with Google, according to someone with knowledge of OpenAI’s plans. The search service would be partly powered by Bing, this person said.</span></blockquote><blockquote><span style="font-family: verdana;">The move to launch a search app comes a year after Microsoft CEO Satya Nadella said his company would “make Google dance” by incorporating artificial intelligence from OpenAI into Microsoft’s Bing search engine. That partnership has failed to dent Google’s search dominance.</span><span style="font-family: verdana;">"</span></blockquote></div><div style="text-align: left;"><span style="font-family: verdana;">In other words, OpenAI is developing a search engine, but has not yet completed this project. Its serious commitment plus Microsoft's vast resources do not guarantee that this project will succeed; but it does guarantee that Google must maintain the "Code Red" high alert that it declared in March 2023 when it first perceived that the Microsoft/OpenAi partnership posed an existential threat.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><div style="background-color: white; font-size: 13.86px;"><div style="font-size: medium;"><div id="Overview" style="color: black;"><span><span style="font-family: verdana;"><b style="font-size: large;">B. </b></span></span><span><span style="font-family: verdana;"><b style="font-size: large;">Top 4 stories in past week ...</b></span></span> </div><div id="Overview"><ol><li><span style="font-family: verdana; font-size: medium;"><a href="#Google" style="color: black; font-size: small; font-weight: 700; text-align: center;"><span style="color: #cc0000;">Google</span></a><br />"Google unveils Gemini 1.5, a next-gen AI model with million-token context window", Michael Nuñez, <b style="color: black;"><i><a href="https://t.co/NSIv07tKnr"><span style="color: #cc0000;">VentureBeat</span></a></i></b>, 2/15/24 <b style="color: black;"><span style="color: #cc0000;">***</span></b><br />-- This story also covered by <b style="color: black;"><i><a href="https://t.co/XBnSoX91dV"><span style="color: #cc0000;">The Verge,</span></a></i></b> <b style="color: black;"><i><a href="https://techcrunch.com/2024/02/15/googles-new-gemini-model-can-analyze-an-hour-long-video-but-few-people-can-use-it/"><span style="color: #cc0000;">TechCrunch</span></a></i></b>, <b style="color: black;"><i><a href="https://t.co/BwL5ditedf"><span style="color: #cc0000;">Bloomberg</span></a></i></b>, <b style="color: black;"><i><a href="https://t.co/pgXGpQqDpm"><span style="color: #cc0000;">Wired</span></a></i></b>, <a href="https://t.co/UWtPBbWo3S" style="color: black; font-style: italic; font-weight: bold;"><span style="color: #cc0000;">Mashable</span></a><b style="color: black; font-style: italic;"> </b>... and <b style="font-style: italic;"><a href="https://blog.google/technology/ai/google-gemini-next-generation-model-february-2024/"><span style="color: #cc0000;">Google</span></a></b><br /></span></li></ol></div><div id="Overview" style="text-align: center;"><span style="font-family: verdana;"><span><span style="font-family: Times;"><ul style="text-align: left;"><li><span style="font-family: verdana; font-size: medium;"><b style="color: black; font-size: small; font-style: normal; text-align: center;"><a href="#OpenAI"><span style="color: #cc0000;">OpenAI</span></a></b><br /><span style="color: black; font-style: normal; font-weight: 400;">"ChatGPT is getting ‘memory’ to remember who you are and what you like" , David Pierce, </span><a href="https://t.co/6VbIlitbUG" style="color: black; font-style: normal; font-weight: 400;"><b><i><span style="color: #cc0000;">The Verge</span></i></b></a><span style="color: black; font-style: normal; font-weight: 400;">, 2/13/24 </span><b style="color: black; font-style: normal; font-weight: 400;"><span style="color: #cc0000;">***</span></b><span style="color: black; font-style: normal; font-weight: 400;"> </span><br /><span style="color: black; font-style: normal; font-weight: 400;">-- This story also covered by </span><b style="color: black; font-style: normal; font-weight: 400;"><i><a href="https://t.co/SQA0ckWvfs"><span style="color: #cc0000;">Bloomberg</span></a></i></b><span style="color: black; font-style: normal; font-weight: 400;">, </span><b style="color: black; font-style: normal; font-weight: 400;"><i><a href="https://t.co/84zBpBqNg5"><span style="color: #cc0000;">Wired</span></a></i></b><span style="color: black; font-style: normal; font-weight: 400;">, </span><a href="https://t.co/ZA5MYa0oE4" style="color: black; font-style: italic; font-weight: bold;"><span style="color: #cc0000;">NY Times</span></a><b style="color: black; font-style: italic; font-weight: 400;"> </b><b><i><a href="https://openai.com/blog/memory-and-new-controls-for-chatgpt"><span style="color: #cc0000;">... and OpenAI</span></a></i></b><br /><br /></span></li><li style="color: black; font-style: italic; font-weight: bold;"><span style="font-family: verdana; font-size: medium;"><b style="font-size: small; font-style: normal; text-align: center;"><a href="#OpenAI"><span style="color: #cc0000;">OpenAI</span></a></b><br /><span style="font-style: normal; font-weight: 400;">"OpenAI’s Sora Turns AI Prompts Into Photorealistic Videos", Steven Levy, </span><b style="font-style: normal;"><i><a href="https://t.co/v6KoRej9zT"><span style="color: #cc0000;">Wired</span></a></i></b><span style="font-style: normal; font-weight: 400;">, 2/13/24 </span><span style="font-style: normal; font-weight: 400;"><span style="color: #cc0000;">*** </span></span><br /><span style="font-style: normal; font-weight: 400;">-- This story also covered by</span><span style="color: #cc0000; font-style: normal; font-weight: 400;"> </span><b style="font-style: normal;"><i><a href="https://t.co/EC6YSbf8WF"><span style="color: #cc0000;">VentureBeat</span></a></i></b><span style="color: #cc0000; font-style: normal;">, </span><b style="font-style: normal;"><i><a href="https://t.co/TccL9vD99L"><span style="color: #cc0000;">The Verge</span></a></i></b><span style="color: #cc0000; font-style: normal;">, </span><b style="font-style: normal;"><i><a href="https://t.co/VZCh44yNSA"><span style="color: #cc0000;">Bloomberg</span></a></i></b><span style="color: #cc0000; font-style: normal; font-weight: 400;">, </span><span style="font-style: normal;"><a href="https://t.co/nmvuaplof9"><span style="color: #cc0000;">TechCrunch,</span></a> </span><b style="font-style: normal;"><i><a href="https://t.co/k84n9w2YdL"><span style="color: #cc0000;">Engadget</span></a></i></b><span style="font-style: normal;">, </span><b style="font-style: normal;"><i><a href="https://t.co/JQB1Jn8qCF"><span style="color: #cc0000;">Mashable</span></a></i></b><span style="font-style: normal; font-weight: 400;">, </span><a href="https://www.nytimes.com/2024/02/15/technology/openai-sora-videos.html?smid=nytcore-android-share"><span style="color: #cc0000;">NY Times</span></a><br /><br /></span></li><li style="color: black; font-style: normal; font-weight: 400;"><span style="font-family: verdana; font-size: medium;"><b style="font-size: small; text-align: center;"><a href="#OpenAI"><span style="color: #cc0000;">OpenAI</span></a></b><br />"OpenAI Develops Web Search Product in Challenge to Google", <b><i><a href="https://www.theinformation.com/articles/openai-develops-web-search-product-in-challenge-to-google?rc=v6kkoz&shared=8b3a659a73e4c6e3"><span style="color: #cc0000;">The Information</span></a></i></b>, 2/13/24</span><span style="font-family: verdana;"><span style="font-size: medium;"> ... Link downloads the full exclusive paywalled article for 3 free reads <b><span style="color: #cc0000;">*** </span></b><br />-- This story also covered by </span><b><i><span style="font-size: medium;"><a href="https://www.bloomberg.com/news/articles/2024-02-15/alphabet-drops-after-report-openai-is-developing-search-product?utm_source=twitter&utm_campaign=socialflow-organic&utm_content=tech&cmpid%3D=socialflow-twitter-tech&utm_medium=social"><span style="color: #cc0000;">Bloomberg</span></a>, <a href="https://t.co/VBgknBNlHZ"><span style="color: #cc0000;">Gizmodo</span></a>,</span></i></b></span></li></ul><i style="color: black;"><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#Top" style="font-weight: 400;"></a></i></span></span></span></div></div><div style="font-size: medium;"><div><span style="color: #1a222a; font-family: verdana; font-size: 13.86px;"><span style="font-size: medium;"><b><br /></b></span></span></div><div><span style="color: #1a222a; font-family: verdana; font-size: 13.86px;"><span style="font-size: medium;"><b>C. </b></span></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="background-color: transparent; color: #336699; font-family: verdana; font-size: large; text-decoration-line: none;"><span style="color: #1a222a; font-size: medium;">➡</span></a><span style="color: #1a222a;"> </span><b style="font-size: 18px;"><i><a href="https://neoskeptics.blogspot.com/2023/12/basic-ai-faqs.html"><span style="color: #cc0000; font-family: verdana;">Dozen Basic AI FAQs</span></a></i></b></div><div style="color: #1a222a;"><span style="font-family: verdana;"><span style="font-size: medium;">This page contains links to responses by Google's Bard chatbot running Gemini Pro to 12 questions that should be asked more frequently, but aren't. As consequence, too many readily understood AI terms have become meaningless buzzwords in the media.</span></span></div></div></div></div>
Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-33764330124661154052024-02-13T17:07:00.043-05:002024-03-27T21:37:32.707-04:00Google runs Microsoft's playbook very badly; well-intended public policies; and Altman's $trillions ...TL;DR and podcast 12Feb24<div style="text-align: right;"><span style="font-family: verdana; font-size: x-small;"><i>Last update: Monday 2/12/24 </i></span></div><div style="text-align: left;"><span><div class="separator" style="clear: both; text-align: center;"><div style="font-family: verdana; text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><span style="font-size: medium;"><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjs1Vrd0jn0sV0kMOsAYhFiXtkgoMrKBI42hNeTpXbH56BnUvxCcB7kP7SpjJbn_lEWfypnstTry6yhRuv0786_JswwL-I-oNsbHH6rVNEdZS1ecImM3gHCHptvNRxS-_6m3hEHbrkRzqIbQAskTUdyWAbezgnt4ZAYz5aE0VjBOuEKBvhjrCaXfkiwjtA/s1200/GoogleInCloud.jpg" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="800" data-original-width="1200" height="213" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjs1Vrd0jn0sV0kMOsAYhFiXtkgoMrKBI42hNeTpXbH56BnUvxCcB7kP7SpjJbn_lEWfypnstTry6yhRuv0786_JswwL-I-oNsbHH6rVNEdZS1ecImM3gHCHptvNRxS-_6m3hEHbrkRzqIbQAskTUdyWAbezgnt4ZAYz5aE0VjBOuEKBvhjrCaXfkiwjtA/s320/GoogleInCloud.jpg" vspace="10" width="320" /></a></div><div style="text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><span style="font-size: medium;"><br /></span></span></span></span></div><div style="text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><span style="font-size: medium;"><br /></span></span></span></span></div>Welcome to our 12Feb</span></span></span></span><span style="font-size: medium;"><span><span style="background-color: white; color: #1a222a;">24</span><span style="background-color: white; color: #1a222a;"> </span><span style="background-color: white; color: #1a222a;">TL;DR summary + podcast about the past week's top AI stories on our "</span><b><i><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html"><span style="color: #cc0000;">Useful AI News</span></a></i></b><span style="background-color: white; color: #1a222a;">" page </span></span><b style="background-color: white; color: #1a222a;">➡</b><span style="background-color: white; color: #1a222a;"> (1) US AI Consortium, (2) FCC, (3) Altman , and (4) Google</span></span></div><div style="font-family: verdana; text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><br /></span></span></span></div><div style="font-family: verdana; text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><br /></span></span></span></div><div style="font-family: verdana; text-align: left;"><br /></div><div style="font-family: verdana; text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><br /></span></span></span></div><div style="font-family: verdana; text-align: left;"><div class="separator" style="clear: both;"><div style="text-align: right;"><div style="color: #1a222a; font-size: 13.86px;"><span style="font-size: medium;">Click link for podcast (opens in new tab)</span> </div><div style="text-align: right;"><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-size: large; text-align: left; text-decoration-line: none;"><span style="color: #1a222a; font-size: medium;">➡</span></a><span><span><span style="color: #1a222a; font-size: 13.86px;"> <b style="background-color: white; color: black; font-family: Verdana, Geneva, sans-serif; font-size: large; text-align: left;"><a href="https://drive.google.com/file/d/1o1CUQudUjCHR-gnK5NVltkKEsMW2gG6D/view?usp=sharing" target="_blank"><span style="color: #cc0000;"><i>Consortium + FCC + Altman + Google</i></span></a><span style="color: #1a222a;"> </span></b><br /><span style="color: black; font-family: Times; font-size: medium; text-align: left;"><span style="font-family: verdana;"></span></span></span></span></span><span style="color: #1a222a; font-size: 13.86px;">Click the "start" button the when the podcast page is loaded</span></div></div></div><div class="separator" style="clear: both;"><div style="color: #1a222a; font-size: 13.86px; text-align: right;"><span style="color: black;"><span style="color: #1a222a;">... </span></span> If audio fails to start, or gets stuck, try reloading that page</div></div></div><div style="background-color: white; text-align: right;"><div class="separator" style="clear: both; text-align: left;"><div style="color: black; font-family: Times;"><span face="Verdana, Geneva, sans-serif" style="color: #1a222a; font-size: medium;">TL;DR link<b> </b></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-family: verdana; font-size: large; text-decoration-line: none;"><span style="color: #1a222a;">➡</span></a><b style="font-family: Verdana, Geneva, sans-serif;"><span style="font-size: medium;"><span style="color: #1a222a;"> </span><i><a href="https://neoskeptics.blogspot.com/2024/02/google-copies-microsofts-playbook-well.html#more"><span style="color: #cc0000;">HERE</span></a></i></span></b></div><div style="color: black; font-family: Times;"><b style="color: #1a222a; font-family: Verdana, Geneva, sans-serif;"><span><a name='more'></a></span></b></div></div></div></div></span></div><div style="text-align: left;"><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;">A. </b><b style="background-color: white; font-family: verdana; font-size: large;">TL;DR ... </b><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;">Top 4 stories in past week ...</b></div><div style="text-align: center;"><b style="background-color: white; font-family: verdana; font-size: large;"><i><span style="color: #cc0000;"><br /></span></i></b></div><div style="text-align: left;"><b><span style="font-family: verdana;">1) </span></b><span style="background-color: white; color: #1a222a; font-family: verdana;"><b>Google, Apple, Meta, Microsoft, Amazon, and everybody else join US consortium to advance responsible AI</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">The U.S. AI Safety Institute Consortium (AISIC) has been convened by the U.S. Department of Commerce and will be housed in its National Institute of Standards and Technology. According to the Dept. of Commerce:</span></div><div style="text-align: left;"><blockquote><span style="font-family: verdana;"><span>"</span><span face=""Source Sans Pro Web", "Helvetica Neue", Helvetica, Roboto, Arial, sans-serif" style="background-color: white; color: #1b1b1b; font-size: 16.8px;">The U.S. government has a significant role to play in setting the standards and developing the tools we need to mitigate the risks and harness the immense potential of artificial intelligence. President Biden directed us to pull every lever to accomplish two key goals: set safety standards and protect our innovation ecosystem. That’s precisely what the U.S. AI Safety Institute Consortium is set up to help us do,” </span><span face=""Source Sans Pro Web", "Helvetica Neue", Helvetica, Roboto, Arial, sans-serif" style="background-color: white; box-sizing: inherit; color: #1b1b1b; font-size: 16.8px;">said Gina Raimondo,</span><span face=""Source Sans Pro Web", "Helvetica Neue", Helvetica, Roboto, Arial, sans-serif" style="background-color: white; color: #1b1b1b; font-size: 16.8px;"> Secretary,U.S. Dept. of Commerce</span></span></blockquote></div><div style="text-align: left;"><span style="font-family: verdana;">High sounding words, but so generic/nonspecific that we won't know what the consortium will do until it actually does something. </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><b><span style="font-family: verdana;">2) The Federal Communications Comission</span></b><span style="background-color: white; color: #1a222a; font-family: verdana;"><b> banned A.I.-Generated Robocalls</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">The Federal Communications Commission announced the unanimous adoption of a Declaratory Ruling that recognizes calls made with AI-generated voices are “artificial” under the Telephone Consumer Protection Act (TCPA). According to the FCC:</span></div><blockquote><div style="text-align: left;"><span style="font-family: verdana;">"The TCPA gives the FCC civil enforcement authority to fine robocallers. The Commission can also take steps to block calls from telephone carriers facilitating illegal robocalls. In addition, the TCPA allows individual consumers or an organization to bring a lawsuit against robocallers in court. Lastly, State Attorneys General have their own enforcement tools which may be tied to robocall definitions under the TCPA. "</span></div></blockquote><div style="text-align: left;"><span style="font-family: verdana;">More high sounding words, but everybody knows that this declaration was driven by concern that the coming elections might be flooded with AI generated misinformation calls. However, we won't know if the fines will be steep enough to discourage robo-callers and/or if the robo-calls can be blocked fast enough until we are much deeper into the election season.</span></div><div style="text-align: left;"><br /></div><div style="text-align: left;"><b><span style="font-family: verdana; font-size: medium;">3) </span></b><span style="font-family: verdana; font-size: medium;"><b>Sam Altman Seeks Trillions (or was that bazillions???) of Dollars to Reshape Business of Chips and AI</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">Mr. Altman acts as though he still believes that bigger language models are always better language models. Bigger models require more graphics processing units (GPUs), the kinds of expensive chips made by Nvidia, chips that will be in increasingly short supply and so become even more costly. </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">But wait!!! In a previous TL;DR we suggested that Mr. Altman is first and foremost a salesman, a salesman with no academic training in AI and no hands-on work experience as an AI developer. </span><span style="font-family: verdana;">As such, his pronouncements about the future of AI technology should not be taken seriously. But as a super salesman,</span><span style="font-family: verdana;"> he is always making a sales pitch to someone. In this case he is probably pitching to the leaders of oil rich countries in the Middle East who are in a desperate race to convert their oil riches into non-fossil riches, while they still can.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Notice that we did not say that Mr. Altman really believes that bigger language models are always better. As the CEO of OpenAI, Microsoft's partner in developing large models, he is surely aware that Microsoft's CEO, Satya Nadella, on a recent widely attended conference call about Microsoft's earnings in the last quarter, loudly proclaimed Microsoft's discovery that small models have surprising power. Indeed, Nadella noted that Microsoft has formed a team of its own best AI experts to develop powerful small models that will be cheaper to build and to operate than the large models that had been developed by OpenAI. This team reports directly to Keven Scott, Microsoft's Chief Technology Officer, the same CTO who managed Microsoft's successful partnership with OpenAI. Readers who are unaware of Scott's crucial role as the top level manager of the successful Microsoft/OpenAI partnership are referred to:</span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">"The Inside Story of Microsoft’s Partnership with OpenAI", Charles Duhigg, <b><i><a href="https://www.newyorker.com/magazine/2023/12/11/the-inside-story-of-microsofts-partnership-with-openai"><span style="color: #cc0000;">New Yorker</span></a></i></b>, 12/1/23</span></li></ul></div><div style="text-align: left;"><b><span style="font-family: verdana; font-size: medium;">4) </span></b><b><span style="font-family: verdana; font-size: medium;">Google runs Microsoft's playbook ... very badly</span></b></div><div style="text-align: left;"><span style="font-family: verdana;">Our last story is a disturbing reminder of our tech media's capacity to misinform us unintentionally by misdirection. The essence of the story was simple: Google changed the name of its chatbot from "Bard" to "Gemini"; it introduced a monthly subscription plan that enables users to gain access to its largest language model (LLM); it changed the name of its largest LLM from "Ultra" to "Advanced"; and it integrated Gemini into its workspace, i.e., it enabled Gemini to help subscribers make more effective use of gmail, Google Docs, Google Sheets, etc -- its email, word processor, spread sheet applications, etc</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">What was misleading in each of the six news articles cited in our headline was their comparison of Google's offerings to OpenAI's offerings. Evidently that was what Google had done in its press briefing. No one should be surprised to read that Google claimed that its new posture was now as strong as, if not stronger than OpenAi's posture. And no one should have reason to double this claim.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">The editor now begs his readers' indulgence while he invokes a comparison to the Super Bowl that he watched this weekend ... along with over 120 million other viewers. Imagine that Google was a professional football team and had made the following claims to a neutral bystander in the state of Nevada in an alternative universe.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><b>Google</b>: We are a college football team at the end of the first half of our championship game in a league of community colleges. We're feeling great as we head to the showers, with a solid lead of 14 to 6.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><b>Bystander</b>: No, no, you're not. You are really a professional football team in the NFL; you are playing the Kansas City Chiefs in the Super Bowl; and they are crushing you at the first half with a score of 49 to 3. </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Does anyone in the real universe believe that Google's main competitor is OpenAI, at most a 100 billion dollar startup? Or is Google really in an existential battle with Microsoft, a 3 trillion dollar behemoth that owns full access to all of OpenAI's technology that was created with its $13 billion investment???</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Yes, Google changed the name of its chatbot, after Microsoft changed the name of its chatbot from "Bing Chat" to "Copilot: .. And Google now offers a subscription plan that provides access to its most powerful LLM; after Microsoft offered subscriptions ... And Google integrated its chatbot into its other applications: again after Microsoft. Whereas Apple used to say "There's an app for that", Microsoft now says "There's a copilot for that" ... for Word, Excel, PowerPoint, Teams, GitHub, etc, etc, etc, etc, etc, etc, etc.</span></div><div style="text-align: left;"><br /></div><div style="text-align: left;"><span style="font-family: verdana;">Meanwhile, back at the Super Bowl, Google ran a commercial; Microsoft ran a commercial; but OpenAi did not. Google's commercial was a touching demonstration of how Google's AI enhanced Pixel smart phone enables visually challenged folk to take better pictures ... touching ... but Samsung makes the lion's share of And</span><span style="font-family: verdana;">roid phones. Why didn't Google tout its AI enhancements to Samsung's new Galaxy S24 phones, AND to its own new Pixel phones, AND to other new Android phones?</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Of course, Microsoft's commercial was on point. It showed all kinds of people trying to do all kinds of things. And guess what the tagline at the end of the commercial told them? Yup!!! There's a copilot that will help all of them do all of those things. <br /><ul style="text-align: left;"><li><span style="font-family: verdana;">For the benefit of the 210 million of you readers who did not watch the Super Bowl, here are links to the commercials ... <b><i><a href="https://t.co/ekd9swAYOU"><span style="color: #cc0000;">Google</span></a></i></b> vs. <b><i><a href="https://t.co/67kOsALG6r"><span style="color: #cc0000;">Microsoft</span></a></i></b></span></li></ul></span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><div style="background-color: white;"><div style="color: black;"><div id="Overview"><div id="Overview" style="color: #1a222a; font-family: Verdana, Geneva, sans-serif; font-size: 13.86px;"><span style="font-family: verdana;"><b style="font-size: large;">B. Top 4 stories in past week ...</b></span></div><div id="Overview" style="color: #1a222a;"><ol><li style="font-family: Verdana, Geneva, sans-serif; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;"><span><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#Policy" style="color: #336699; font-weight: 700; text-align: center; text-decoration-line: none;"><span style="color: #cc0000; font-size: x-small;">Public Policy</span></a><br /><span>"Google, Apple, Meta and other huge tech companies join US consortium to advance responsible AI", Sharon Goldman, </span><b><i><a href="https://t.co/ni4o4yflII" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Engadget</span></a></i></b><span> 2/8/24</span><br /><span>-- This story also covered by </span><a href="https://t.co/ni4o4yflII" style="color: #336699; font-style: italic; font-weight: bold; text-decoration-line: none;"><span style="color: #cc0000;">VentureBeat</span></a><span> ... and the </span><i><b><a href="https://www.commerce.gov/news/press-releases/2024/02/biden-harris-administration-announces-first-ever-consortium-dedicated" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">U.S. Dept of Commerce</span></a><br /><br /></b></i></span></span></li><li style="margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;"><span style="font-size: x-small;"><span style="font-weight: 700; text-align: center;"><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#Policy" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Public Policy</span></a></span><span style="font-weight: 700; text-align: center;"> </span></span><br /><span face="Verdana, Geneva, sans-serif" style="font-size: medium;">"F.C.C. Bans A.I.-Generated Robocalls", Cecilia Kang, </span><b style="font-family: Verdana, Geneva, sans-serif; font-size: large;"><i><a href="https://t.co/DHxjoEml8o" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">NY Times</span></a></i></b><span face="Verdana, Geneva, sans-serif" style="font-size: medium;">, 2/8/24 </span><b style="font-family: Verdana, Geneva, sans-serif; font-size: large;"><span style="color: #cc0000;">***</span></b><span face="Verdana, Geneva, sans-serif" style="font-size: medium;"> </span><br /><span face="Verdana, Geneva, sans-serif" style="font-size: medium;">-- This story also covered by </span><b style="font-family: Verdana, Geneva, sans-serif; font-size: large;"><i><a href="https://www.cnn.com/2024/02/08/tech/fcc-scam-robocalls-ai-generated-voices/index.html" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">CNN</span></a></i></b><span face="Verdana, Geneva, sans-serif" style="font-size: medium;">, </span><b style="font-family: Verdana, Geneva, sans-serif; font-size: large;"><i><span style="color: #cc0000;">Forbes</span></i></b><span face="Verdana, Geneva, sans-serif" style="font-size: medium;">, </span><a href="https://www.npr.org/2024/02/08/1230052884/the-fcc-says-ai-voices-in-robocalls-are-illegal" style="color: #336699; font-family: Verdana, Geneva, sans-serif; font-size: large; font-style: italic; font-weight: bold; text-decoration-line: none;"><span style="color: #cc0000;">NPR</span></a><b style="font-family: Verdana, Geneva, sans-serif; font-size: large; font-style: italic;"> ... </b><span face="Verdana, Geneva, sans-serif" style="font-size: medium;">and the </span></span><b style="font-family: Verdana, Geneva, sans-serif; font-size: 13.86px;"><i><a href="https://docs.fcc.gov/public/attachments/DOC-400393A1.pdf" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000; font-family: verdana; font-size: medium;">FCC</span></a><br /></i></b><br /></li><li style="font-family: Verdana, Geneva, sans-serif; font-size: 13.86px; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;"><span style="font-size: medium;"><b style="font-size: small; text-align: center;"><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#OpEds" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">OpEds</span></a></b><br />"(Dr. Evil Voice) Sam Altman Wants 7 Trillion Dollars", Michelle Cheng, <b><i><a href="https://t.co/FmlzTnTUgv" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Gizmodo</span></a></i></b>, 2/9/24 ... <b><span style="color: #cc0000;">*** </span></b> <i>... LOL, ROTF, LMFAO ... :-)</i><br />-- This story also covered by <b><i><a href="https://www.wsj.com/tech/ai/sam-altman-seeks-trillions-of-dollars-to-reshape-business-of-chips-and-ai-89ab3db0" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Wall Street Journal</span></a></i></b>, <b><i><a href="https://arstechnica.com/information-technology/2024/02/report-sam-altman-seeking-trillions-for-ai-chip-fabrication-from-uae-others/" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Ars Technica</span></a></i></b>, <b><i><a href="https://www.computerworld.com/article/3712684/sam-altman-wants-to-raise-trillions-of-dollars-for-ai-chip-initiative.html" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Computer World</span></a><br /><br /></i></b></span></span></li><li style="font-family: Verdana, Geneva, sans-serif; font-size: 13.86px; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;"><span style="font-size: medium;"><b><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#Google" style="color: #336699; font-size: small; text-align: center; text-decoration-line: none;"><span style="color: #cc0000;">Google</span></a><br style="font-size: 13.86px; font-weight: 400;" /><span style="font-size: medium; font-weight: 400;">"Google rebrands Bard chatbot as Gemini, rolls out paid subscription", Jeffrey Dastin, <b><i><a href="https://t.co/p2AkKLGRNx" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Reuters</span></a></i></b>, 2/8/24 <b><span style="color: #cc0000;">***</span></b> <br />-- This story also covered by <b><i><a href="https://t.co/qrUPLSl6Up" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">VentureBeat</span></a></i></b>, <b><i><a href="https://t.co/UE2KXFGz8x" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">NY Times</span></a></i></b>, <b><i><a href="https://www.zdnet.com/article/google-rebrands-bard-to-gemini-now-available-for-the-first-time-in-mobile/" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">ZDNet</span></a></i></b>, <b><i><a href="https://www.engadget.com/google-rebrands-its-bard-ai-chatbot-as-gemini-which-now-has-its-own-android-app-151303210.html" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Engadget</span></a></i></b>, <b><i><a href="https://t.co/6kd6yLvFKr" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">MIT Tech Review</span></a></i></b></span></b></span></span></li></ol></div></div></div><div style="font-size: medium;"><div style="color: #1a222a;"><span style="font-family: verdana; font-size: 13.86px;"><span style="font-size: medium;"><b><br /></b></span></span></div><div><span style="color: #1a222a; font-family: verdana; font-size: 13.86px;"><span style="font-size: medium;"><b>C. </b></span></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="background-color: transparent; color: #336699; font-family: verdana; font-size: large; text-decoration-line: none;"><span style="color: #1a222a; font-size: medium;">➡</span></a><span style="color: #1a222a;"> </span><b style="font-size: 18px;"><i><a href="https://neoskeptics.blogspot.com/2023/12/basic-ai-faqs.html"><span style="color: #cc0000; font-family: verdana;">Dozen Basic AI FAQs</span></a></i></b></div><div style="color: #1a222a;"><span style="font-family: verdana;"><span style="font-size: medium;">This page contains links to responses by Google's Bard chatbot running Gemini Pro to 12 questions that should be asked more frequently, but aren't. As consequence, too many readily understood AI terms have become meaningless buzzwords in the media.</span></span></div></div></div></div>
Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-75281143781261380842024-02-06T22:19:00.018-05:002024-03-27T21:37:43.406-04:00Big Tech's investor conference calls ... TL;DR and podcast 5Feb24<div style="text-align: right;"><span style="font-family: verdana; font-size: x-small;"><i>Last update: Wednesday 2/7/24 </i></span></div><div style="text-align: left;"><span style="font-family: verdana;"><div class="separator" style="clear: both; text-align: center;"><div style="text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><span style="font-size: medium;"><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgeClLPv35C1srAzx2eRPJ5i_V7Nw9eTbT3ICJJn4u2u7H3N_5gnHo-k2hgOHUzIvoMQ2stoQLw5yXAx9bwWp2iijgH8HLCJz89poy6X4s_AfomZeFiIEZlEINL3AnPWbZSyO3LeXII0Xy4qxPG0fdyLjRyNSl3JhlUkHUJzSdjVMGYhefXTd8llelLW8s/s1800/GoogleMicrosoftBannerRedBackground.jpeg" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="1012" data-original-width="1800" height="180" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgeClLPv35C1srAzx2eRPJ5i_V7Nw9eTbT3ICJJn4u2u7H3N_5gnHo-k2hgOHUzIvoMQ2stoQLw5yXAx9bwWp2iijgH8HLCJz89poy6X4s_AfomZeFiIEZlEINL3AnPWbZSyO3LeXII0Xy4qxPG0fdyLjRyNSl3JhlUkHUJzSdjVMGYhefXTd8llelLW8s/s320/GoogleMicrosoftBannerRedBackground.jpeg" vspace="10" width="320" /></a></div>Welcome to our 5Jan</span></span></span></span><span style="font-size: medium;"><span><span style="background-color: white; color: #1a222a;">24</span><span style="background-color: white; color: #1a222a;"> </span><span style="background-color: white; color: #1a222a;">TL;DR summary + podcast about the past week's top AI stories on our "</span><b style="color: #1a222a;"><i><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html">Useful AI News</a></i></b><span style="background-color: white; color: #1a222a;">" page </span></span><b style="background-color: white; color: #1a222a;">➡</b><span style="background-color: white; color: #1a222a;"> GenAI related comments by CEOs during investor conference calls about the latest quarterly reports from Google, Apple, Amazon, and Microsoft </span></span></div><div style="text-align: left;"><div><span style="color: #cc0000; font-family: verdana;"><b style="background-color: white; font-size: large;"><br /></b></span></div></div><div style="text-align: left;"><div class="separator" style="clear: both; font-size: 13.86px;"><div style="text-align: right;"><div style="color: #1a222a;"><span style="font-size: medium;">Click link for podcast (opens in new tab)</span> </div><div><span style="color: #1a222a;"> </span><span style="color: #cc0000;"><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="background-color: white; color: #336699; font-size: large; text-align: left; text-decoration-line: none;"><span style="color: #1a222a; font-size: medium;">➡</span></a> <i><span style="font-size: medium;"><b><a href="https://drive.google.com/file/d/1PIEG_yw0M34RtFOVTbnsnUeOJmnjYPRQ/view?usp=sharing" target="_blank">Google + Apple + Amazon + Microsoft</a></b></span></i><i><span style="font-size: medium;"><b> </b></span></i></span></div></div></div><div class="separator" style="clear: both;"><div style="color: #1a222a; font-size: 13.86px; text-align: right;"><span style="color: black;"><span style="color: #1a222a;">Click the "start" button the when the podcast page is loaded<br />... </span></span> If audio fails to start, or gets stuck, try reloading the page</div></div></div><div style="background-color: white; text-align: right;"><div class="separator" style="clear: both; text-align: left;"><span face="Verdana, Geneva, sans-serif" style="color: #1a222a; font-size: medium;">TL;DR link<b> </b></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-size: large; text-decoration-line: none;"><span style="color: #1a222a;">➡</span></a><b style="font-family: Verdana, Geneva, sans-serif;"><span style="font-size: medium;"><span style="color: #1a222a;"> </span><i><a href="https://neoskeptics.blogspot.com/2024/02/big-techs-investor-conference-calls.html#more"><span style="color: #cc0000;">HERE</span></a></i></span></b></div></div></div></span></div><span><span style="font-family: verdana;"><a name='more'></a></span></span><div style="text-align: left;"><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;">A. </b><b style="background-color: white; font-family: verdana; font-size: large;">TL;DR ... </b><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;">Top 4 stories in past week ...</b></div><div style="text-align: left;"><span style="font-family: verdana;">Our top stories last week were about the Big Tech Five minus Facebook, i.e., about Amazon, Apple, Google, and Microsoft. To be specific, our stories related to GenAI comments made by their CEOs during investors' conference calls that discussed their last quarterly earnings reports. </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">The challenge to the Big Tech CEOs was straight-forward: If a company was making billion dollar investments in generative AI, its CEO had to (1) identify a large market segment wherein the CEO's company was dominant, and (2) convince potential investors that the company had a strategy that would enable it to use generative AI to derive substantially larger revenue streams from that segment.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">At this point, </span><span style="font-family: verdana;">Facebook has invested billions upon billions into the virtual reality of the "metaverse", but not in generative AI; so its massive increase in profits in the last quarter was not related to its celebrity chatbots or to any other applications of GenAI.</span></div><div style="text-align: left;"><b><span style="font-family: verdana; font-size: medium;"><br /></span></b></div><div style="text-align: left;"><b><span style="font-family: verdana; font-size: medium;">1) Google's large, but diappointing earnings</span></b></div><div style="text-align: left;"><span style="font-family: verdana;">According to the Wall Street Journal, "Sales in Google’s cloud-computing business climbed 26%, recovering somewhat following a disappointing third quarter that some investors viewed as a sign of weak demand for the company’s AI services."</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Similarly, the NY Times commented that "Alphabet, Google’s parent company, on Tuesday reported search revenue and a profit margin for its latest quarter that fell short of Wall Street’s expectations, in a sign that growth in its flagship business and recent layoffs, intended to cut costs, were not enough to offset its growing investment in artificial intelligence."</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Google dominates search and has a substantial share of cloud computing; however its search and mapping services are free to end users, but paid for by advertisers. Unfortunately, it was unable to convince investors that its increases in cloud sales were powered by increases in AI-driven search.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><b><span style="font-family: verdana; font-size: medium;">2) Apple GenAI coming "later this year"</span></b></div><div style="text-align: left;"><span style="font-family: verdana;">Apple has not shared any data about its investments in generative AI. So CEO Tim Cook's following statements, as quoted by The Verge, command our attention:</span></div><div style="text-align: left;"><blockquote><span style="font-family: verdana;">"Our M.O., if you will, has always been to to do work and then talk about work, and not to get out in front of ourselves. And so we’re going to hold that to this as well. But we have got some things that we’re incredibly excited about, that we’ll be talking about later this year ... Let me just say that I think there’s a huge opportunity for Apple with generative AI and with AI, without getting into many more details or getting out ahead of myself,” Cook said to conclude the call.</span></blockquote></div><div style="text-align: left;"><span style="font-family: verdana;">Many questions come to mind: Why has Apple taken so long to get to the starting line of this race? Why is Cook so confident that Apple will start offering genAI services soon? And what changed between March 2023 (the launch of GPT-4) when the other four Big Five Techs got into the race and now?</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Lack of data encourages speculation. Throughout most of 2023, generative AI has been dominated by large language models (LLMs), models that needed to know everything about everything, models that were therefore trained on all accessible digital information, even when the developers had questionable rights to access such data, even when the data entailed violations of individual privacy, even when the models required extensive "prompting" for users to obtain the desired results, and even when developers had to remind users to check all results for possible errors and hallucinations. All of this sounds clumsy, rude, and very un-Apple, a company whose highly paid designers work overtime to ensure the smoothest possible user experience wherein "everything just works" because of "Apple magic".</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Apple's iCloud is housed in Google's cloud. According to <b><i><a href="Apple’s Spending on Google Cloud Storage On Track to Soar 50% This Year">The Infomation</a></i></b>, Apple is the largest corporate client for Google’s storage. If Apple pursued LLMs, it would have to pay the premium prices Google charges for its graphic processing units (GPUs), the costs that make LLMs so expensive to operate, no matter whose cloud houses them.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">However, as noted in reports in various tech media since the summer of 2023, many observers have identified the surprising power of small language models (SLMs), i.e., small focused models that exhibit the same kinds of emergent cognitive skills as large language models (LLMs), but don't know "everything".</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">So we have to speculate that Apple will finally launch its GenAI initiatives later this year because it will mainly deploy relatively inexpensive small language models housed on its users' iPhones and Macs, rather than expensive LLMs housed on Google's cloud. But Apple will probably make comparatively larger investments in the design of elegant user interfaces for its chatbots, "elegant" as in simple, but powerful interfaces that will not require users to bumble through prompt engineering.</span></div><div style="text-align: left;"><br /></div><div style="text-align: left;"><b><span style="font-family: verdana; font-size: medium;">3) </span></b><b><span style="font-family: verdana; font-size: medium;">Amazon's new chatbot</span></b></div><div style="text-align: left;"><span style="font-family: verdana;">A few weeks ago, Amazon announced a new chatbot. called "Q", but nobody cared. What market segment dominated by Amazon would provide a user base for Q? Amazon offered no answers. Indeed, what market segment does Amazon dominate, besides the cloud? Duh ... online shopping? Correct-a-mundo!!! </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">This week Amazon offered a new chatbot called "Rufus" that will be its version of a Microsoft copilot. Whereas Microsoft's copilots help users use Microsoft's office productivity apps more effectively, Rufus will help users go shopping on Amazon more efficiently. According to the NY Times:</span></div><div style="text-align: left;"><blockquote><span style="font-family: verdana;">"Customers can ask the tool, Rufus, product questions directly in the search bar of the company’s mobile app, Amazon said in a blog post. The A.I. will then provide answers in a conversational tone. The examples provided in the announcement included comparing different kinds of coffee makers, recommendations for gifts and a follow-up question about the durability of running shoes."</span></blockquote><p><span style="font-family: verdana;">Before Rufus, most customers used Google to search for information about the kinds of products they were interested in buying; then they went to Amazon to buy the most cost-effective version of that product in their preferred size, color, etc.</span> <span style="font-family: verdana;">The NY Times does not indicate whether Rufus will be based on a large language model or a small language model. But common sense suggests that powerful small models would be the most profitable option for Amazon</span></p></div><div style="text-align: left;"><b><span style="font-family: verdana; font-size: medium;">4) Microsoft's current LLM success finances its future SLM success</span></b></div><div style="text-align: left;"><span style="font-family: verdana;">We assume that most readers of this blog understand the secret of Microsoft's success in 2023:</span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">Microsoft dominates the world-wide market for office productivity apps</span></li><li><span style="font-family: verdana;">So Microsoft converted OpenAI's ChatGPT into copilots for each of its productivity apps and charged a monthly subscription fee for each member of the enterprise who used a copilot for a productivity app.</span></li></ul><div><span style="font-family: verdana;">Unfortunately, it cost Microsoft billions of dollars to underwrite OpenAI's development of ChatGPT based on large language models (GPT-3.5, GPT-4), and billions more to operate its copilots. Fortunately Microsoft's own researchers documented the surprising power of small language models by the end of 2023. Indeed, during the conferencer call, The Information reported that Microsoft's CEO, Satya Nadella:</span></div><div><blockquote><span style="font-family: verdana;">"... </span><span style="background-color: white; font-family: verdana;">also touted the success of Microsoft’s “small language models” that are a less costly option for customers, compared to OpenAI’s large language models. Nadella said AT&T and Thomson Reuters were testing out the SLMs on Azure <i>[Microsoft's cloud]</i>. As we </span><b style="background-color: white; font-family: verdana;"><i><a href="https://www.theinformation.com/articles/microsoft-forms-team-to-make-cheaper-generative-ai?rc=v6kkoz&shared=abbca979dbb6e4c8">reported</a></i></b><span style="background-color: white; font-family: verdana;"> last week, the company formed a new group dedicated to SLMs."</span></blockquote></div></div><div style="text-align: left;"><span style="font-family: verdana;">Nadella's successful </span><span style="font-family: verdana;">presentation led to yet another rise in the price of Microsoft's stock</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><i><b>Correction</b>: The TL;DR (and podcast) asserted that all four top stories were based on CEOs' comments during conference calls. This was not true for Amazon's story. The NY Times based its report on an announcement by Amazon. No other major tech publication covered this announcement by the time the editor composed the TL;DR.<br /><br /></i></span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><div style="background-color: white; color: #1a222a;"><div style="color: black; font-size: medium;"><div id="Overview"><span><span style="font-family: verdana;"><b style="font-size: large;">B. <b>Top 4 stories in past week ...</b><div id="Overview" style="font-family: Times; font-size: medium; font-weight: 400;"><ol><li><span style="font-size: medium;"><span style="font-family: verdana;"><a href="#Google" style="font-size: small; font-weight: 700; text-align: center;"><span style="color: #cc0000;">Google</span></a><br />"Google’s Ad Sales Fall Short of Wall Street’s Lofty Expectations", Miles Kruppa, <b><i><a href="https://t.co/Jo4TAL1Qja"><span style="color: #cc0000;">Wall Street Journal</span></a></i></b>, 1/30/24 <br />-- This story also covered by <b><i><a href="https://t.co/c1F499A2kb"><span style="color: #cc0000;">NY Times</span></a></i></b>, <b><i><a href="https://t.co/T78FFg8Gus"><span style="color: #cc0000;">Forbes</span></a>,</i></b> and </span><b style="font-family: verdana;"><i><a href="https://www.theinformation.com/articles/microsoft-dishes-on-ai-revenue-google-ceo-says-agents-are-coming?rc=v6kkoz&shared=cf317c231125d962"><span style="color: #cc0000;">The Information</span></a></i></b><span style="font-family: verdana;">, 1/31/24 ... Link downloads the full paywalled article for 3 free reads.<br /><br /></span></span></li><li><span style="font-family: verdana;"><a href="#OtherLLMs" style="font-size: small; font-weight: 700; text-align: center;"><span style="color: #cc0000;">Other Models</span></a><br /><span style="font-size: medium;">"Tim Cook confirms Apple’s generative AI features are coming ‘later this year’", Chris Welch, <b><i><a href="https://t.co/j5ynWuPPb2"><span style="color: #cc0000;">The Verge</span></a></i></b>, 2/1/24 <b><span style="color: #cc0000;">***</span></b> <br />-- This story also covered by </span><b><i><span style="font-size: medium;"><a href="https://www.cnbc.com/2024/02/01/tim-cook-teases-apple-ai-announcement-later-this-year.html"><span style="color: #cc0000;">CNBC</span></a>, <a href="https://www.zdnet.com/article/apple-reportedly-eyeing-generative-ai-push-and-siri-overhaul-for-the-iphone/"><span style="color: #cc0000;">ZDNet</span></a>, <a href="https://t.co/aIAuJ6L0Y7"><span style="color: #cc0000;">Computerworld</span></a><br /><br /></span></i></b></span></li><li><span style="font-family: verdana;"><span style="font-size: medium;">"Amazon Enters Chatbot Fray With Shopping Tool", Karen Weise<b style="font-style: italic;">, </b><b><i><a href="https://t.co/QNnzCZjMWw"><span style="color: #cc0000;">NY Times</span></a></i></b><span style="font-size: medium;">, 2/3/24 </span><span style="font-size: medium;"> </span><br /></span><b style="font-style: italic;"> </b></span></li><li><span style="font-size: medium;"><span style="font-family: verdana;"><a href="#Microsoft" style="font-size: small; text-align: center;"><b><span style="color: #cc0000;"><span>M</span>icrosoft</span></b></a><br />"Microsoft Dishes on AI Revenue; Google CEO Says ‘Agents’ Are Coming", Aaron Holmes and Jon Victor, <b><i><a href="https://www.theinformation.com/articles/microsoft-dishes-on-ai-revenue-google-ceo-says-agents-are-coming?rc=v6kkoz&shared=cf317c231125d962"><span style="color: #cc0000;">The Information</span></a></i></b>, 1/31/24 ... Link downloads the full paywalled article for 3 free reads<br />-- "CEO Satya Nadella also touted the success of Microsoft’s “small language models” that are a less costly option for customers, compared to OpenAI’s large language models. Nadella said AT&T and Thomson Reuters were testing out the SLMs on Azure. As we <b><i><a href="https://www.theinformation.com/articles/microsoft-forms-team-to-make-cheaper-generative-ai?rc=v6kkoz&shared=abbca979dbb6e4c8">reported</a></i></b> last week, the company formed a new group dedicated to SLMs."<br />-- Note: As per its title, this article also discusses Google's earnings call<br />-- Microsoft's quarterly earnings also reported by <b><i><a href="https://t.co/kZrUG7BHAi"><span style="color: #cc0000;">Reuters</span></a></i></b>, <b><i><a href="https://t.co/HE7dmXrn7i"><span style="color: #cc0000;">NY Times</span></a></i></b>, <br /><br /></span></span></li></ol></div></b></span></span></div></div><div style="color: black;"><div style="color: #1a222a;"><div style="font-family: verdana; font-size: 13.86px;"><b><span id="Basics" style="font-size: medium;">C. Basics </span></b></div><div><ul style="line-height: 1.4; list-style-image: initial; list-style-position: initial; margin: 0.5em 0px; padding: 0px 2.5em;"><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;">"Watch an A.I. Learn to Write by Reading Nothing but Shakespeare or Harry Potter or Jane Austen or Star Trek or Moby Dick", Aatish Bhatia, <i><a href="https://www.nytimes.com/interactive/2023/04/26/upshot/gpt-from-scratch.html?campaign_id=29&emc=edit_up_20230427&instance_id=91193&nl=the-upshot&regi_id=59528930&segment_id=131497&te=1&user_id=e4c6f77d7013b535754fe6ac3b8162ff" style="color: #336699; text-decoration-line: none;"><b>NY Times</b></a></i><b>,</b> 4/27/23 </span></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;">"What Is a Large Language Model, the Tech Behind ChatGPT?", Kurt Muehmel, <b><i><a href="https://blog.dataiku.com/large-language-model-chatgpt?utm_id=6324578214330--6324578214130--6324581648930&utm_source=nam-facebook&utm_medium=paid-social&utm_campaign=GLO+CONTENT+ChatGPT+%26+LLMs+March+2023&fbclid=PAAablIjJk6gcf25O1Gr65_Ch0-yDP70GcYS1ft8ygHUBzlJ9ugQnkb4_inzE_aem_th_AdrlIXtqBGq2TriUVchowbil7U6Mxz2isdEGCUVdY1UYednqTeuBB2E2Jhf7F7sjQgPAirGHQNL3bj7ZRq1AlBZt" style="color: #336699; text-decoration-line: none;">Data Iku</a></i></b>, 6/7/23 </span></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;">"Let's learn about artificial intelligence -- A series about AI, machine learning, ChatGPT, and more", Mark Wiemer, <b><i><a href="https://markwiemer.medium.com/lets-learn-about-artificial-intelligence-b74ca3b68eb5" style="color: #336699; text-decoration-line: none;">Medium</a></i></b>, 3/21/23<br /><br /></span></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;">"Textbooks Are All You Need II: phi-1.5 technical report", Yuanzhi Li, Sébastien Bubeck, Ronen Eldan, Allie Del Giorno, Suriya Gunasekar, Yin Tat Lee, <b><i><a href="https://www.microsoft.com/en-us/research/publication/textbooks-are-all-you-need-ii-phi-1-5-technical-report/" style="color: #336699; text-decoration-line: none;">Microsoft Research</a></i></b>, September 2023</span></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;">"Phi-2: The surprising power of small language models", Mojan Javaheripi and Sébastien Bubeck , <b><i><a href="https://www.microsoft.com/en-us/research/blog/phi-2-the-surprising-power-of-small-language-models/" style="color: #336699; text-decoration-line: none;">Microsoft Research</a></i></b>, 12/12/24<br /><br /></span></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;"><span style="color: #cc0000; font-style: italic; font-weight: bold;">Dozen Basic AI FAQs</span><br />This page contains links to responses by Google's Bard chatbot to 12 questions that should be asked more frequently, but aren't. </span></li></ul></div></div></div></div></div>
Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-2206882446894412452024-01-29T16:05:00.024-05:002024-03-27T21:37:56.826-04:00Microsoft's new “small” AI team ... TL;DR and podcast 28Jan24 <div style="text-align: right;"><span style="font-family: verdana; font-size: x-small;"><i>Last update: Monday 1/31/24 </i></span></div><div style="text-align: left;"><span style="font-family: verdana;"><div class="separator" style="clear: both; text-align: center;"><div style="text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><span style="font-size: medium;"><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi-HqqoAp3YjC45UbsG1xW41weaELKIjJIwRKcsLds6hxLIXCA7_-R4BZKaHm_paq4d6lE1wuipVxs-mcBFULBIIn6I3Z8UBMIwHeJil1au9b71wCemos3hOqQ3QUAi0YUiRrFKbb5SL3byawW1hQjYPpt9_nabWYVWm0-Q6qQFrGw1yCipsTegq_m-PvE/s1200/MicrosoftClassroom.webp" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="675" data-original-width="1200" height="180" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi-HqqoAp3YjC45UbsG1xW41weaELKIjJIwRKcsLds6hxLIXCA7_-R4BZKaHm_paq4d6lE1wuipVxs-mcBFULBIIn6I3Z8UBMIwHeJil1au9b71wCemos3hOqQ3QUAi0YUiRrFKbb5SL3byawW1hQjYPpt9_nabWYVWm0-Q6qQFrGw1yCipsTegq_m-PvE/s320/MicrosoftClassroom.webp" vspace="10" width="320" /></a></div><br />Welcome to our 28Jan</span></span></span></span><span style="font-size: medium;"><span><span style="background-color: white; color: #1a222a;">24</span><span style="background-color: white; color: #1a222a;"> </span><span style="background-color: white; color: #1a222a;">TL;DR summary + podcast about the past week's top AI story on our "</span><b style="color: #1a222a;"><i><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html">Useful AI News</a></i></b><span style="background-color: white; color: #1a222a;">" page </span></span><b style="background-color: white; color: #1a222a;">➡</b><span style="background-color: white; color: #1a222a;"> Microsoft creates new team to build “small” AI, </span></span></div><div style="text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><br /></span></span></span></div><div style="text-align: left;"><div class="separator" style="clear: both;"><div style="text-align: right;"><div style="color: #1a222a;"><span>Click link for podcast (opens in new tab)</span><span style="font-size: 13.86px;"> </span></div><div><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-size: large; text-align: left; text-decoration-line: none;"><span style="color: #1a222a; font-size: medium;">➡</span></a><span style="font-size: medium;"><span style="color: #1a222a;"> </span><b><i><span style="background-color: white; text-align: left;"><a href="https://drive.google.com/file/d/1P7HCdvaqrmH8OpkzHQYXbr3aLCj9YSLg/view?usp=sharing" target="_blank"><span style="color: #cc0000;">Microsoft creates new “small” AI team</span></a><span style="color: #1a222a;"> </span></span></i></b></span><i style="color: #1a222a;"><span style="font-size: medium;"><b><span style="font-size: 13.86px;"> </span></b></span></i></div></div></div><div class="separator" style="clear: both;"><div style="color: #1a222a; font-size: 13.86px; text-align: right;"><span style="color: black;"><span style="color: #1a222a;">Click the "start" button when the podcast page is loaded<br />... </span></span> If audio fails to start, or gets stuck, try reloading that page.</div></div></div><div style="background-color: white; text-align: right;"><div class="separator" style="clear: both; text-align: left;"><span face="Verdana, Geneva, sans-serif" style="color: #1a222a; font-size: medium;">TL;DR link<b> </b></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-size: large; text-decoration-line: none;"><span style="color: #1a222a;">➡</span></a><b style="font-family: Verdana, Geneva, sans-serif;"><span style="font-size: medium;"><span style="color: #1a222a;"> </span><i><a href="https://neoskeptics.blogspot.com/2024/01/microsofts-new-small-ai-team-tldr-and.html#more"><span style="color: #cc0000;">HERE</span></a></i></span></b></div></div></div></span></div><span><span style="font-family: verdana;"><a name='more'></a></span></span><div style="text-align: left;"><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;">A. </b><b style="background-color: white; font-family: verdana; font-size: large;">TL;DR ... </b><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;">Top story in past week ... </b></div><div style="text-align: left;"><span style="font-family: verdana;">Although this edition of our TL;DR only identified one "must-read" top story on our <b><i><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html">Useful AI News</a></i></b> page last week, our readers should understand that our news page also identified other articles that they might also find of interest. </span><span style="font-family: verdana;">For example:</span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">The U.S. Department of Commerce announced its intention to impose a requirement that AI companies notify the U.S. government of their intentions to develop language models that exceeded a specified size <b>before</b> they developed those models. Of course this restriction was conceived <b>before</b> it was generally recognized that small language models had surprising power. So small models, smaller than the required notification level, might also pose unacceptable risks.</span></li></ul></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">Microsoft announced a pro version of Microsoft Copilot that would provide support for the apps in its productivity suite and would be available to all users who paid a monthly subscription fee<br /><br /></span></li><li><span style="font-family: verdana;">Google announced Lumiere, a space-time diffusion model for generating realistic AI videos<br /><br /></span></li><li><span style="font-family: verdana;">Stability AI announced a small language model (1.6 billion parameters) that's smaller than Microsoft's Phi-2 (2.7 bilion parameters)</span></li></ul></div><div style="text-align: left;"><span style="font-family: verdana;">Now let's turn to our top story</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><b><span style="font-family: verdana; font-size: medium;">1) </span></b><b style="color: #1a222a; font-family: verdana; font-size: large; text-align: right;"><span id="slm" style="background-color: white; text-align: left;">Microsoft creates a new small language model (SLM) team </span></b></div><div style="text-align: left;"><span style="font-family: verdana;">This story was reported exclusively in a tech publication called "The Information", with all but the first few words of the article posted behind an expensive paywall. </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Another publication, Windows Central, published a summary of the paywalled piece. A third tech publication, The Verge, published a less extensive summary that was consistent with the summary posted by Windows Central. Fortunately, the high significance of Microsoft's initiative only depends on two key points:</span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">Microsoft has formed a team of some of its best techs to develop small language models (SLMs).<br /></span></li><li><span style="font-family: verdana;">The team will report directly to Kevin Scott, Microsoft's chief technology officer (CTO), the same visionary top level manager who managed Microsoft's partnership with OpenAI that resulted in the successful transfer and transformation of OpenAI's large language model technology into copilots for Microsoft's office productivity apps.</span></li></ul></div><div style="text-align: left;"><span style="font-family: verdana;">That's all we need to know. Why? Because Microsoft's recently announced small language model, <b><i><a href="https://www.microsoft.com/en-us/research/blog/phi-2-the-surprising-power-of-small-language-models/">Phi-2</a></i></b>, has already provided Microsoft with proof-of-concept that SLM's trained on high quality data have surprising power that is comparable to the power of LLM's that are 500 times as large. </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">It costs Microsoft billions to fund OpenAI's development of its large language GPT models and is costing billions to operate. If powerful SLMs prove to only cost millions to develop and millions to operate, they will provide Microsoft with more customers, more revenue, and more profits.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">That Microsoft should decide to quietly launch a "Manhattan Project" to develop powe</span><span style="font-family: verdana;">rful SLMs using its own in-house tech talent, instead of subcontracting this development to OpenAI, is to be expected. The stunning instability shown by OpenAI during the "</span><b style="font-family: verdana;"><i><a href="https://neoskeptics.blogspot.com/2023/11/the-ai-big-bang-aka-fall-and-rise-of.html">AI Big Bang</a></i></b><span style="font-family: verdana;">" weekend in late November 2023, when OpenAI's CEO Sam Altman was abruptly fired without any warning then rehired a few days later, might have had catastrophic consequences for Microsoft. Fortunately for Microsoft, this catastrophe was averted by the masterful management of the crisis by Satya Nadella, Microsoft's CEO.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><div style="background-color: white; color: #1a222a;"><div style="color: black;"><div id="Overview" style="font-size: medium;"><span><span style="font-family: verdana;"><b style="font-size: large;">B. </b></span></span><b style="color: #1a222a; font-family: verdana; font-size: large;">Top story in past week </b><b style="font-family: verdana; font-size: large;">... out of 9 new stories </b></div><div id="Overview" style="color: #1a222a; font-family: Verdana, Geneva, sans-serif;"><ol><li style="margin: 0px 0px 0.25em; padding: 0px;"><span style="font-size: medium;"><span style="font-family: verdana;"><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#Microsoft" style="color: #336699; font-size: small; text-align: center; text-decoration-line: none;"><b><span style="color: #cc0000;">Microsoft</span></b></a><br /></span></span><span style="font-family: verdana;"><span style="font-size: medium;">"Exclusive: Microsoft has created a new team to build “small” AI that’s cheaper than OpenAI’s.", Aaron Holmes, <b><i><a href="https://t.co/yILK3YitED" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">The Information</span></a></i></b>, 1/23/24 <b><span style="color: #cc0000;">***</span></b> ... Link now presents the full paywalled article for 3 free reads 1/31/24<br /><br />-- An extensive summary of this "exclusive" was published by <a href="https://t.co/nggWdO4HxS" style="color: #336699; font-style: italic; font-weight: bold; text-decoration-line: none;"><span style="color: #cc0000;">Windows Central</span></a>. A brief summary also appeared in <a href="https://t.co/IwjlQwLAjQ" style="color: #336699; font-style: italic; font-weight: bold; text-decoration-line: none;"><span style="color: #cc0000;">The Verge</span></a></span><br /><br /></span></li></ol></div></div><div style="color: black;"><div style="color: #1a222a;"><span style="font-family: verdana; font-size: 13.86px;"><span style="font-size: medium;"><b>C. SLMs and other background for understanding Microsoft's "Manhattan Project"</b></span></span></div><div style="color: #1a222a;"><span style="font-family: verdana;"><span><ul style="line-height: 1.4; list-style-image: initial; list-style-position: initial; margin: 0.5em 0px; padding: 0px 2.5em;"><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;">"Textbooks Are All You Need II: phi-1.5 technical report", Yuanzhi Li, Sébastien Bubeck, Ronen Eldan, Allie Del Giorno, Suriya Gunasekar, Yin Tat Lee, <b><i><a href="https://www.microsoft.com/en-us/research/publication/textbooks-are-all-you-need-ii-phi-1-5-technical-report/" style="color: #336699; text-decoration-line: none;">Microsoft Research</a></i></b>, September 2023</li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;">"Phi-2: The surprising power of small language models", Mojan Javaheripi and Sébastien Bubeck , <b><i><a href="https://www.microsoft.com/en-us/research/blog/phi-2-the-surprising-power-of-small-language-models/" style="color: #336699; text-decoration-line: none;">Microsoft Research</a></i></b>, 12/12/24</li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;">"The Inside Story of Microsoft’s Partnership with OpenAI", Charles Duhigg, <b><i><a href="https://www.newyorker.com/magazine/2023/12/11/the-inside-story-of-microsofts-partnership-with-openai" style="color: #336699; text-decoration-line: none;">The New Yorker</a></i></b>, 12/1/23 </li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;">"Exploring The Future: 5 Cutting-Edge Generative AI Trends In 2024", Janakiram MSV, <b><i><a href="https://t.co/VT8wmx5RJ8" style="color: #336699; text-decoration-line: none;">Forbes</a></i></b>, 1/2/24</li></ul></span></span></div><div style="color: #1a222a;"><span style="font-family: verdana; font-size: 13.86px;"><span style="font-size: medium;"><b><br /></b></span></span></div><div style="color: #1a222a;"><span style="font-family: verdana; font-size: 13.86px;"><span style="font-size: medium;"><b>D. </b></span></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="background-color: transparent; color: #336699; font-family: verdana; font-size: large; text-decoration-line: none;"><span style="color: #1a222a; font-size: medium;">➡</span></a> <b style="font-size: 18px;"><i><a href="https://neoskeptics.blogspot.com/2023/12/basic-ai-faqs.html"><span style="font-family: verdana;">Dozen Basic AI FAQs</span></a></i></b></div><div style="color: #1a222a; font-size: medium;"><span style="font-family: verdana;"><span style="font-size: medium;">This page contains links to responses by Google's Bard chatbot running Gemini Pro to 12 questions that should be asked more frequently, but aren't. As consequence, too many readily understood AI terms have become meaningless buzzwords in the media.</span></span></div></div></div></div>
Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-30396380555444190662024-01-21T21:28:00.050-05:002024-03-27T21:39:17.011-04:00Sam Altman (AI's Oracle at Davos) and AI smartphones ... TL;DR and podcast 21Jan24 <div style="text-align: right;"><span style="font-family: verdana; font-size: x-small;"><i>Last update: Sunday 1/21/24</i></span></div><div style="text-align: left;"><span style="font-family: verdana;"><div class="separator" style="clear: both; text-align: center;"><div style="text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><span style="font-size: medium;"><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiYs_JhZXwpGDEQZgzjbchO7to0nROof3o39iB5UafGQKf0AkOosqxuCpTkHdDVKB5Hf_NzfZkCw0R91Q7iwJ_fOtnnRvPyQTtFFF78Lhzf_t7engB3oMCupFHNTqmzMykINc3iFUg8rhW-4zH_pqyUW5w8i9aTY2hWiYO8tbGPKQXd-JPKsevtQMTQynI/s1562/SamAltManAtDavos2024.png" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="762" data-original-width="1562" height="156" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiYs_JhZXwpGDEQZgzjbchO7to0nROof3o39iB5UafGQKf0AkOosqxuCpTkHdDVKB5Hf_NzfZkCw0R91Q7iwJ_fOtnnRvPyQTtFFF78Lhzf_t7engB3oMCupFHNTqmzMykINc3iFUg8rhW-4zH_pqyUW5w8i9aTY2hWiYO8tbGPKQXd-JPKsevtQMTQynI/s320/SamAltManAtDavos2024.png" vspace="10" width="320" /></a></div>Welcome to our 21Jan</span></span></span></span><span style="font-size: medium;"><span><span style="background-color: white; color: #1a222a;">24</span><span style="background-color: white; color: #1a222a;"> </span><span style="background-color: white; color: #1a222a;">TL;DR summary + podcast about the past week's top AI stories on our "</span><b style="color: #1a222a;"><i><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html">Useful AI News</a></i></b><span style="background-color: white; color: #1a222a;">" page <b><br /></b></span></span><span style="background-color: white; color: #1a222a;"><b><br /></b>1) AI's Oracle at Davos, and (2) Samsung's new AI smartphones</span></span></div><div style="text-align: left;"><br /></div><div style="text-align: right;"><span style="background-color: white; color: #1a222a;">Click link for podcast (opens in new tab) </span></div><div style="text-align: right;"><span style="background-color: white; color: #1a222a; font-size: 13.86px;"> </span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-size: large; text-align: left; text-decoration-line: none;"><span style="color: #1a222a; font-size: medium;">➡</span></a> <a href="https://drive.google.com/file/d/1P60AXe-gPKinLTMlWu147NxiTtY-5lXh/view?usp=sharing" target="_blank"><b><i><span style="color: #cc0000; font-size: medium;">AltmanAtDavos + SamsungAIPhones</span></i></b></a></div><div style="text-align: right;"><span style="text-align: left;"><span style="background-color: white; color: #1a222a;"><span>Click the "start" button the when the podcast page is loaded </span></span></span></div><div style="text-align: right;"><span style="text-align: left;"><span style="background-color: white; color: #1a222a;"><span>... </span></span></span><span style="background-color: white; color: #1a222a; text-align: left;"> </span><span style="background-color: white; color: #1a222a; text-align: left;">If audio fails to start, or gets stuck, try reloading that page</span><span style="background-color: white; color: #1a222a; font-size: 13.86px; text-align: left;">.</span></div><div style="background-color: white; color: #1a222a; font-family: Verdana, Geneva, sans-serif; text-align: left;"><span style="font-family: verdana; font-size: medium;"></span></div><div style="background-color: white; font-family: Verdana, Geneva, sans-serif; text-align: right;"><div class="separator" style="clear: both; color: #1a222a; text-align: left;"><span style="font-family: verdana; font-size: medium;">TL;DR link<b> </b></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-family: verdana; font-size: large; text-decoration-line: none;"><span style="color: #1a222a;">➡</span></a><b style="color: black;"><span style="font-size: medium;"><span style="color: #1a222a;"> </span><i><a href="https://neoskeptics.blogspot.com/2024/01/sam-altman-ais-oracle-at-davos-and-ai.html#more"><span style="color: #cc0000;">HERE</span></a></i></span></b></div></div></div></span></div><span><span style="font-family: verdana;"><a name='more'></a></span></span><div style="text-align: left;"><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;">A. </b><b style="background-color: white; font-family: verdana; font-size: large;">TL;DR ... </b><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;">Top 2 stories in past week ...</b></div><div style="text-align: left;"><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;"><br /></b></div><div style="text-align: left;"><span style="font-family: verdana;"><b style="font-size: large;">1) Thus spake the AI Oracle at Davos, </b><b style="font-size: large;">Sam Altman</b><br />Sam Altman, the renowned CEO of Open AI, was interviewed at Bloomberg House in Davos while he attended the annual global conference of the wise and wealthy. A link to the video of the full 30 minute interview is embedded in the Bloomberg article referenced on this page's headline section (below). We have only noted three of Altman's many oracular proclamations in this summary. </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">The first expresses his disappointment that the New York Times had filed a massive copyright violation suit against Open AI and Microsoft last week:</span></div><div style="text-align: left;"><blockquote><span style="background-color: white; font-family: verdana;">"There is this belief held by some people that you need all my training data and my training data is so valuable"</span></blockquote></div><div style="text-align: left;"><span style="background-color: white; font-family: verdana;"><i><span style="font-style: normal;">Altman's comments suggest that he thinks that it can't be wrong to steal data from the New York Times as long as he steals enough data from everybody else so that OpenAI's models would be just about as effective without the NY Times data. </span></i></span></div><div style="text-align: left;"><span style="background-color: white; font-family: verdana;"><i><span style="font-style: normal;"><br /></span></i></span></div><div style="text-align: left;"><span style="background-color: white; font-family: verdana;"><i><span style="font-style: normal;">He would apply this same logic to every other publisher that threatened to sue OpenAI. </span></i><i><span style="font-style: normal;"><i><i><i><span style="font-style: normal;"><i><span style="font-style: normal;"><i><span style="font-style: normal;"><i><span style="font-style: normal;"><i><span style="font-style: normal;">In other words, under his leadership, OpenAI will not stop stealing data until so many publishers threaten to sue OpenAI simultaneously that OpenAi won't be able to create effective language models with whatever data remains undefended</span></i></span></i></span></i></span></i></span></i></i></i></span></i></span></div><div style="text-align: left;"><br /></div><div style="text-align: left;"><span style="font-family: verdana;">His position is clearly unethical; but we will have to wait for the courts to decide whether it is also illegal. This situation has ample precedent. For example, in the nineteenth century, employers forced their employees to work under abusive conditions. Employee abuse only ceased when workers banded together in unions whose strikes forced the employers to engage in collective bargaining. Further restraints were imposed on employers by legislators when unions became politically active.</span></div><div style="text-align: left;"><span style="background-color: white; font-family: verdana;"><br />Altman's next comment provides an implicit recognition that, going forward, OpenAI might be able to produce effective models that were trained on a substantially smaller amount of data.</span></div><div style="text-align: left;"><blockquote><span style="background-color: white; font-family: verdana;">"<i><span style="font-style: normal;">A lot of our research is how do we learn more from smaller amounts of very high quality data,”</span></i></span></blockquote></div><div style="text-align: left;"><span style="background-color: white; font-family: verdana;">Our research?, No. Microsoft's research;</span><span style="background-color: white; font-family: verdana;"> </span><span style="background-color: white; font-family: verdana;">Microsoft's research; </span><span style="background-color: white; font-family: verdana;">Microsoft's research; </span><span style="background-color: white; font-family: verdana;">not OpenAI's research. Microsoft's research has led to its recent publication that proclaims "</span><b style="background-color: white; font-family: verdana; font-style: italic;"><a href="https://www.microsoft.com/en-us/research/blog/phi-2-the-surprising-power-of-small-language-models/">the surprising power of small language models</a></b><span style="background-color: white; font-family: verdana;"><span style="font-size: small;">". This recent finding was not credible a few years ago when OpenAi started training its </span><span><span>GPT large language models,</span></span><span style="font-size: small;"> without permission, on copyrighted data from the NY Times, Reddit, etc, etc, etc.</span></span></div><div style="text-align: left;"><span style="background-color: white; font-family: verdana;"><br /><span style="font-size: small;">Here is, perhaps, his most telling comment, transcribed from the video of the interview. If you watch the video, you will see Altman making this heart-felt proclamation at </span></span><span style="background-color: white; font-family: verdana;">3 minutes, 45 seconds:</span></div><div style="text-align: left;"><span style="background-color: white; font-family: verdana;"><blockquote><span style="font-size: small;"> "</span>I believe that America is going to be fine, no matter what happens in this election" </blockquote></span></div><div style="text-align: left;"><span style="background-color: white; font-family: verdana;">Hmmmmm ... If Sam Altman were a chatbot, most of us -- whether Democrats, Republicans, or Independents -- would dismiss his assessment as a political hallucination ... but he isn't a chatbot ... A Stanford dropout, Altman received no academic training in AI as an undergraduate nor did he subsequently acquire any hands-on work experience as an AI developer. So what is he?</span></div><div style="text-align: left;"><span style="background-color: white; font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="background-color: white; font-family: verdana;">Although his official title is "CEO", chief executive officer, his de facto main role at OpenAI is "CSO", chief sales officer, because h</span><span style="background-color: white; font-family: verdana;">e is a brilliant salesman. </span><span style="background-color: white; font-family: verdana;">Recall that his salesmanship landed a $13 billion dollar contract from Microsoft to develop a series of expensive GPT models. So his comment about the forthcoming election is just part of his sales pitch to the next president, whoever that turns out to be. </span></div><div style="text-align: left;"><span style="background-color: white; font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="background-color: white; font-family: verdana;">As a retired tech, the editor of this blog still recalls the following lessons learned long ago. Techs know the product; salesmen know how to sell the product. If the product's quality falls below the quality of competitors' products, techs will change jobs; salesmen will merely change their sales pitches. </span></div><div style="text-align: left;"><span style="background-color: white; font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="background-color: white; font-family: verdana;">Anyone who really wants to know how a technology is most likely to evolve should ask the best techs, not the best salesmen. For AI we should trust world renowned AI experts, like Google's Demis Hassabis, or Facebook's Yann LeCun, or OpenAi's Ilya Sutskever. Indeed,</span><span style="background-color: white; font-family: verdana;"> back in November 2023, Sutskever </span><span style="background-color: white; font-family: verdana;">voted to fire Altman because he didn't trust him. So why should anyone else?</span></div><div style="text-align: left;"><span style="background-color: white; font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="background-color: white; font-family: verdana;">None of Altman's technical forecasts should be taken as plausible possibilities, but merely as a salesman's sales-pitch-of-the-moment. Like all great salesmen, Altman is always pitching.<br /></span></div><div style="text-align: left;"><b><span style="font-family: verdana; font-size: medium;"><br /></span></b></div><div style="text-align: left;"><span style="font-size: medium;"><b><span style="font-family: verdana;">2)</span><span style="font-family: verdana;"> </span></b><span style="background-color: white; color: #1a222a; font-family: verdana;"><b>Samsung's new AI smartphones</b></span></span></div><div style="text-align: left;"><span style="font-family: verdana;">Samsung hosted its most recent Galaxy Unpacked event on January 17, at the the SAP Center in San Jose, California. Most of the company's announcements related to its new S24 top of the line smartphones: Galaxy S24, Galaxy S24+, and Galaxy S24 Ultra. But from the focused perspective of this AI news page, its most important announcement was <b>Galaxy AI</b>, a set of AI-powered tools designed to work alongside the trio of next-generation phones. </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">More specifically, in the context of the AI boom launched last year as a result of the unexpected power of large language models, e.g., GPT-4, and the recently recognized surprising</span><span style="font-family: verdana;"> power of small language models, the editor of this blog wanted to find out which language models would be accessed by the Galaxy Ai tools and for which purposes.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Here are some key AI-driven capabilities noted in <b><i><a href="https://news.samsung.com/global/galaxy-unpacked-2024-highlights-from-galaxy-unpacked-the-promise-of-a-new-beginning-with-galaxy-ai">Samsung's press release</a></i></b>:</span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">New Ways To Connect: <br />-- <b>Live Translate</b> real-time, two-way call translations within the native call app.<br />-- <b>Chat Assist</b> immediately translates the message, allowing users to easily text with friends in the language they’re most comfortable speaking.<br />-- <b>Transcript Assist</b> enables users to create and share short summaries of recorded conversations <br />-- <b>Notes Assist </b>automatically formats and summarizes your notes with bullet points.<br /><br /></span></li><li><span style="font-family: verdana;">Google's<b> Circle to Search</b> which allows users to perform Google searches on anything they circle without switching apps </span></li></ul></div><div style="text-align: left;"><span style="font-family: verdana;">Which language models do the new phones use and when? The press release only indicates that Google is Samsung's primary AI partner, but h</span><span style="font-family: verdana;">ere's a quote from <b><i><a href="https://www.wired.com/story/samsung-galaxy-unpacked-2024/#intcid=_wired-verso-hp-trending_3ef88d6f-4ec2-4377-be30-c6605a00e83b_popular4-1">Wired</a></i></b>:</span></div><div style="text-align: left;"><blockquote><span style="font-family: verdana;"><span>"</span><span style="background-color: white; color: #1a1a1a; letter-spacing: 0.108px;">There's a mix of on-device (</span>via Gemini Nano<span style="background-color: white; color: #1a1a1a; letter-spacing: 0.108px;">) and cloud-based AI smarts (</span>via Gemini Pro<span style="background-color: white; color: #1a1a1a; letter-spacing: 0.108px;">) on the Galaxy S24 series, though Samsung leans more heavily on the latter.</span>"</span></blockquote></div><div style="text-align: left;"><span style="font-family: verdana;">Indeed, the Samsung press release only specifies that Circle to Search uses the on device model as the default option, but users can also decide to use the cloud version. Why does an edge device, like a smartphone, especially a high end smartphone, make such intensive use of the cloud?.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">This puzzling design suggests that Samsung, like Google and Amazon at the end of 2023, has rushed its new AI features to the market somewhat prematurely in order to challenge the de facto success of ChatGPT's large model as quickly as possible and thereby allay any concerns of Samsung's shareholders.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Finally, full disclosure requires that the editor of this blog declare that he has multiple Macs and only Macs in his home office, but he has been pleased by the performance of his Samsung S23 Ultra smartphone. However, he will switch to an iPhone in a nanosecond if Apple releases an iPhone that makes intensive use of small language models that are self-contained on his phone. Why? Because powerful small models won't need to share his healthcare and/or other personal AI queries with a corporate cloud.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><br /></div><div style="text-align: left;"><div style="background-color: white;"><div><div id="Overview" style="color: black; font-size: medium;"><span><span style="font-family: verdana;"><b style="font-size: large;">B. </b></span></span><b style="font-family: verdana; font-size: large;">Top stories in past week ...</b></div><div id="Overview"><ol><li><span style="font-family: verdana; font-size: medium;"><span style="color: black; font-size: medium;"><a href="#LLMNews" style="color: #cc0000; font-size: small; font-weight: bold; text-align: center;"><span style="color: #cc0000;">LLM News</span></a><br />"OpenAI Doesn’t Want to Train on New York Times Data After Lawsuit, Altman Says", Brad Stone and Jake Rudnitsky, <span style="font-size: medium;"><a href="https://www.bloomberg.com/news/articles/2024-01-16/openai-ceo-altman-says-it-doesn-t-want-to-train-on-new-york-times-after-lawsuit"><span style="color: #cc0000;"><b>Bloomberg</b></span></a></span><span style="color: black; font-size: medium;">, 1/18/24 </span><b style="font-size: medium;"><span style="color: #cc0000;">***</span></b><span style="color: black; font-size: medium;">-</span></span><br /><span style="color: black;">-- This story also covered by <b><a href="https://www.cnbc.com/2024/01/18/openai-ceo-on-nyt-lawsuit-ai-models-dont-need-publishers-data-.html"><span style="color: #cc0000;">CNN</span></a></b><span>,<br /><br /></span></span></span></li><li style="font-size: medium;"><span style="font-family: verdana; font-size: medium;"><a href="#LLMNews" style="color: black; font-size: small; font-weight: 700; text-align: center;"><span style="color: #cc0000;">LLM News</span></a><span style="color: black; font-size: medium;"><br /></span>"Here are the key differences between the Samsung Galaxy S24 phones", Sheena Vasani, <b><i><a href="https://www.theverge.com/24036693/samsung-galaxy-s24-plus-ultra-specs-comparison-price"><span style="color: #cc0000;">The Verge</span></a></i></b>, 1/18/24<br style="font-size: medium;" /><span style="font-size: medium;">-- This story also covered by <b><i><a href="https://t.co/1ses1a24kX"><span style="color: #cc0000;">Mashable</span></a></i></b>, </span><b style="color: black; font-size: medium;"><i><a href="https://www.cnet.com/tech/mobile/everything-announced-at-samsung-unpacked-s24-phones-galaxy-ai-and-one-surprise-reveal/"><span style="color: #cc0000;">CNET</span></a></i></b><span style="color: black; font-size: medium;">, </span><b style="color: black; font-size: medium;"><i><a href="https://www.theverge.com/24036693/samsung-galaxy-s24-plus-ultra-specs-comparison-price"><span style="color: #cc0000;">The Verge,</span></a></i></b><span style="color: black; font-size: medium;"> </span><b style="color: black; font-size: medium;"><i><a href="https://www.wired.com/story/samsung-galaxy-unpacked-2024/#intcid=_wired-verso-hp-trending_3ef88d6f-4ec2-4377-be30-c6605a00e83b_popular4-1"><span style="color: #cc0000;">Wired</span></a></i></b><span style="font-size: medium;">, <a href="https://www.bloomberg.com/news/articles/2024-01-17/samsung-seeks-double-digit-sales-growth-with-new-flagship-phone?utm_medium=social&cmpid%3D=socialflow-twitter-tech&utm_campaign=socialflow-organic&utm_source=twitter&utm_content=tech" style="color: black; font-style: italic; font-weight: bold;"><span style="color: #cc0000;">Bloomberg</span></a><b style="color: black; font-style: italic;"> </b>... and <b style="font-style: italic;"><a href="https://news.samsung.com/global/galaxy-unpacked-2024-highlights-from-galaxy-unpacked-the-promise-of-a-new-beginning-with-galaxy-ai"><span style="color: #cc0000;">Samsung</span></a></b></span></span></li></ol></div><div id="Overview" style="color: black; font-size: medium; text-align: center;"><span></span></div></div><div style="color: black; font-size: medium;"><div style="color: #1a222a;"><span style="font-family: verdana; font-size: 13.86px;"><span style="font-size: medium;"><b><br /></b></span></span></div><div style="color: #1a222a;"><span style="font-family: verdana; font-size: 13.86px;"><span style="font-size: medium;"><b>C. </b></span></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="background-color: transparent; color: #336699; font-family: verdana; font-size: large; text-decoration-line: none;"><span style="color: #1a222a; font-size: medium;">➡</span></a> <b style="font-size: 18px;"><i><a href="https://neoskeptics.blogspot.com/2023/12/basic-ai-faqs.html"><span style="font-family: verdana;">Dozen Basic AI FAQs</span></a></i></b></div><div style="color: #1a222a;"><span style="font-family: verdana;"><span style="font-size: medium;">This page contains links to responses by Google's Bard chatbot running Gemini Pro to 12 questions that should be asked more frequently, but aren't. As consequence, too many readily understood AI terms have become meaningless buzzwords in the media.</span></span></div></div></div></div>
Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-91099837152053899312024-01-14T21:22:00.426-05:002024-03-27T21:39:37.245-04:00Massive copyright violations and a ChatGPT app store ... TL;DR and podcast 14Jan24 <div style="text-align: right;"><span style="font-family: verdana; font-size: x-small;"><i>Last update: Tuesday 1/16/24 </i></span></div><div style="text-align: left;"><span style="font-family: verdana;"><div class="separator" style="clear: both; text-align: center;"><div style="text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><span style="font-size: medium;"><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiGSCKliE-ORgvughSga8q8mkV1I2bHPtrxoZ9SyquCFpIhUGgp0yWh_EWcBcRu62VDwAhXGS_gWjx-pbuR2U2XW8QdPEgtvHaRluXtk8YGsIdcT3KxTzOkloFlh8homEByjhexN6C8CsVxCHzm6zY9xz027n3bbI1LZ9tZ5zFy3nFtMNLkzf6u3-yXDz8/s1400/NYTimesBuildingInNYC.jpeg" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="932" data-original-width="1400" height="213" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiGSCKliE-ORgvughSga8q8mkV1I2bHPtrxoZ9SyquCFpIhUGgp0yWh_EWcBcRu62VDwAhXGS_gWjx-pbuR2U2XW8QdPEgtvHaRluXtk8YGsIdcT3KxTzOkloFlh8homEByjhexN6C8CsVxCHzm6zY9xz027n3bbI1LZ9tZ5zFy3nFtMNLkzf6u3-yXDz8/s320/NYTimesBuildingInNYC.jpeg" vspace="10" width="320" /></a></div>Welcome to our </span></span></span></span><span style="font-size: medium;"><span><span><span style="background-color: white; color: #1a222a;">14Jan24</span><span style="background-color: white; color: #1a222a;"> </span><span style="background-color: white; color: #1a222a;">TL;DR summary + podcast about the top two stories </span></span></span><span style="background-color: white; color: #1a222a;">that were posted on our "</span><b style="background-color: white; color: #1a222a;"><i><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html" style="color: #336699; text-decoration-line: none;">Useful AI News</a></i></b><span style="background-color: white; color: #1a222a;">" page during the first two weeks of 2024:</span></span></div><div style="text-align: left;"><span style="font-size: medium;"><span><span style="background-color: white; color: #1a222a;"><br /></span></span></span></div><div style="text-align: left;"><span style="font-size: medium;"><span style="background-color: white; color: #1a222a;">1) T<span style="font-size: medium;">he massive copyright violation lawsuit filed by the New York Times against OpenAi and Microsoft, and</span> (2) </span></span><span style="font-size: medium;"><span style="background-color: white; color: #1a222a; font-family: verdana;">OpenAI's strategy to convert ChatGPT</span><span style="background-color: white; color: #1a222a; font-family: verdana;"> into a revenue generating platform.</span></span><span style="background-color: white; color: #1a222a; font-size: 13.86px;">.</span></div><div style="background-color: white; text-align: right;"><div class="separator" style="clear: both; text-align: left;"><div style="text-align: right;"><div><span style="font-size: medium;"><span style="color: #1a222a;">Click link for podcast (opens in new tab)</span></span> </div><div><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-size: large; text-align: left; text-decoration-line: none;"><span style="color: #1a222a; font-size: medium;">➡</span></a> <a href="https://drive.google.com/file/d/1LNpm_pms1mojZ9GbBvUiuWe_sjN5xui7/view?usp=sharing" target="_blank"><i><span style="color: #cc0000; font-size: medium;"><b>Massive copyright violations + ChatGPT app store</b></span></i></a></div></div></div><div class="separator" style="clear: both; color: #1a222a; text-align: left;"><div style="text-align: right;"><span style="color: black;"><span style="color: #1a222a;">Click the "start" button the when the podcast page is loaded <br />... </span></span><span style="color: #1a222a;"> </span><span style="color: #1a222a;">If audio fails to start, or gets stuck, try reloading that page</span></div><span face="Verdana, Geneva, sans-serif" style="font-size: medium;">TL;DR link<b> </b></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-size: large; text-decoration-line: none;"><span style="color: #1a222a;">➡</span></a><b style="color: black; font-family: Verdana, Geneva, sans-serif;"><span style="font-size: medium;"><span style="color: #1a222a;"> </span><i><a href="https://neoskeptics.blogspot.com/2024/01/massive-copyright-violations-and.html#more"><span style="color: #cc0000;">HERE</span></a></i></span></b></div><div class="separator" style="clear: both; color: #1a222a; text-align: left;"><b style="color: black; font-family: Verdana, Geneva, sans-serif;"><span style="font-size: medium;"><br /></span></b></div></div></div></span></div><span><span style="font-family: verdana;"><a name='more'></a></span></span><div style="text-align: left;"><span style="background-color: white; color: #1a222a;"><span style="font-family: verdana;"><b style="font-size: large;">A.</b></span></span><b style="background-color: white; font-family: verdana; font-size: large;">TL;DR ... </b><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;"> Top 2 stories in past 2 weeks ...</b></div><div style="text-align: left;"><span style="background-color: white; color: #1a222a;"><span style="font-family: verdana;">Last year ended with a bang, tremors, and aftershocks; so it should come as no surprise that the new year has begun with further consequences of the raucous ending of the old. </span></span></div><div style="text-align: left;"><span style="background-color: white; color: #1a222a;"><span style="font-family: verdana;"><br /></span></span></div><div style="text-align: left;"><b><span style="font-family: verdana; font-size: medium;">1) </span></b><span style="background-color: white; color: #1a222a;"><span style="font-family: verdana; font-size: medium;"><b>New York Times lawsuit(s) against OpenAi and Microsoft</b></span></span></div><div style="text-align: left;"><span style="font-family: verdana;">A few days after Christmas, the New York Times filed lawsuits against OpenAI and Microsoft, lawsuits that alleged that their use of millions of articles published in the Times to train the GPT large language models without the Times' permission constituted a massive violation of the Times' copyrights. According to the Wall Street Journal:</span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">"The Times is seeking damages, in addition to asking the court to stop the tech companies from using its content and to destroy data sets that include the Times’ work ... In its suit, the Times said the fair-use argument shouldn’t apply because the AI tools can serve up, almost verbatim, large chunks of text from Times news articles."</span></li></ul><div><span style="font-family: verdana;">Although other publishers have filed similar claims, the Associated Press and Axel Springer have already worked out payment arrangements with OpenAI. If the Times were to win its suit, OpenAi and Microsoft would have to make substantial payments, perhaps billions of dollars. </span></div><div><span style="font-family: verdana;"><br /></span></div><div><span style="font-family: verdana;">Given the high stakes involved, this case will take years before a final decision is rendered by the courts, probably by the U.S. Supreme Court. Meanwhile, these same high stakes might provide the Times with sufficient leverage to "persuade" OpenAi and Microsoft to reach a much lower, but mutually acceptable out-of-court settlement.</span></div><div><span style="font-family: verdana;"><br /></span></div><div><span style="font-family: verdana;">On the other hand, as per one of the related articles, as a result the immediate success of its $13 billion investment in OpenAI, Microsoft has now passed Apple to become the most valuable public company in the world. To be specific, Microsoft's market value has exceeded $3 trillion. As such, it could probably afford to make multi-billion dollar payments to the Times. But if the courts compelled Microsoft to make multi-billion dollar payments, Microsoft's Big Tech competitors would also be compelled to make similar payments ... which some of them might not be able to afford.</span></div></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><div><span style="font-family: verdana;">Another related link is a podcast featuring an interview in which an editor of The Verge interviewed Dana Rao, the general counsel for Adobe. Among other things, the counsel discusses Adobe's decision to negotiate fees to the owners of the visual materials Adobe used to train its AI image generator called "Firefly" before it trained the generator. The courts might consider Adobe's strategy as evidence that OpenAI and Microsoft might have been able to produce their own image and text generators by following a similar strategy of negotiating payments before they started training their models</span></div><div><span style="font-family: verdana;"><br /></span></div><div><span style="font-family: verdana;">Finally, our previous TL;DR and podcast -- <b><i><a href="https://neoskeptics.blogspot.com/2024/01/happy-new-year-top-ai-stories-of-2023_9.html#prediction"><span>Top AI stories of 2023</span></a></i></b> -- predicted that "the rapid and widespread proliferation of small language models (SLMs) will be the dominant AI news stories throughout 2024" because a growing body of research has shown that small models have surprising power. Had OpenAI and Microsoft anticipated this finding, they might have produced language modules as powerful as their GPT series using a fraction of the training materials. This would have entailed a fraction of the copyrights involved ... and more affordable court ordered payments to the NY Times and other data owners.</span></div><div><span style="font-family: verdana;"><br /></span></div></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><b><span style="font-family: verdana; font-size: medium;">2) Conversion of ChatGPT into a revenue generating platform</span></b></div><div style="text-align: left;"><span style="font-family: verdana;">Last summer, OpenAI announced its intention to open an app store for extension to ChatGPT. It affirmed this intention at its developer conference in December. And it made its official announcement on Wednesday January 10th. As per our headline story in the NY Times, OpenAI "was opening an app store for people to share customized versions of its popular chatbot, ChatGPT, as the artificial intelligence company works to expand the reach of its flagship technology and turn it into a moneymaker</span><span style="font-family: verdana;">". </span><span style="font-family: verdana;">In other words, OpenAI wants to turn ChatGPT into a </span><span style="font-family: verdana;">platform.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">In so doing, OpenAI hopes to follow the same path to greater financial success that has been followed by Facebook, YouTube, TikTok, Google's Play Store, and Apple's App Store, the same path that Elon Musk wants Twitter to follow by his conversion of Twitter to "X". </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">If you develop a great app, it will attract many users; but if you convert your app into a platform, lots of other developers and other creators will use their ingenuity to create lots of new content for your platform, e.g., apps and extensions. This new content will attract far larger numbers of users and thereby generate far more income from new subscription fees and/or advertising fees ... fees from which you can extract a very substantial share.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Why is OpenAI launching its app store at this time? On the one hand, it cost OpenAI (and Microsoft) billions of dollars to develop and more billions to operate its GPT models. On the other hand, most ChatGPT users are still using the free GPT-3.5 model. So OpenAI needs more revenue. As a first step, OpenAI will only allow access to its app store to its "plus" users, i.e., users who pay the $20 per month subscription fee for access to GPT-4.</span></div><div style="text-align: left;"><div><div id="Overview"><span style="font-family: verdana;"><span><span><b style="font-size: large;"><br class="Apple-interchange-newline" />B. </b></span></span><b style="background-color: white; color: #1a222a; font-size: large;">Top stories ...</b></span></div><div id="Overview" style="background-color: white; color: #1a222a; font-size: 13.86px;"><ol><li style="margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;"><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#LLMNews" style="color: #336699; font-size: small; font-weight: 700; text-align: center; text-decoration-line: none;"><span style="color: #cc0000;">LLM News</span></a><br /><span style="font-size: medium;">"The New York Times sued Microsoft and OpenAI for alleged copyright infringement, touching off a legal fight over generative-AI technologies, with implications for the future of the news business", Alexandra Bruell, <b><i><a href="https://t.co/oJcTzpzXLq" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Wall Street Journal</span></a></i></b>, 12/27/23 <b><span style="color: #cc0000;">***</span></b> <br />-- <i><b><span style="font-style: normal; font-weight: 400;">This story also covered by</span><span style="font-style: normal; font-weight: 400;"> </span></b></i><a href="https://neoskeptics.blogspot.com/p/--%20%20This%20story%20also%20covered%20by" style="color: #336699; font-style: italic; font-weight: bold; text-decoration-line: none;"><span style="color: #cc0000;">The Verge</span></a><b style="font-style: italic;">, </b><a href="https://t.co/1EnIwwbdVW" style="color: #336699; font-style: italic; font-weight: bold; text-decoration-line: none;"><span style="color: #cc0000;">BBC</span></a><b style="font-style: italic;">, </b><a href="https://t.co/TvlkbRmiKj" style="color: #336699; font-style: italic; font-weight: bold; text-decoration-line: none;"><span style="color: #cc0000;">Bloomberg</span></a><b style="font-style: italic;">, </b><a href="https://t.co/ondZISqf6q" style="color: #336699; font-style: italic; font-weight: bold; text-decoration-line: none;"><span style="color: #cc0000;">Reuters</span></a><b style="font-style: italic;">, </b><a href="https://www.npr.org/2023/12/27/1221821750/new-york-times-sues-chatgpt-openai-microsoft-for-copyright-infringement" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;"><b><i>NPR</i></b></span></a><b style="font-style: italic;">, <a href="https://t.co/Ywn8TDT4Dp" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Engadget</span></a>, <a href="https://neoskeptics.blogspot.com/p/TechCrunch," style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">TechCrunch</span></a> ... </b>and<b style="font-style: italic;"> </b><a href="https://www.nytimes.com/2023/12/27/business/media/new-york-times-open-ai-microsoft-lawsuit.html" style="color: #336699; text-decoration-line: none;"><b><i><span style="color: #cc0000;">NY Times</span></i></b></a><br />-- Here's a related story <b>➡ </b><span style="color: black;">"Inside the News Industry’s Uneasy Negotiations With OpenAI", Benjamin Mullin, </span><b style="color: black;"><i><a href="https://www.nytimes.com/2023/12/29/business/media/media-openai-chatgpt.html?smid=nytcore-android-share"><span style="color: #cc0000;">NY Times</span></a></i></b><span style="color: black;">, 12/29/23</span><br style="color: black;" /><span style="color: black;">-- Here's another related story </span><b>➡ </b><span style="color: black;">"Microsoft Tops Apple to Become Most Valuable Public Company", Tripp Mickle and Karen Weise, </span><b style="color: black;"><i><a href="https://www.nytimes.com/2024/01/12/technology/microsoft-apple-most-valuable-company.html?smid=nytcore-android-share"><span style="color: #cc0000;">NY Times</span></a></i></b><span style="color: black;">, 1/13/24</span><br style="color: black;" /><span style="color: black;">-- And a 3rd related story (podcast) </span><b>➡ </b><span style="color: black;"><span style="color: #1a222a;">"How Adobe is managing the AI copyright dilemma, with general counsel Dana Rao", Nilay Patel, </span><b><i><a href="https://www.theverge.com/24027198/adobe-dana-rao-ai-copyright-fair-use-figma-acquisition-deal-decoder-interview"><span style="color: #cc0000;">The Verge</span></a></i></b><span style="color: #1a222a;">, 1/9/24</span></span><br /><br /></span></span></li><li style="margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;"><b style="font-size: small; text-align: center;"><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#OpenAI" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">OpenAI</span></a></b><br />"<span style="font-size: medium;">OpenAI Unveils App Store for Custom Chatbots", Cade Metz, <b><i><a href="https://www.nytimes.com/2024/01/10/technology/openai-app-store-chatgpt.html?smid=nytcore-android-share" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">NY Times</span></a></i></b>, 1/8/24<b><span style="color: #cc0000;"> ***</span></b><br />-- This story also covered by </span><span style="font-size: medium; font-style: italic; font-weight: bold;"><a href="https://techcrunch.com/2024/01/10/openai-launches-a-store-for-custom-ai-powered-chatbots/" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">TechCrunch</span></a>, <a href="https://www.bloomberg.com/news/articles/2024-01-10/openai-launches-gpt-store-for-users-to-share-custom-chatbots?utm_medium=social&utm_content=tech&cmpid%3D=socialflow-twitter-tech&utm_source=twitter&utm_campaign=socialflow-organic" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Bloomberg</span></a>, <a href="https://www.theverge.com/2024/1/10/24032144/openai-chatgpt-gpt-store-ai-launch" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">The Verge</span></a>, <br /></span><span style="font-size: medium;">-- The following commentary focused on the app store as OpenAI's strategy for transforming ChatGPT into a lucrative platform, like Facebook, because ChatGPT is not earning enough revenue from subscriptions <b>➡ </b>"ChatGPT’s FarmVille Moment", David Karpf, <b><i><a href="https://www.theatlantic.com/technology/archive/2024/01/openai-gpt-store-farmville/677115/" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">TheAtlantic</span></a></i></b>, 1/12/24<br /><br /></span></span></li></ol></div></div><div><div style="color: #1a222a;"><span style="font-family: verdana; font-size: 13.86px;"><span style="font-size: medium;"><b><br /></b></span></span></div><div style="color: #1a222a;"><span style="font-family: verdana;"><span style="font-size: 13.86px;"><span style="font-size: medium;"><b>C. </b></span></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-size: large; text-decoration-line: none;"><span style="color: #1a222a; font-size: medium;">➡</span></a> <b style="font-size: 18px;"><i><a href="https://neoskeptics.blogspot.com/2023/12/basic-ai-faqs.html"><span>Dozen Basic AI FAQs</span></a></i></b></span></div><div style="color: #1a222a;"><span style="font-family: verdana;"><span style="font-size: medium;">This page contains links to responses by Google's Bard chatbot running Gemini Pro to 12 questions that should be asked more frequently, but aren't. As consequence, too many readily understood AI terms have become meaningless buzzwords in the media.</span></span></div></div></div>
Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-26551350157105587872024-01-09T13:03:00.040-05:002024-03-27T21:39:52.272-04:00Happy New Year!!! Top AI stories of 2023 ... TL;DR and podcast 9Jan24 <div style="text-align: right;"><span style="font-family: verdana; font-size: x-small;"><i>Last update: <b><span style="color: red;">Friday 1/12/24 </span></b></i></span></div><div style="text-align: left;"><span style="font-family: verdana;"><div class="separator" style="clear: both; text-align: center;"><div style="text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><span style="font-size: medium;"><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj0pryiWnKCzA7bzN99Ptp7cbWB07muOxOfVVsis0mad2BD5A3go4WGZaDAO5_7V-QbYV_kdDybBxSk49_x8nz4E-FQcFkL_6UkLz664w6j9RSW_3sNVS1xY-eL3YKmZpRleBF6TceXwHtStF2odQAEfa2LvqEy004F0mfl9igPhx5oeeJdTGWBtZ7FSkw/s2048/HappyNewYear2024.jpg" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="2048" data-original-width="2048" height="320" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj0pryiWnKCzA7bzN99Ptp7cbWB07muOxOfVVsis0mad2BD5A3go4WGZaDAO5_7V-QbYV_kdDybBxSk49_x8nz4E-FQcFkL_6UkLz664w6j9RSW_3sNVS1xY-eL3YKmZpRleBF6TceXwHtStF2odQAEfa2LvqEy004F0mfl9igPhx5oeeJdTGWBtZ7FSkw/s320/HappyNewYear2024.jpg" vspace="10" width="320" /></a></div>Happy New Year!!! ... And welcome to our </span></span></span></span><span style="font-size: medium;"><span><span style="background-color: white; color: #1a222a;">TL;DR summary + podcast about the top AI stories that were posted on our "</span><b style="color: #1a222a;"><i><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html">Useful AI News</a></i></b><span style="background-color: white; color: #1a222a;">" page in 2023. We selected the stories that we thought will prove to be the most useful for predicting what the top AI stories will be in 2024.</span></span></span></div><div style="text-align: left;"><span style="font-size: medium;"><span><span style="background-color: white; color: #1a222a;"><br /></span></span></span></div><div style="text-align: left;"><span style="font-size: medium;"><span><span><span style="background-color: white; color: #1a222a;">Our first top story is OpenAI's movement of ChatGPT to its new GPT-4 model in March 2023, a move that led to ChatGPT's domination of the AI news sector for the rest of the year. Our second top story is about the most significant Big Tech challenges to OpenAI's supremacy in the final weeks of 2023, i.e., challenges from Amazon, Google, and Microsoft ... yes, Microsoft. <br /><br />These top stories framed our prediction of the top stories for 2024. Spoiler alert, here's our prediction: </span></span></span><span style="background-color: white; color: #1a222a;">Whereas large language models (LLMs) dominated AI news throughout 2023, the rapid and widespread proliferation of small language models (SLMs) will be the dominant AI news stories throughout 2024</span><span style="background-color: white; color: #1a222a;">.</span><span style="background-color: white; color: #1a222a;"> </span></span></div><div style="text-align: left;"><span style="font-size: medium;"><span><span style="background-color: white; color: #1a222a;"><br /></span></span></span></div><div style="text-align: left;"><span><span style="background-color: white; color: #1a222a;"><span>Click the "start" button in the audio control (below) to hear the podcast ... </span></span></span><span style="background-color: white; color: #1a222a;"> </span><span style="background-color: white; color: #1a222a;">If audio fails to start, or gets stuck, try reloading the page</span><span style="background-color: white; color: #1a222a; font-size: 13.86px;">.</span></div><div style="background-color: white; color: #1a222a; font-family: Verdana, Geneva, sans-serif; text-align: left;"><span style="font-family: verdana; font-size: medium;"></span></div><div style="background-color: white; font-family: Verdana, Geneva, sans-serif; text-align: right;"><span style="color: #1a222a; font-family: verdana; font-size: medium;">
<iframe allow="autoplay" height="50" src="https://drive.google.com/file/d/1LAt0O_zq6e1z2tQkIQc0SruG23NZ4JWp/preview?usp=sharing" width="400"></iframe></span><div class="separator" style="clear: both; color: #1a222a; text-align: left;"><div style="color: black; font-family: Times;"><span style="font-family: verdana;"><div class="separator" style="clear: both; text-align: center;"><div style="text-align: right;"><div class="separator" style="clear: both; color: #1a222a; text-align: left;"><span face="Verdana, Geneva, sans-serif" style="font-size: medium;">TL;DR link<b> </b></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-size: large; text-decoration-line: none;"><span style="color: #1a222a;">➡</span></a><b style="color: black; font-family: Verdana, Geneva, sans-serif;"><span style="font-size: medium;"><span style="color: #1a222a;"> </span><i><a href="https://neoskeptics.blogspot.com/search?updated-max=2024-01-14T21:22:00-05:00&max-results=5&start=5&by-date=false"><span style="color: #cc0000;">HERE</span></a></i></span></b></div><div><b style="color: black; font-family: Verdana, Geneva, sans-serif;"><span style="font-size: medium;"><br /></span></b></div></div></div></span></div><span style="color: black; font-family: Times;"><span style="font-family: verdana;"></span></span></div></div></div></span></div><span><span style="font-family: verdana;"><a name='more'></a></span></span><div style="text-align: left;"><div id="Overview"><span><span style="font-family: verdana;"><b style="font-size: large;">A. </b></span></span><b style="background-color: white; font-family: verdana; font-size: large;">TL;DR ... </b><b style="font-family: verdana; font-size: large;">Top AI stories in 2023</b></div><div id="Overview"><span><span style="font-family: verdana;"><b style="font-size: large;"><div id="Overview" style="font-family: Times; font-size: medium; font-weight: 400; text-align: center;"><br /></div></b></span></span></div><div id="Overview"><span><span style="font-family: verdana;"><b><span style="font-size: medium;">1. </span></b></span></span><span><span style="background-color: white; font-family: verdana;"><b style="font-size: large;">OpenAI releases ChatGPT's new GPT-4 model</b><br />When OpenAi released ChatGPT running on the GPT-3.5 model in November 2022, it was so impressive that one hundred million users subscribed to its services within three months. </span></span></div><div id="Overview"><br /></div><div id="Overview"><span><span style="background-color: white; font-family: verdana;">When OpenAI introduced its new GPT-4 model in March 2023, it quickly blew GPT- 3.5 out of the water because its cognitive skills were far more powerful. </span></span><span style="background-color: white; font-family: verdana;">We were then told that GPT-4 was far more powerful than GPT-3.5 because it had been trained on far more data. Bigger was better.</span></div><div id="Overview"><span style="background-color: white; font-family: verdana;"><br /></span></div><div id="Overview"><span style="font-family: verdana;">Unfortunately, this "explanation" didn't really explain the new cognitive skills that users of these models have found to be most impressive and/or the most useful, e.g., the models' ability to answer complex questions, solve problems, write poems about a suggested theme in a specified literary style, crack jokes in the style of well known comedians, debug computer code, summarize documents, etc, etc, etc.</span></div><div id="Overview"><span style="font-family: verdana;"><br /></span></div><div id="Overview"><span style="font-family: verdana;">Fortunately, many AI experts involved in the creation of these and other large language models frankly admitted that the training of these models was not designed to imbue the models with these higher level skills. Their sudden unexpected emergence is still a mystery. Note that "Emergent" skills are skills that a model somehow acquires, but was not trained to acquire</span></div><div id="Overview"><ul style="text-align: left;"><li><span style="font-family: verdana;">"From GPT-1 to GPT-4: A Comprehensive Analysis and Comparison of OpenAI’s Evolving Language Models", Dhanshree Shripad Shenwai, <b><i><a href="https://www.marktechpost.com/2023/07/05/from-gpt-1-to-gpt-4-a-comprehensive-analysis-and-comparison-of-openais-evolving-language-models/">Marktechpost</a></i></b>,, 7/5/23</span></li></ul></div><div id="Overview"><br /></div><div id="Overview"><span><span style="background-color: white; font-family: verdana;"><br /><b style="font-size: large;">2. Big Tech's separate challenges to OpenAI's dominance<br /></b><br /><b>2a)</b> </span><span style="background-color: white; font-family: verdana;"><b>Microsoft's most important announcements at its Ignite 2023 conference for developers and IT professionals (Nov 14-17, Seattle, WA)</b></span><b style="background-color: white; font-family: verdana; font-size: large;"><span style="color: #cc0000;"><br /></span></b></span></div><div id="Overview"><span><span style="background-color: white; font-family: verdana;"><div style="color: #1a222a; font-family: Verdana, Geneva, sans-serif;"><ul style="line-height: 1.4; list-style-image: initial; list-style-position: initial; margin: 0.5em 0px; padding: 0px 2.5em;"><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;">Rather than use ChatGPT as the orimary user interface for its LLM, Microsoft confirmed that it would provide a customized interface, called a "Copilot", for each of its office productivity apps. The copilots would help their users to use each productivity app more effectively. So there would be a copilot for Word, for Excel, for Teams, for GitHub, etc. Enterprise customers who subscribed to Microsoft's office apps would need an additional subscription for each copilot<br /></span></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;">It renamed "Bing Chat" to "Microsoft Copilot" and subsequently announced that Microsoft Copilot running GPT-4 would be free, whereas OpenAI charged its "plus" users a $20 per month subscription fee.</span></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;">It announced its intention to build its own AI chips rather than buy AI chips from Nvidia that were in short supply</span></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;">It enabled Microsoft's Azure cloud to support open source LLMs, e.g., Facebook's LLaMa</span></li></ul><div><span style="font-family: verdana;">These initiatives enhanced Microsoft's strategic independence from OpenA's LLMs and Nvidia's chips, and offered potential opportunities for Microsoft to earn fees from open source LLMs that used its Azure cloud. </span></div><div><ul style="text-align: left;"><li><span style="font-family: verdana;">"The Inside Story of Microsoft’s Partnership with OpenAI", Charles Duhigg, </span><b style="font-family: verdana;"><i><a href="https://www.newyorker.com/magazine/2023/12/11/the-inside-story-of-microsofts-partnership-with-openai" style="color: #336699; text-decoration-line: none;">The New Yorker</a></i></b><span style="font-family: verdana;">, 12/1/22</span></li></ul></div></div></span></span></div><div id="Overview"><br /></div><div id="Overview"><span><b><span style="background-color: white; font-family: verdana;">2b)</span><span style="background-color: white; font-family: verdana;"> </span></b><b style="background-color: white; font-family: verdana;">The "AI Big Bang" </b><span style="background-color: white; font-family: verdana;">a/k/a The fall and rise of Sam Altman .</span></span><span style="background-color: white; font-family: verdana;">.. 11/17/23 to 11/21/23 <br /><div style="text-align: center;">... Note that this event is only included as a top AI story of 2023 because of its potential influence on Microsoft's actions in 2024 ....</div></span></div><div id="Overview"><span style="background-color: white;"><span><p style="font-family: verdana; font-size: large;"></p><ul style="font-family: verdana; text-align: left;"><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="color: #1a222a;">The tumultuous long weekend began on Friday evening when the OpenAI board abruptly notified Sam Altman that he was fired. In its dismissal notice the OpenAI board said that it no longer trusted Sam Altman, but it did not say why. However, the </span><b><i><a href="https://www.washingtonpost.com/technology/2023/11/22/sam-altman-fired-y-combinator-paul-graham/" style="text-decoration-line: none;"><span style="color: #2b00fe;">Washington Post</span></a><span style="color: #1a222a;"> </span></i></b><span style="color: #1a222a;">reported that<br /></span><span style="color: #1a222a;"><br /></span>"Four years ago, one of Altman’s mentors, Y Combinator founder Paul Graham, flew from the United Kingdom to San Francisco to give his protégé the boot, according to three people familiar with the incident, which has not been previously reported."<br /><br />"Graham had surprised the tech world in 2014 by tapping Altman, then in his 20s, to lead the vaunted Silicon Valley incubator. Five years later, he flew across the Atlantic with concerns that the company’s president put his own interests ahead of the organization — worries that would be echoed by OpenAI’s board."<br /><br /></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="color: #1a222a;">On Tuesday evening, Altman returned to OpenAI as CEO; three out of the four members of the board were dismissed; and two new outsiders were appointed to the board. </span></li></ul><div style="text-align: left;"><span style="text-align: center;"><span style="font-family: verdana;">Microsoft's CEO Satya Nadella's confident leadership won this "battle" handily, especially his </span></span><span style="font-family: verdana; text-align: center;">reassuring </span><span style="font-family: verdana; text-align: center;">assertions that Microsoft's $13 billion investment in OpenAI gave it full access to OpenAI's revolutionary LLM technology, regardless of what happened to Sam Altman or to OpenAI. Indeed, Microsoft's stock hit an all-time high after Altman's return was announced. Of course Sam Altman won, but the three board members who were dismissed lost, and Open AI lost. OpenAI lost because it failed to fulfill its non-profit mission. OpenAI is now a de facto profit-oriented operation beholden to Microsoft and other investors.</span></div><div style="text-align: left;"><span style="text-align: center;"><span style="font-family: verdana;"><br /></span></span></div></span></span></div><div id="Overview"><span style="background-color: white; font-family: verdana;">2c)<b> </b></span><span style="font-family: verdana; font-weight: bold;">Amazon's "re:Invent 2023" customer conference (Nov 27-Dec 1, Las Vegas, NV) </span></div><div id="Overview"><span style="font-family: verdana;">Amazon played </span><span style="font-family: verdana;">catchup with regards to the deployment of generative AI technology; more specifically, Amazon played catchup to Microsoft. </span></div><div id="Overview"><span style="font-family: verdana;"><br /></span></div><div id="Overview"><span style="font-family: verdana;">Few were surprised by Amazon's announcements of an array of generative AI services that it would provide for enterprise customers of its AWS cloud services. </span><span style="font-family: verdana;">Its most memorable announcement were for its chatbot and its image generator:</span></div><div id="Overview"><ul style="text-align: left;"><li><span style="font-family: verdana;">Amazon's chatbot is called "Q"</span></li><li><span style="font-family: verdana;">Its image generator is called "Titan Image Generator"</span></li></ul></div><div id="Overview"><span style="font-family: verdana;">Note that "Titan" is the name of Amazon's LLM. Also note that Amazon had previously announced its intention to produce its own GPU chips called "Trainium". Amazon also announced its development of more powerful "Trainium" chips </span></div><div id="Overview"><b style="background-color: white; font-family: verdana;"><span style="font-size: medium;"><br /></span></b></div><div id="Overview"><span style="background-color: white; font-family: verdana; font-size: medium;"><b>2d)</b></span><span style="background-color: white; font-family: verdana;"> </span><b style="color: #1a222a; font-family: verdana;">Google announced "Gemini", its new language models, on 12/6/23</b></div><div style="background-color: white; color: #1a222a; font-family: Verdana, Geneva, sans-serif;"><span style="font-family: verdana;"><div>Gemini comes in three versions: Gemini <b>Nano</b>, Gemini <b>Pro</b>, and Gemini <b>Ultra</b></div><div><ul style="line-height: 1.4; list-style-image: initial; list-style-position: initial; margin: 0.5em 0px; padding: 0px 2.5em;"><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><b>Nano,</b> the lightest version, is a small language model (SLM) that will eventually run on all Android devices, but currently only runs on Google's Pixel smartphones. It will not require Internet connectivity<br /><br /></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><b>Pro</b> is an LLM, a more powerful model than Nano, that and currently runs Bard on all devices. Bard can be accessed at <a href="http://bard.google.com/" style="color: #336699; text-decoration-line: none;"><b><i>bard.google.com</i></b></a><br /><br /></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><b>Ultra</b> is the most powerful LLM, but Google has not yet announced a release date<br /></li></ul>At present Nano and Pro accept text input, then deliver text output, but all versions will eventually be <b>multimodal</b>, accepting input as text, images, video, audio, and code.</div><div><br />All models were trained on Google's own chips (Tensor Processing Units) and will run in data centers on those chips</div><div><br /><b><i><a href="https://deepmind.google/technologies/gemini/#capabilities" style="color: #336699; text-decoration-line: none;">Google's official announcement</a> </i></b>presented the results of 18 <b>benchmarks</b> tests that compared Gemini Ultra with GPT-4. Gemini performed better than GPT-4 by small margins on all but one test.<br /><ul style="line-height: 1.4; list-style-image: initial; list-style-position: initial; margin: 0.5em 0px; padding: 0px 2.5em;"><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><b>Stock market </b>... the most important benchmark???<br />"Alphabet soars as Wall Street cheers arrival of AI model Gemini", Aditya Soni, <b><i><a href="https://t.co/tKAFY9a7zp" style="color: #336699; text-decoration-line: none;">Reuters</a></i></b><span style="color: black;">, 12/7/23</span></li></ul></div></span></div><div id="Overview"><b style="background-color: white; font-family: verdana;"><span style="font-size: medium;"><br /></span></b></div><div id="Overview"><span style="background-color: white; font-family: verdana;"><b>2e) Microsoft announced Phi-2, a small language model (SLM) on 12/12/23</b><br /><div style="text-align: center;">... Note that this event is included as a top AI event of 2023 because of its potential influence on everybody's actions in 2024 ...</div></span></div><div id="Overview"><span style="background-color: white; font-family: verdana; font-size: medium;"><br /></span></div><div id="Overview"><span style="background-color: white; font-family: verdana;"><span style="background-color: transparent; color: #1a222a;">Microsoft published the results of a Phi-2 vs. Ultra test on its blog in a note titled: "</span><b style="background-color: transparent; color: #1a222a;"><i><a href="https://www.microsoft.com/en-us/research/blog/phi-2-the-surprising-power-of-small-language-models/" style="color: #336699; text-decoration-line: none;">The surprising power of small language models</a></i></b><span style="background-color: transparent; color: #1a222a;">". </span></span><span style="color: #1a222a; font-family: verdana;">Phi-2 is not a large language model (LLM); Phi-2 is a small language model (SLM). Whereas Gemini Ultra contains 1.56 trillion parameters, Phi-2 only contains 2.7 billion parameters; so Ultra is about 580 times as large as Phi-2.</span></div><div id="Overview"><br /></div><div id="Overview"><span style="background-color: white; font-family: verdana;"><span style="color: #1a222a;">Microsoft posed the kind of question to Phi-2 and to Gemini Ultra that students might encounter in an introduction to physics course, specifically: to calculate the speed of a skier when it reaches the bottom of a hill, given the skier's mass, the gravitational constant, and the height of the hill. Phi-2 calculated the correct answer, but Ultra's answer was wrong. Adding insult to injury, Microsoft asked Phi-2 to identify the logical error in Ultra's calculations ... which it did. </span></span><span style="background-color: white; color: #1a222a; font-family: verdana;">How could this happen? </span><span style="background-color: white; color: #1a222a; font-family: verdana;">The answer implied by Microsoft's blog note is plausible, </span></div><div id="Overview"><span style="background-color: white; font-family: verdana;"><blockquote style="color: #1a222a;"><div id="Overview">"The massive increase in the size of language models to hundreds of billions of parameters has unlocked a host of emerging capabilities that have redefined the landscape of natural language processing. A question remains whether such emergent abilities can be achieved at a smaller scale using strategic choices for training, e.g., data selection ...</div></blockquote><blockquote style="color: #1a222a;"><div id="Overview">... training data quality plays a critical role in model performance. This has been known for decades, but we take this insight to its extreme by focusing on “textbook-quality” data, following upon our prior work “<b><i><a href="https://www.microsoft.com/en-us/research/publication/textbooks-are-all-you-need/" style="color: #336699; text-decoration-line: none;">Textbooks Are All You Need.</a></i></b>” </div></blockquote></span></div><div id="Overview"><span style="background-color: white; font-family: verdana;"><span style="color: #1a222a;">By spending a couple of hours searching the Internet for other notes about small language models, the editor of this blog discovered that many other researchers in the AI community had published similar findings about the surprisng power of SLMs. Contrary to the conventional wisdom that dominated most discussions of language models throughout 2023, bigger is not necessarily better.</span></span></div><div id="Overview"><ul style="text-align: left;"><li style="background-color: white; font-family: verdana;">"The Rise of Small Language Models— Efficient & Customizable", Bijit Ghosh, <a href="https://medium.com/@bijit211987/the-rise-of-small-language-models-efficient-customizable-cb48ddee2aad#id_token=eyJhbGciOiJSUzI1NiIsImtpZCI6IjFmNDBmMGE4ZWYzZDg4MDk3OGRjODJmMjVjM2VjMzE3YzZhNWI3ODEiLCJ0eXAiOiJKV1QifQ.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.CrodzVbvpWsfgDphqhte8dTLLam8G6UsHB4lv952wXzNFxNyI9bjPU0nCzfwxf-mtQyd02wN3bu4LEoUPFYsCm81w4jt_Q8ExgYSxIqAgQ5rm9zVANdSeHlQe2M9IF8_wiTSoCLQzBXX5JBf69RwLcBvPEmAHDKV9Xd6EyToYoeC9AWOEvdrl_v19hRhcybl4wYshK-i2T0Rv9CERLjk7PKF1YLRO_7w7Ji3QjGXh5wsYsunP-i3Fz61y5of1_aFcm0ECgZA0cmvJPg8giopftA2c04asPqZRRwVIOSJyU5gNuF7aXsHjLVa5z-yC6cSYIDoKdj6vET73XSkNd6hlQ"><b><i>Medium</i></b></a>, 11/26/23</li><li style="background-color: white; font-family: verdana;">"9 Best Small Language Models Released in 2023", Sandhra Jayan, <b><i><a href="https://analyticsindiamag.com/9-best-small-language-models-released-in-2023/">AIM</a></i></b>, 12/7/23</li><li style="background-color: white; font-family: verdana;">"Small language models an emerging GenAI force", Antone Gonsalves, <b><i><a href="https://www.techtarget.com/searchenterpriseai/news/366563445/Small-language-models-an-emerging-GenAI-force">TechTarget</a></i></b>, 12/15/23</li><li style="background-color: white; font-family: verdana;">"Everything You Need to Know about Small Language Models (SLM) and its Applications", Tanya Malhotra, <b><i><a href="https://www.marktechpost.com/2023/12/05/everything-you-need-to-know-about-small-language-models-slm-and-its-applications/">MarTechPost</a></i></b>, 12/5/23</li><li style="background-color: white; font-family: verdana;">"7 Steps to Running a Small Language Model on a Local CPU", Aryan Garg, <b><i><a href="https://www.kdnuggets.com/7-steps-to-running-a-small-language-model-on-a-local-cpu">KDnuggets</a></i></b>, 11/14/23</li></ul></div><div id="Overview"><span style="background-color: white; font-family: verdana;"><div id="Overview"><span style="font-family: verdana;"><b><span style="font-size: medium;"><br /><br /></span></b></span></div><div id="Overview"><span style="font-family: verdana;"><b><span id="prediction" style="font-size: medium;">B. Our prediction for top AI stories in 2024 ...</span></b></span></div><div id="Overview" style="text-align: left;"><b style="color: #1a222a;"><br />Whereas large language models (LLMs), especially ChatGPT's GPT-4, dominated AI news in 2023, the rapid and widespread proliferation of small language models (SLMs) will be the dominant AI news stories throughout 2024</b><span style="color: #1a222a;">.</span><br /><span style="color: #1a222a;"><br />LLMs know everything about everything, so they currently have trillions of parameters. By contrast, SLMs are focused; their knowledge domains are limited; so they only need billions, not trillions, of parameters. Nevertheless, as noted in </span><b style="color: #1a222a;"><i><a href="https://www.microsoft.com/en-us/research/blog/phi-2-the-surprising-power-of-small-language-models/">Microsoft's report</a></i></b><span style="color: #1a222a;"> and elsewhere, SLMs have surprising power, sometimes comparable to LLMs.<br /></span><span style="color: #1a222a;"><br /></span>-- SLMs should cost a fraction of the cost of LLMs to develop and operate. Whereas LLMs cost billions to develop and operate, one would expect that SLMs could be developed and operated for hundreds of millions, perhaps tens of millions. <br /><br />-- SLMs can be trained on high quality private data which their developers license from the owners of the data, licenses that do not violate property rights or invade individual privacy.<br /><br />Therefore Big Tech's expensive LLMs will no longer dominate the language model space. Less expensive, more cost-effective SLM's will be developed by scores of smaller firms, e.g., Anthropic, firms that can afford to develop highly effective, but far less expensive smaller models<br /><br />With so many potential competitors, the shared development of open source platforms offers greater potential profitability to the contributors than if each developer pursued the separate development of its own closed platform. Indeed, open source development would be more likely to resolve the mystery of emergent properties.</div><div id="Overview" style="text-align: left;"><br /></div><div id="Overview" style="text-align: left;">None of the above implies that the Big Tech firms will not produce their own SLMs in 2024. Indeed, Google's Nano is an SLM. Nevertheless, Big Tech will face stiff competition from Smaller Tech. However, given its public confirmation of the surprising power of SLMs, it is surprising that Microsoft did not announce its own commercial SLM. Instead, the Microsoft report makes the following declarations:</div><blockquote><div id="Overview" style="text-align: left;">"With its compact size, Phi-2 is an ideal playground for researchers, including for exploration around mechanistic interpretability, safety improvements, or fine-tuning experimentation on a variety of tasks. We have made Phi-2(opens in new tab) available in the Azure AI Studio model catalog to foster research and development on language models."<br /></div></blockquote><p>Given that Microsoft, in effect, had outsourced a $13 billion contract to OpenAI to develop a commercial LLM for Microsoft, why didn't Microsoftt award another contract to OpenAI to develop a commercial SLM? Perhaps Microsoft is reluctant to extend its dependence on OpenAI at this time, given the stunning instability that OpenAI displayed during the "AI Big Bang" weekend just a few weeks ago. </p><p></p></span></div><div id="Overview"><span><span style="font-family: verdana;"><b style="font-size: large;"><br /></b></span></span></div><div id="Overview"><span><span style="font-family: verdana;"><b style="font-size: large;">C. Top stories of 2023 ...</b></span></span></div><div id="Overview"><ol><li><span style="background-color: white; font-family: verdana;"><span style="font-size: medium;"><b>OpenAI releases ChatGPT's new GPT-4 model</b><br />"GPT-4 has arrived. It will blow ChatGPT out of the water.", Drew Harwell and Nitasha Tiku, <b><i><a href="https://www.washingtonpost.com/technology/2023/03/14/gpt-4-has-arrived-it-will-blow-chatgpt-out-water/">Washington Post</a></i></b>, <b><span>3/14/23</span></b><br />-- This story also covered by<b><i><a href="https://www.nytimes.com/2023/03/14/technology/openai-gpt4-chatgpt.html"> NY Times</a></i></b>, <b><i><a href="https://www.theatlantic.com/technology/archive/2023/03/gpt4-release-rumors-hype-future-iterations/673396/">The Atlantic</a></i></b>, <b><i><a href="https://www.technologyreview.com/2023/03/14/1069823/gpt-4-is-bigger-and-better-chatgpt-openai/">MIT Tech Review,</a></i></b> <b><i><a href="https://techcrunch.com/2023/03/14/5-ways-gpt-4-outsmarts-chatgpt/">TechCrunch</a></i></b>, <b><i><a href="https://www.theverge.com/2023/3/14/23638033/openai-gpt-4-chatgpt-multimodal-deep-learning">The Verge,</a></i></b> <b><i><a href="https://www.forbes.com/sites/forbesbusinesscouncil/2023/03/28/gpt-4-released-what-it-means-for-the-future-of-your-business/?sh=6913428d2dc6">Forbes</a></i></b>, <b><i><a href="https://www.nature.com/articles/d41586-023-00816-5">Nature</a></i></b> ... and <b><i><a href="https://openai.com/research/gpt-4">OpenAI</a></i></b> <br /><br /></span></span></li><li><span style="font-family: verdana; font-size: medium;"><span style="background-color: white;"><span><span><b>Big Tech's most significant challenges to OpenAI and GPT-4 in the final weeks of 2023</b> ... Only our main headline stories are displayed, but links to related articles can be found on the TL;DR+ Podcast page for each week<br /><br />a) <span style="font-size: medium;"><b>Microsoft's Ignite 2023 conference for developers and IT professionals (Nov 14-17, Seattle, WA)</b></span><b style="font-size: medium;"><span style="color: #cc0000;"><br /></span></b><span style="font-size: medium;">-- Overview: "Microsoft Ignite 2023: Copilot AI expansions, custom chips and all the other announcements", </span><b style="font-size: medium;"><i><a href="https://t.co/8Rq4pok5eF"><span>TechCrunch</span></a></i></b><span style="font-size: medium;">, 11/15 & 16/23 ... This page contains TechCrunch summaries of </span><span style="font-size: medium;">every</span><span style="font-size: medium;"> announcement made by Microsoft. </span> <br />-- Other related stofies <b style="color: #1a222a; font-size: medium;">➡</b> <b><i><a href="https://neoskeptics.blogspot.com/2023/11/microsofts-dev-con-googles-graphcast.html">TL;DR + Podcast 19Nov23</a><br /></i></b><br />b) <b>The "AI Big Bang" a/k/a The fall and rise of Sam Altman<br /></b>-- "OpenAI Ousts CEO Sam Altman", Will Knight, <b><i><a href="https://t.co/s4RBLzVqjj" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Wired</span></a></i></b>, 11/17/23 <br /><span style="color: #1a222a;">-- "Sam Altman Is Reinstated as OpenAI’s Chief Executive", Cade Metz, Mike Isaac, Tripp Mickle, Karen Weise and Kevin Roose, <b><i><a href="https://www.nytimes.com/2023/11/22/technology/openai-sam-altman-returns.html?smid=nytcore-android-share" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">NY Times</span></a></i></b>, 11/22/23<br /></span>-- Other related stofies<b style="font-style: italic;"> </b><b style="color: #1a222a; font-size: medium; font-style: normal; font-weight: bold;">➡</b><a href="https://neoskeptics.blogspot.com/2023/11/the-ai-big-bang-aka-fall-and-rise-of.html" style="font-style: italic; font-weight: bold;">TL;DR and podcast 26Nov23</a><br /></span><br /><span>c)<b> </b></span></span></span><span style="background-color: white;"><span><b>Amazon's re:Invent 2023 conference (Nov 27-Dec 1, Las Vegas, NV) </b><br style="color: #1a222a; font-size: medium;" /><span style="color: #1a222a; font-size: medium;">-- Overview "Here’s everything Amazon Web Services announced at AWS re:Invent", Christine Hall, </span><b style="color: #1a222a; font-size: medium;"><i><a href="https://techcrunch.com/2023/11/29/heres-everything-aws-reinvent-2023-so-far/" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">TechCrunch</span></a></i></b><span style="color: #1a222a; font-size: medium;">, 11/29/23</span><br />-- Other related stofies<b style="font-style: italic;"> </b><b style="color: #1a222a; font-size: medium;">➡</b><a href="https://neoskeptics.blogspot.com/2023/12/new-tldr-podcast-3dec23-tldr-summary.html" style="font-style: italic; font-weight: bold;">TL;DR and Podcast 3Dec23</a><br /></span><br /><span face="Arial, Tahoma, Helvetica, FreeSans, sans-serif">d)</span></span><span><span style="background-color: white;"> </span><b style="background-color: white; color: #1a222a;">Google announced "Gemini"... its new LLM<br /></b>-- "Google DeepMind's Demis Hassabis Says Gemini Is a New Breed of AI", Will Knight, <i style="background-color: white; color: #1a222a; font-size: medium;"><b><a href="https://www.wired.com/story/google-deepmind-demis-hassabis-gemini-ai/" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Wired</span></a></b></i><span style="background-color: white; color: #1a222a; font-size: medium;">, 12/6/23 </span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; text-decoration-line: none;"><br />-- Other related stories<span style="background-color: white; color: black;"> </span><span style="background-color: white; color: #1a222a; font-size: medium;">➡</span><b style="font-style: italic;">TL;DR and podcast 10Dec23</b></a><br /><br />e) <span><b>Small Language Models (SLMs)<br /></b>-- "Microsoft releases Phi-2, a small language model AI that outperforms Llama 2, Mistral 7B", Carl Franzen, <b style="background-color: white; color: #1a222a; font-size: medium;"><i><a href="https://venturebeat.com/ai/microsoft-releases-phi-2-a-small-language-model-ai-that-outperforms-llama-2-mistral-7b/" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">VentureBeat</span></a></i></b><span style="background-color: white; color: #1a222a; font-size: medium;">, 12/12/23 ... <i>Note: Phi-2 also performed better than Gemini Ultra on a test of logical reasoning.</i></span><br />-- "The surprising power of Small Language Models (SLMs)", Mojan Javaheripi and Sébastien Bubeck, Microsoft Research, 12/12/23<br />-- Other related stofies<b style="background-color: white; font-style: italic;"> </b><b style="background-color: white; color: #1a222a; font-size: medium;">➡</b> <b><i><a href="https://neoskeptics.blogspot.com/2023/12/last-update-thursday-122123-welcome-to.html">TL;DR and podcast 21Dec23</a><br /></i></b></span></span></span></li></ol><span style="font-family: verdana;"><b><span style="font-size: medium;"><br /></span></b></span></div></div><div style="text-align: left;"><div style="background-color: white; color: #1a222a;"><span style="font-family: verdana; font-size: 13.86px;"><span style="font-size: medium;"><b>D. </b></span></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="background-color: transparent; color: #336699; font-family: verdana; font-size: large; text-decoration-line: none;"><span style="color: #1a222a; font-size: medium;">➡</span></a> <b style="font-size: 18px;"><i><a href="https://neoskeptics.blogspot.com/2023/12/basic-ai-faqs.html"><span style="font-family: verdana;">Dozen Basic AI FAQs</span></a></i></b></div><div style="background-color: white; color: #1a222a;"><span style="font-family: verdana;"><span style="font-size: medium;">This page contains links to responses by Google's Bard chatbot running Gemini Pro to 12 questions that should be asked more frequently, but aren't. As consequence, too many readily understood AI terms have become meaningless buzzwords in the media.</span></span></div></div>
Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-32939951584051290012023-12-27T18:47:00.198-05:002024-03-27T19:37:18.462-04:00Dozen Basic AI FAQs<div style="text-align: right;"><span style="font-family: verdana; font-size: x-small;"><i>Last update: <b><span style="color: red;">Friday 1/12/24 </span></b></i></span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><div class="separator" style="clear: both; text-align: center;"><span style="margin-left: 1em; margin-right: 1em;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgs8VmVjCD6P5d6dcelNJiqiI4xoFwBe2WIWivMKrIYMXHalIs4R15QiCo0OEUM2bRYO-GKjRX3rV52maaNIJiN3yE8ISWxzs7gl_GeF0fvM0VM3PxW15klxYJ60-g2gpXypBjMBeHLnuCDXJeCRz69gItoBDRO7KfC-dODFpsMEAx1eq6VqZZAJFveBpc/s1024/NetworkWithGrayNodes.png" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="1024" data-original-width="1024" height="320" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgs8VmVjCD6P5d6dcelNJiqiI4xoFwBe2WIWivMKrIYMXHalIs4R15QiCo0OEUM2bRYO-GKjRX3rV52maaNIJiN3yE8ISWxzs7gl_GeF0fvM0VM3PxW15klxYJ60-g2gpXypBjMBeHLnuCDXJeCRz69gItoBDRO7KfC-dODFpsMEAx1eq6VqZZAJFveBpc/s320/NetworkWithGrayNodes.png" vspace="10" width="320" /></a></span></div>This page contains links to responses by Google's Bard chatbot to 12 questions that should be asked more frequently, but aren't. As consequence, too many readily understood AI terms have become meaningless buzzwords in the media. </span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;">The questions on this page are ordered by increasing specificity, so they should be read in numerical order. Most readers are less likely to understand Bard's answers to later questions if they haven't read its answers to questions that appear earlier in the list.</span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><span face="Verdana, Geneva, sans-serif" style="background-color: white; color: #1a222a; font-size: medium;">Read more<b> </b></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="background-color: white; color: #336699; text-decoration-line: none;"><span style="color: #1a222a;">➡</span></a><b style="background-color: white; font-family: Verdana, Geneva, sans-serif; font-size: medium;"><span style="font-size: medium;"><span style="color: #1a222a;"> </span><i><a href="https://neoskeptics.blogspot.com/2023/12/basic-ai-faqs.html#more"><span style="color: #cc0000;">HERE</span></a></i></span></b></span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b style="background-color: white; font-family: Verdana, Geneva, sans-serif; font-size: medium;"><span style="font-size: medium;"><br /></span></b></span></div><span><span style="font-family: verdana;"><a name='more'></a><span><div style="text-align: center;"><span style="font-size: large;">Instructions </span><b style="background-color: white; color: #1a222a; text-align: left;">➡</b><span style="font-size: large;"> Click the linked terms in each question to load Bard's response to the question<br /><br /></span></div></span></span></span><div><span><span style="font-family: verdana;"><span style="font-size: medium;"><span><b>Question 1:</b></span> What is <b><i><a href="https://neoskeptics.blogspot.com/p/q1-what-is-artificial-intelligence-and.html">artificial intelligence</a></i></b> (AI) and what are its major subfields?</span></span></span><b style="font-family: verdana; font-size: large;"> </b></div><div><div><span style="font-family: verdana; font-size: medium;"><br /></span></div><div><span style="font-size: medium;"><span style="font-family: verdana;"><b>Question 2:</b> </span><span style="font-family: verdana;">What is <b><i><a href="https://neoskeptics.blogspot.com/p/q2-what-is-artificial-general.html">artificial general intelligence</a></i></b> (AGI)?</span></span></div><div><br /></div><div><span style="font-family: verdana;"><span style="font-size: medium;"><b>Question 3:</b> What is <b><i><a href="https://neoskeptics.blogspot.com/p/what-is-natural-language-processing.html">natural language processing</a></i></b> and what are the most important developments in its history?</span></span></div><div style="text-align: center;"><br /></div><div><span style="font-size: medium;"><span style="font-family: verdana;"><span><b>Question 4: </b></span></span><span style="font-family: verdana;"> <span>What is <b><i><a href="https://neoskeptics.blogspot.com/p/q4-what-is-machine-learning-how-does-it.html">machine learning?</a></i></b> How does it differ from rule-based learnng? Please provide examples</span></span></span></div><div><span style="font-family: verdana; font-size: medium;"><br /></span></div><div><span style="font-family: verdana; font-size: medium;"><b>Question 5:</b> <span style="font-size: medium;"><span>What impact has <b><i><a href="https://neoskeptics.blogspot.com/p/q5-what-impact-has-machine-learning-had.html">machine learning had on natural language processing</a></i></b>? Which kind of machine learning has had the largest impact?</span></span><b><span style="color: #999999;"><br /></span><br /></b></span></div><div style="text-align: left;"><span style="font-size: medium;"><span><span style="font-family: verdana;"><b>Question 6:</b> </span><span style="font-family: verdana;">What is a <b><i><a href="https://neoskeptics.blogspot.com/p/q6-what-is-language-model-and-what-are.html">language model</a> </i></b></span></span><span style="background-color: white; color: #1a222a;"><span style="font-family: verdana;">and what are the most important architectural structures for language models? Provide examples of recent language models that embody each architecture.</span></span></span></div><div style="text-align: center;"><span style="color: #999999; font-family: verdana;"><span style="font-family: verdana; font-size: medium;"><b style="color: red;"><br /></b></span></span></div><div><span style="font-family: verdana;"><span style="font-family: verdana; font-size: medium;"><b>Question 7:<span style="color: #999999;"> </span></b><span style="color: #999999;"> </span></span><span style="font-family: verdana; font-size: medium;">In the development of language models, what are<span style="color: #999999;"> </span><a href="https://neoskeptics.blogspot.com/p/q6-in-development-of-natural-language.html"><b style="font-style: italic;">emergent properties</b>?</a> </span></span></div><div style="text-align: center;"><br /></div><div><span style="font-size: medium;"><span style="font-family: verdana;"><span style="font-family: verdana;"><b><span><span>Question 8:</span></span><span> </span></b></span></span><span style="font-family: verdana;"><b><i><a href="https://neoskeptics.blogspot.com/p/q8-generative-language-models-can.html">Generative language models</a></i></b> can create original text, images, videos, sound, and code. What is done to language models to make them generative? And can small language models be generative?</span></span></div><div><span style="color: #999999; font-family: verdana; font-size: medium;"><b><br /></b></span></div><div><span style="font-family: verdana; font-size: medium;"><b>Question 9:</b> </span><span style="font-family: verdana; font-size: medium;">Are all emergent properties of a model the <b><i><a href="https://neoskeptics.blogspot.com/p/q9-are-all-emergent-properties-of-model.html">result of its generativity?</a></i></b></span></div><div style="text-align: center;"><span style="color: #999999; font-family: verdana; font-size: medium;"><b style="color: red;"><br /></b></span></div><div><span style="font-family: verdana; font-size: medium;"><b>Question 10:</b> What are the most important emergent properties that have been exhibited by <b><i><a href="https://neoskeptics.blogspot.com/p/q10-what-are-most-important-emergent.html">large language models</a></i></b>?</span></div><div style="text-align: center;"><br /></div><div><span><span style="font-family: verdana; font-size: medium;"><b>Question 11:</b> What are the most important emergent properties that have been exhibited by <b><i><a href="https://neoskeptics.blogspot.com/p/q11-what-are-most-important-emergent.html">small language models</a></i></b>?</span></span><div><div style="text-align: left;"><span style="color: #999999; font-family: verdana;"><span style="font-size: medium;"><br /></span></span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><span><span style="font-weight: bold;">Question 12</span></span></span><span style="font-family: verdana; font-size: medium;"><b> <span style="color: #cc0000;">via ChatGPT</span>: </b>Please identify <b><a href="https://neoskeptics.blogspot.com/p/q12-via-chatgpt-please-identify-seven.html"><i>seven small language</i> models</a> </b>thst have less than 80 billion parameters that were released in 2022 or 2023. <i>Note: Using ChatGPT because Bard (before Gemini Pro) unable to search Internet for recent content.</i></span></div><div style="text-align: center;"><br /></div><div style="text-align: left;"><br /></div><div style="text-align: left;"><span style="font-family: verdana;">____________________________________</span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b>Links to related notes: </b></span></div><div style="text-align: left;"><ul style="text-align: left;"><li style="background-color: white; border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana; font-size: medium;">"The Rise of Small Language Models— Efficient & Customizable", Bijit Ghosh, <a href="https://medium.com/@bijit211987/the-rise-of-small-language-models-efficient-customizable-cb48ddee2aad#id_token=eyJhbGciOiJSUzI1NiIsImtpZCI6IjFmNDBmMGE4ZWYzZDg4MDk3OGRjODJmMjVjM2VjMzE3YzZhNWI3ODEiLCJ0eXAiOiJKV1QifQ.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.CrodzVbvpWsfgDphqhte8dTLLam8G6UsHB4lv952wXzNFxNyI9bjPU0nCzfwxf-mtQyd02wN3bu4LEoUPFYsCm81w4jt_Q8ExgYSxIqAgQ5rm9zVANdSeHlQe2M9IF8_wiTSoCLQzBXX5JBf69RwLcBvPEmAHDKV9Xd6EyToYoeC9AWOEvdrl_v19hRhcybl4wYshK-i2T0Rv9CERLjk7PKF1YLRO_7w7Ji3QjGXh5wsYsunP-i3Fz61y5of1_aFcm0ECgZA0cmvJPg8giopftA2c04asPqZRRwVIOSJyU5gNuF7aXsHjLVa5z-yC6cSYIDoKdj6vET73XSkNd6hlQ" style="color: #336699; text-decoration-line: none;"><b><i>Medium</i></b></a>, 11/26/23</span></li><li style="background-color: white; border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana; font-size: medium;">"9 Best Small Language Models Released in 2023", Sandhra Jayan, <b><i><a href="https://analyticsindiamag.com/9-best-small-language-models-released-in-2023/" style="color: #336699; text-decoration-line: none;">AIM</a></i></b>, 12/7/23</span></li><li style="background-color: white; border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana; font-size: medium;">"Small language models an emerging GenAI force", Antone Gonsalves, <b><i><a href="https://www.techtarget.com/searchenterpriseai/news/366563445/Small-language-models-an-emerging-GenAI-force" style="color: #336699; text-decoration-line: none;">TechTarget</a></i></b>, 12/15/23</span></li><li style="background-color: white; border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana; font-size: medium;">"Everything You Need to Know about Small Language Models (SLM) and its Applications", Tanya Malhotra, <b><i><a href="https://www.marktechpost.com/2023/12/05/everything-you-need-to-know-about-small-language-models-slm-and-its-applications/" style="color: #336699; text-decoration-line: none;">MarTechPost</a></i></b>, 12/5/23</span></li><li style="background-color: white; border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana; font-size: medium;">"7 Steps to Running a Small Language Model on a Local CPU", Aryan Garg, <b><i><a href="https://www.kdnuggets.com/7-steps-to-running-a-small-language-model-on-a-local-cpu" style="color: #336699; text-decoration-line: none;">KDnuggets</a></i></b>, 11/14/23</span></li></ul></div></div></div></div>Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-88758508097698850982023-12-20T22:35:00.017-05:002024-03-27T19:20:54.518-04:00The surprising power of Small Language Models (SLMs) ... TL;DR and podcast 21Dec23<div style="text-align: right;"><span style="font-family: verdana; font-size: x-small;"><i>Last update: Thursday 12/21/23 </i></span></div><div style="text-align: left;"><span style="font-family: verdana;"><div class="separator" style="clear: both; text-align: center;"><div style="text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><span style="font-size: medium;"><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgKlFuZVsFJnHXSiBskG7thoLT4teY4sarTLN-Z5e42kgleL4W7h904bX2iNEepDoBdW0LU_eCyTweK42vay_z7q1rRinubOTDmtyJwqHGlJZikt_BR3sCToJfUI5H-aiWgnm-skyHTDTYp7XAlN54RY2CLOQLMyzbOdZpYXnlPbb_W8DWhEnnMm_Qn46c/s749/SmallRobotInPalmOfWhiteHand.webp" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="420" data-original-width="749" height="179" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgKlFuZVsFJnHXSiBskG7thoLT4teY4sarTLN-Z5e42kgleL4W7h904bX2iNEepDoBdW0LU_eCyTweK42vay_z7q1rRinubOTDmtyJwqHGlJZikt_BR3sCToJfUI5H-aiWgnm-skyHTDTYp7XAlN54RY2CLOQLMyzbOdZpYXnlPbb_W8DWhEnnMm_Qn46c/s320/SmallRobotInPalmOfWhiteHand.webp" vspace="10" width="320" /></a></div>Welcome to our 21Dec23</span></span></span></span><span style="font-size: medium;"><span><span style="background-color: white; color: #1a222a;"> </span><span style="background-color: white; color: #1a222a;">TL;DR summary + podcast about the past week's top AI story on our "</span><b style="color: #1a222a;"><i><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html">Useful AI News</a></i></b><span style="background-color: white; color: #1a222a;">" page </span></span><b style="background-color: white; color: #1a222a;">➡</b><span style="background-color: white; color: #1a222a;"> <span style="color: black; font-size: medium;">Microsoft's newest small language model (SLM) and its benchmarks</span> <br /><br /></span></span></div><div style="text-align: left;"><br /></div><div style="text-align: left;"><span><span style="background-color: white; color: #1a222a;"><span>Click the "start" button in the audio control (below) to hear the podcast ... </span></span></span><span style="background-color: white; color: #1a222a;"> </span><span style="background-color: white; color: #1a222a;">If audio fails to start, or gets stuck, try reloading the page</span><span style="background-color: white;"><span style="color: #1a222a; font-size: 13.86px;">.</span><br /><div style="text-align: right;"><br /></div></span></div><div style="background-color: white; color: #1a222a; font-family: Verdana, Geneva, sans-serif; text-align: left;"><span style="font-family: verdana; font-size: medium;"></span></div><div style="background-color: white; color: #1a222a; font-family: Verdana, Geneva, sans-serif; text-align: right;"><span style="font-family: verdana; font-size: medium;"> <iframe allow="autoplay" height="50" src="https://drive.google.com/file/d/1Kwl4Wwbcrfa7ISZl5a5gmq3G0wSF2D3S/preview?usp=sharing" width="400"></iframe></span><div class="separator" style="clear: both; text-align: left;"><span face="Verdana, Geneva, sans-serif" style="font-family: verdana; font-size: medium;">TL;DR link<b> </b></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-family: verdana; font-size: large; text-decoration-line: none;"><span style="color: #1a222a;">➡</span></a><b style="color: black;"><span style="font-size: medium;"><span style="color: #1a222a;"> </span><i><a href="https://neoskeptics.blogspot.com/2023/12/last-update-thursday-122123-welcome-to.html#more"><span style="color: #cc0000;">HERE</span></a></i></span></b></div></div></div></span></div><span><span style="font-family: verdana;"><a name='more'></a></span></span><div style="text-align: left;"><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;"><br /></b></div><div style="text-align: left;"><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;">A. Top story in past 2 weeks ...</b></div><div style="text-align: left;"><span style="background-color: white;"><span style="color: #1a222a; font-family: verdana;"><div>As the reader will see, we only have one top story. Given that Google had rolled out tables of benchmark tests two weeks ago that proclaimed the superiority of its biggest LLM cannon, Gemini Ultra, over Microsoft's biggest LLM cannon, GPT-4, the editor was not surprised when he learned that the Microsoft "Empire" had decided to strike back by firing at the biggest cannon in the Google "Empire". </div><div><br /></div><div>But the editor was stunned when he found Microsoft firing test results wherein its smallest cannon, Phi-2, outperformed Gemini Ultra. Phi-2 is not a large language model (LLM); Phi-2 is a small language model (SLM). Whereas Gemini Ultra contains 3.5 trillion parameters, Phi-2 only contains 2.7 billion parameters; so Ultra is about 580 times as large as Phi-2. Wow!!! How can such a little David outperform such a big Goliath at anything? Indeed, Microsoft published the benchmarks on its blog in a note titled: "<b><i><a href="https://www.microsoft.com/en-us/research/blog/phi-2-the-surprising-power-of-small-language-models/">The surprising power of small language models</a></i></b>"</div><div><br /></div><div>This story was obviously far more important that any other AI news item in the past two weeks. Indeed, Microsoft's announcement that small language models could support the emergence of the same kinds of dazzling capabilities that had emerged from large language models might be the most important generative AI news story since the publication of ChatGPT running GPT 3.5 in November 2022,</div><div><br /></div></span></span></div><div style="text-align: left;"><span style="background-color: white; font-family: verdana;"><b>Microsoft's newest small language model (SLM) and benchmarks</b></span></div><div style="text-align: left;"><span style="font-family: verdana;"><span style="background-color: white;">On December 12, 2023, staff from the Microsoft Research unit published a report on Microsoft's blog that included tables of benchmarks and other performance indicators that supported the assertion of the note's title as to "The surprising power of small language models." The report's Table 2 was the most relevant table for our discussion because it compares Microsoft's Phi-2 with data that Google published about its Gemini Nano 2 SLM. A copy of the report's Table 2 appears below. </span></span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">The second column shows that Nano 2 is about the same size as Phi-2, i.e., it has about the same number of parameters. The remaining four columns in the table provide four performance measres for each SLM. Phi-2 outperforms Nano 2 on all four measures, but it clobbers Nano 2 on MBPP (column 5), Phi-2's value 59.1 being more than twice as large as Nano 2's 29.1 value.</span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;"><b>What is MBPP?</b><br />The MBPP (Mostly Basic Python Problems) dataset consists of around 1,000 crowd-sourced Python programming problems, designed to be solvable by entry level programmers, covering programming, fundamentals, standard library functionality, and so on. Each problem consists of a task.</span></li></ul></div><div style="text-align: left;"><figure class="wp-block-table" style="background-color: white; border-collapse: collapse; box-sizing: border-box; margin: 40px 0px; overflow-x: auto;"><table style="border-collapse: collapse; font-family: "Segoe UI", SegoeUI, "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 0.875rem; margin-bottom: 0px; min-width: auto; width: 798.875px;"><thead style="border-bottom: 1px solid rgb(47, 47, 47); box-sizing: border-box;"><tr style="border-bottom: 1px solid rgb(200, 200, 200); box-sizing: border-box;"><th style="background-color: #002050; border: 1px solid rgb(204, 204, 204); box-sizing: border-box; color: white; padding: 6px 8px; text-align: inherit;"></th><th style="background-color: #002050; border: 1px solid rgb(204, 204, 204); box-sizing: border-box; color: white; padding: 6px 8px; text-align: inherit;">Size</th><th style="background-color: #002050; border: 1px solid rgb(204, 204, 204); box-sizing: border-box; color: white; padding: 6px 8px; text-align: inherit;">BBH</th><th style="background-color: #002050; border: 1px solid rgb(204, 204, 204); box-sizing: border-box; color: white; padding: 6px 8px; text-align: inherit;">BoolQ</th><th style="background-color: #002050; border: 1px solid rgb(204, 204, 204); box-sizing: border-box; color: white; padding: 6px 8px; text-align: inherit;">MBPP</th><th style="background-color: #002050; border: 1px solid rgb(204, 204, 204); box-sizing: border-box; color: white; padding: 6px 8px; text-align: inherit;">MMLU</th></tr></thead><tbody style="box-sizing: border-box;"><tr style="border-bottom: 1px solid rgb(200, 200, 200); box-sizing: border-box;"><td style="border: 1px solid rgb(204, 204, 204); box-sizing: border-box; padding: 6px 8px; vertical-align: top;">Gemini Nano 2</td><td style="border: 1px solid rgb(204, 204, 204); box-sizing: border-box; padding: 6px 8px; vertical-align: top;">3.2B</td><td style="border: 1px solid rgb(204, 204, 204); box-sizing: border-box; padding: 6px 8px; vertical-align: top;">42.4</td><td style="border: 1px solid rgb(204, 204, 204); box-sizing: border-box; padding: 6px 8px; vertical-align: top;">79.3</td><td style="border: 1px solid rgb(204, 204, 204); box-sizing: border-box; padding: 6px 8px; vertical-align: top;">27.2</td><td style="border: 1px solid rgb(204, 204, 204); box-sizing: border-box; padding: 6px 8px; vertical-align: top;">55.8</td></tr><tr style="border-bottom: 1px solid rgb(200, 200, 200); box-sizing: border-box;"><td style="border: 1px solid rgb(204, 204, 204); box-sizing: border-box; padding: 6px 8px; vertical-align: top;">Phi-2</td><td style="border: 1px solid rgb(204, 204, 204); box-sizing: border-box; padding: 6px 8px; vertical-align: top;">2.7B</td><td style="border: 1px solid rgb(204, 204, 204); box-sizing: border-box; padding: 6px 8px; vertical-align: top;">59.3</td><td style="border: 1px solid rgb(204, 204, 204); box-sizing: border-box; padding: 6px 8px; vertical-align: top;">83.3</td><td style="border: 1px solid rgb(204, 204, 204); box-sizing: border-box; padding: 6px 8px; vertical-align: top;">59.1</td><td style="border: 1px solid rgb(204, 204, 204); box-sizing: border-box; padding: 6px 8px; vertical-align: top;">56.7</td></tr></tbody></table><figcaption class="wp-element-caption" style="box-sizing: border-box; margin-top: 0.25rem; padding: 0.5rem 1rem 0.5rem 0px;"><center style="box-sizing: border-box;"><span style="font-family: verdana;"><span face=""Segoe UI", SegoeUI, "Helvetica Neue", Helvetica, Arial, sans-serif" style="box-sizing: border-box; font-weight: bolder;">Table 2.</span> Comparison between Phi-2 and Gemini Nano 2 Model on Gemini’s reported benchmarks</span></center></figcaption></figure></div><div style="text-align: left;"><span style="font-family: verdana;"><div id="Overview" style="background-color: white; color: #1a222a;">Now let's pause and take a couple of steps backwards. One of the most impressive emergent skills shown by ChatGPT, Bard, and other LLM chatbots is their ability to generate valid code in various languages, e.g., Python. Somehow the editor of this blog got the impression that only large language models could generate code, models that were at least 500 times as large as the pip-squeaks in Table 2</div><div id="Overview" style="background-color: white; color: #1a222a;"><ul style="text-align: left;"><li>ChatGPT's GPY-4 LLM has about 1.76 trillion parameters; ans</li><li>Gemini's Ultra has about 1.56 trillion parameters. </li></ul>Nevertheless, the data in Table 2 validates the title of the Mocrosoft blog note Small language models have surprising power.</div><div id="Overview" style="background-color: white; color: #1a222a;"><br /></div><div id="Overview" style="background-color: white; color: #1a222a;">As for Phi-2 being more than twice as powerful as Nano, this result should not be taken too seriously. Phi-2 is still a research project, whereas Google has already added Nano to its Pixel smartphone production lines. Indeed, Google has issued assurances that it will installi Nano on all android devices wherein they will function as stand alone models without connections to the Internet cloud. Although the editor could glibly wave aside this "victory" of Microsoft's SLM over Google's SLM, Microsoft's next victorious challenge cannot be ignored. </div><div id="Overview" style="background-color: white; color: #1a222a;"><br /></div><div id="Overview" style="background-color: white; color: #1a222a;">Microsoft posed the kind of question to Phi-2 and to Gemini Ultra that students might encounter in an introduction to physics course, specifically: to calculate the speed of a skier when it reaches the bottom of a hill, given the skier's mass, the gravitational constant, and the height of the hill. Phi-2 calculated the correct answer, but Ultra's answer was wrong. Adding insult to injury, Microsoft asked Phi-2 to identify the mistake in Ultra's answer ... which it did. </div><div id="Overview" style="background-color: white; color: #1a222a;"><br /></div><div id="Overview" style="background-color: white; color: #1a222a;">If the models had been asked to select the correct answer given multiple choiices, Phi-2 might have made a lucky guess. But this question, like all good questions in intro to physics courses, was an open question that required students to calculate the right answer and to show how they had made their calculation. Nevertheless, the SLM got it right; the LLM got it wrong; and the SLM identified the error in the LLM's logic. </div><div id="Overview" style="background-color: white; color: #1a222a;"><br /></div><div id="Overview" style="background-color: white; color: #1a222a;"><b>High quality data can empower SLMs</b></div><div id="Overview" style="background-color: white; color: #1a222a;">How could this happen? The answer implied by the authors of Microsoft's blog note, is plausible, snarky, and initially disturbing. </div><blockquote><div id="Overview" style="background-color: white; color: #1a222a;">"The massive increase in the size of language models to hundreds of billions of parameters has unlocked a host of emerging capabilities that have redefined the landscape of natural language processing. A question remains whether such emergent abilities can be achieved at a smaller scale using strategic choices for training, e.g., data selection ...</div></blockquote><blockquote><div id="Overview" style="background-color: white; color: #1a222a;">... training data quality plays a critical role in model performance. This has been known for decades, but we take this insight to its extreme by focusing on “textbook-quality” data, following upon our prior work “<b><i><a href="https://www.microsoft.com/en-us/research/publication/textbooks-are-all-you-need/">Textbooks Are All You Need.</a></i></b>” </div></blockquote><div id="Overview" style="background-color: white; color: #1a222a;">In other words, a substantial chunk of Phi-2's surprising power derives from its having been trained exclusively on high quality data, which the authors describe as "textbook" quality</div><div id="Overview" style="background-color: white; color: #1a222a;"><ul style="text-align: left;"><li>This is a plausible converse to one of information technology's oldest maxims <b style="font-size: large;">➡</b> garbage in, garbage out.<br /> </li><li>It's snarky because it suggests that Gemini Ultra's weaker performance might be due to its having been trained on inferior data. Indeed, if Gemini Ultra could not correctly answer a simple intro to physics question, what other shortfalls might its users encounter that might also follow from inferrior training data? <br /><br /></li><li>And it is initially disturbing because it suggests that Big Tech corporations may be currently engaged in a race to the bottom because of their de facto assumption that bigger is better. So they beg, borrow, and steal all of the data they can get their hands on, as fast as they can, from anywhere and everywhere, regardless of copyright restrictions, privacy violations, and gender/racial bias.</li></ul></div><div id="Overview" style="background-color: white; color: #1a222a;"><b>SLMs can empower competition</b></div><div id="Overview" style="background-color: white; color: #1a222a;">However, on second thought, initial disturbance quickly gives way to elation. If bigger was always better, generative AI would be forever dominated by Big Tech -- Apple, Amazon, Facebook/Meta, Google/Alphabet, and Microsoft -- because no other firms would be rich enough to afford this competition's ever increasing entry fees.</div><div id="Overview" style="background-color: white; color: #1a222a;"><br /></div><div id="Overview" style="background-color: white; color: #1a222a;">But common sense suggests that if SLMs provide viable foundations for generative AI, then the cost of producing small models should be orders of magnitude lower then the costs of producing models that were 500 times as large. Small models should only require investments measured in hundreds of millions or even tens of millions compared to the billions required to produce large models. Instead of five competitors, we might see 500. More competitors should produce more effective models that could address more important challenges much faster, at lower costs, and more safely if a substantial segment of the producers worked on open source models.</div><div id="Overview" style="background-color: white; color: #1a222a;"><br /></div><div id="Overview" style="background-color: white;"><b><span style="font-size: medium;">This is our last TL;DR/podcast for 2023 ... See you next year ... :-)</span></b></div><div id="Overview" style="background-color: white; text-align: center;"><b><i><span style="color: #cc0000; font-size: medium;">S</span><span style="color: #38761d; font-size: medium;">e</span><span style="color: #cc0000; font-size: medium;">a</span><span style="color: #674ea7; font-size: medium;">s</span><span style="color: #cc0000; font-size: medium;">ons </span><span style="color: #38761d; font-size: medium;">G</span><span style="color: #cc0000; font-size: medium;">ree</span><span style="color: #3d85c6; font-size: medium;">t</span><span style="color: #cc0000; font-size: medium;">in</span><span style="color: #2b00fe; font-size: medium;">g</span><span style="color: #cc0000; font-size: medium;">s ... </span><span style="color: #3d85c6; font-size: medium;">M</span><span style="color: #cc0000; font-size: medium;">e</span><span style="color: #38761d; font-size: medium;">r</span><span style="color: #cc0000; font-size: medium;">ry </span><span style="color: #38761d; font-size: medium;">C</span><span style="color: #cc0000; font-size: medium;">h</span><span style="color: #2b00fe; font-size: medium;">r</span><span style="color: #cc0000; font-size: medium;">is</span><span style="color: #a64d79; font-size: medium;">t</span><span style="color: #cc0000; font-size: medium;">mas ... </span><span style="color: #3d85c6; font-size: medium;">H</span><span style="color: #cc0000; font-size: medium;">appy </span><span style="color: #38761d; font-size: medium;">C</span><span style="color: #cc0000; font-size: medium;">ha</span><span style="color: #3d85c6; font-size: medium;">n</span><span style="color: #cc0000; font-size: medium;">u</span><span style="color: #674ea7; font-size: medium;">k</span><span style="color: #cc0000; font-size: medium;">ah </span></i></b></div><div id="Overview" style="background-color: white; text-align: center;"><b><i><span style="color: #cc0000; font-size: medium;">... </span><span style="color: #3d85c6; font-size: medium;">H</span><span style="color: #cc0000; font-size: medium;">a</span><span style="color: #38761d; font-size: medium;">p</span><span style="color: #cc0000; font-size: medium;">py </span><span style="color: #3d85c6; font-size: medium;">K</span><span style="color: #e69138; font-size: medium;">w</span><span style="color: #cc0000; font-size: medium;">a</span><span style="color: #674ea7; font-size: medium;">n</span><span style="color: #cc0000; font-size: medium;">za ... </span><span style="color: #45818e; font-size: medium;">H</span><span style="color: #cc0000; font-size: medium;">a</span><span style="color: #c27ba0; font-size: medium;">p</span><span style="color: #cc0000; font-size: medium;">py </span><span style="color: #0b5394; font-size: medium;">N</span><span style="color: #e69138; font-size: medium;">e</span><span style="color: #cc0000; font-size: medium;">w </span><span style="color: #6aa84f; font-size: medium;">Y</span><span style="color: #cc0000; font-size: medium;">ear!!!</span></i></b></div><div id="Overview" style="background-color: white; color: #1a222a;"><br /></div><div id="Overview" style="background-color: white; color: #1a222a; font-family: Verdana, Geneva, sans-serif; font-size: 13.86px;"><span style="font-family: verdana;"><span style="font-size: large;"><b>B. Top story in past 2 weeks ...</b></span></span></div><div id="Overview" style="background-color: white; color: #1a222a; font-family: Verdana, Geneva, sans-serif;"><ol><li style="margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;"><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#Microsoft" style="color: #336699; font-size: small; text-align: center; text-decoration-line: none;"><b><span style="color: #cc0000;">Microsoft</span></b></a><span style="font-size: medium;"><br /> </span><span style="font-size: medium;">"Microsoft releases Phi-2, a small language model AI that outperforms Llama 2, Mistral 7B", Carl Franzen, <b><i><a href="https://venturebeat.com/ai/microsoft-releases-phi-2-a-small-language-model-ai-that-outperforms-llama-2-mistral-7b/" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">VentureBeat</span></a></i></b>, 12/12/23 <b><span style="color: #cc0000;">***</span></b> <br />-- This story also covered by <b><i><a href="https://www.zdnet.com/article/microsoft-unveils-phi-2-a-small-language-model-that-packs-power/" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">ZDNet</span></a></i></b>, <b><i><a href="https://www.techrepublic.com/article/microsoft-phi-2-ai/" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">TechRepublic</span></a></i></b>, <b><i><a href="https://medium.com/@AITutorMaster/the-power-of-small-llm-microsofts-phi-2-and-the-future-of-ai-0204513c4692" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Medium</span></a></i></b>, <b><i><a href="https://t.co/VfkA061mo3" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Computerworld</span></a> </i></b>... and </span><b><i><a href="https://www.microsoft.com/en-us/research/blog/phi-2-the-surprising-power-of-small-language-models/" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;"><span style="font-size: medium;">Microsoft</span></span></a></i></b></span></li></ol></div></span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><div style="background-color: white; color: #1a222a; font-size: 13.86px;"><span style="font-family: verdana;"><span style="font-size: medium;"><b>C. Microsoft Research -- a component of Microsoft</b></span></span></div><div style="background-color: white;"><span><span style="color: #1a222a; font-family: verdana;"><div>The editor of this blog asked Microsoft Copilot the following questions:</div><div><b><br /></b></div><div><b>Question 1</b>: What does the Microsoft Research unit do? </div><div><b>Response 1</b>: </div><div>"Microsoft Research is a division of Microsoft Corporation that focuses on advancing science and technology to benefit humanity. The mission statement of Microsoft Research is to expand the state of the art in each of the areas in which it does research, rapidly transfer innovative technologies into Microsoft products, and ensure that Microsoft products have a future"</div><div><br /></div><div><b>Question 2: Please describe Microsoft's small language models</b></div><div><b>Reaponse 2:</b></div><div>"Microsoft Research has been working on developing small language models that can outperform larger models in conversational tasks </div><div><br /></div><div>1. In 2023, Microsoft Research introduced Orca 2, a pair of compact language models that have proven themselves by either matching or surpassing the capabilities of much larger language models </div><div><br /></div><div>2. Phi-2 is another small language model developed by Microsoft Research that demonstrates outstanding reasoning and language understanding capabilities, showcasing state-of-the-art performance among base language models with less than 13 billion"</div><div><br /></div><div><br /></div></span></span></div></div><div style="text-align: left;"><span style="font-family: verdana;"><div style="font-family: Times;"><div><span style="font-family: verdana;"><b>D. Links to ChatGPT's responses to five basic AI questions:</b> </span></div><div><span style="font-family: verdana;">ChatGPT's answer to each question provides the framework for its response to the next question. So most readers should start with Question 1.</span></div><div><span style="font-family: verdana;"><div style="color: #1a222a; font-size: 13.86px;"><ul style="line-height: 1.4; list-style-image: initial; list-style-position: initial; margin: 0.5em 0px; padding: 0px 2.5em;"><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><b>Question 1: <a href="https://neoskeptics.blogspot.com/2023/11/what-is-artificial-intelligence.html" style="color: #336699; text-decoration-line: none;">What is Artificial Intelligence?</a><br /><br /></b></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><b><h3 class="post-title entry-title" itemprop="name" style="font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-feature-settings: normal; font-kerning: auto; font-optical-sizing: auto; font-size: 18px; font-stretch: normal; font-variant-alternates: normal; font-variant-east-asian: normal; font-variant-numeric: normal; font-variant-position: normal; font-variation-settings: normal; line-height: normal; margin: 0px; position: relative;">Question 2: <a href="https://neoskeptics.blogspot.com/2023/11/question-2-what-is-machine-learning.html" style="color: #336699; text-decoration-line: none;">What is Machine Learning?</a></h3> </b></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-size: medium;"><b><h3 class="post-title entry-title" itemprop="name" style="font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-feature-settings: normal; font-kerning: auto; font-optical-sizing: auto; font-size: 18px; font-stretch: normal; font-variant-alternates: normal; font-variant-east-asian: normal; font-variant-numeric: normal; font-variant-position: normal; font-variation-settings: normal; line-height: normal; margin: 0px; position: relative;">Question 3: <a href="https://neoskeptics.blogspot.com/2023/11/question-4-what-are-neural-networks.html" style="color: #336699; text-decoration-line: none;">What are Neural Networks?</a> </h3><br /></b></span></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><h3 class="post-title entry-title" itemprop="name" style="font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-feature-settings: normal; font-kerning: auto; font-optical-sizing: auto; font-size: 18px; font-stretch: normal; font-variant-alternates: normal; font-variant-east-asian: normal; font-variant-numeric: normal; font-variant-position: normal; font-variation-settings: normal; line-height: normal; margin: 0px; position: relative;">Question 4: <a href="https://neoskeptics.blogspot.com/2023/11/question-5-what-are-gpt-models-chatgpts.html" style="color: #336699; text-decoration-line: none;">What are GPT models? </a><br /><br /></h3></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-size: medium;"><h3 class="post-title entry-title" itemprop="name" style="font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-feature-settings: normal; font-kerning: auto; font-optical-sizing: auto; font-size: 18px; font-stretch: normal; font-variant-alternates: normal; font-variant-east-asian: normal; font-variant-numeric: normal; font-variant-position: normal; font-variation-settings: normal; line-height: normal; margin: 0px; position: relative;">Question 5: <a href="https://neoskeptics.blogspot.com/2023/11/question-5-have-models-other-than-gpt.html" style="color: #336699; text-decoration-line: none;">Have models other than GPT made substantial progress towards AGI?</a> <span style="font-weight: normal;">... ChatGPT's response limited to its knowledge of models in 2022.</span></h3></span></li></ul></div></span></div></div></span></div>
Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-19203400549350587282023-12-10T11:58:00.116-05:002024-03-27T19:22:28.355-04:00First came the AI Big Bang; now comes the aftershocks and tremors... TL;DR and podcast 10Dec23 <div style="text-align: right;"><span style="font-family: verdana; font-size: x-small;"><i>Last update: Sunday 12/10/23 </i></span></div><div style="text-align: left;"><span style="font-family: verdana;"><div class="separator" style="clear: both; text-align: center;"><div style="text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><span style="font-size: medium;"><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhl5kWzUtqKIvOBTE6zuvYA8tI8hDVrKeOU5k37H-UT7HaUYdgB6A7wNiBBORGJqA0UNC_OxKEzqu8B_ca2eSEhsPQbdOcw7MrM9ri0Lf3z1Tg4O3YiHzXyS4eiTiZZ0LL2-7WChjeYsvkXnyH5pjPBx0sPY9Qjh-6cbdoox5-e5yW7KiUvQmNgKhiEndM/s2240/Demis-Hassabis--1760047727.webp" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="1497" data-original-width="2240" height="214" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhl5kWzUtqKIvOBTE6zuvYA8tI8hDVrKeOU5k37H-UT7HaUYdgB6A7wNiBBORGJqA0UNC_OxKEzqu8B_ca2eSEhsPQbdOcw7MrM9ri0Lf3z1Tg4O3YiHzXyS4eiTiZZ0LL2-7WChjeYsvkXnyH5pjPBx0sPY9Qjh-6cbdoox5-e5yW7KiUvQmNgKhiEndM/s320/Demis-Hassabis--1760047727.webp" vspace="10" width="320" /></a></div><br />Welcome to our </span></span></span></span><span style="font-size: medium;"><span><span style="background-color: white; color: #1a222a;">12Dec23</span><span style="background-color: white; color: #1a222a;"> </span><span style="background-color: white; color: #1a222a;">TL;DR + podcast about the past week's top 3 AI stories on our "</span><b style="color: #1a222a;"><i><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html">Useful AI News</a></i></b><span style="background-color: white; color: #1a222a;">" page </span></span><span style="background-color: white; color: #1a222a;">1) The biggest aftershock: Google's Gemini, (2) Tremor: Meta and IBM lead an alliance, and (3) Another tremor: Apple’s new framework </span></span></div><div style="text-align: left;"><span><span style="background-color: white; color: #1a222a;"><span><span style="font-size: medium;"><br /></span></span></span></span></div><div style="text-align: left;"><br /></div><div style="text-align: left;"><span><span style="background-color: white; color: #1a222a;"><span><span style="font-size: medium;"><b><br /></b></span></span></span></span></div><div style="text-align: left;"><span><span style="background-color: white; color: #1a222a;"><span><span style="font-size: medium;"><b>Demis Hassabis, PhD -- CEO, Google DeepMind</b></span></span></span></span></div><div style="text-align: left;"><span><span style="background-color: white; color: #1a222a;"><span><span><br /><div style="text-align: left;"><span style="background-color: transparent;"><span style="color: #1a222a;"><span>Click the "start" button in the audio control (below) to hear the podcast<span style="font-size: large;"> ... </span></span></span></span><span style="font-size: medium;"> If audio fails to start or gets stuck, try reloading the page</span><span><span style="font-size: 13.86px;">.</span></span><span style="color: #cc0000; font-size: large; text-align: right;"> </span></div></span></span></span></span></div><div style="background-color: white; font-family: Verdana, Geneva, sans-serif; text-align: right;"><span style="color: #1a222a; font-family: verdana; font-size: medium;"><iframe allow="autoplay" height="50" src="https://drive.google.com/file/d/1KhriHvgQnU6b8xwqiPM-nUNCNpprsqDU/preview?usp=sharing" width="400"></iframe></span><div class="separator" style="clear: both; color: #1a222a; text-align: left;"><span face="Verdana, Geneva, sans-serif" style="font-family: verdana; font-size: medium;">TL;DR link<b> </b></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-family: verdana; font-size: large; text-decoration-line: none;"><span style="color: #1a222a;">➡</span></a><b style="color: black;"><span style="font-size: medium;"><span style="color: #1a222a;"> </span><i><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html#more"><span style="color: #cc0000;">HERE</span></a></i></span></b></div><div class="separator" style="clear: both; text-align: left;"><span style="font-size: medium;"><span style="color: #1a222a;"></span><b><span><a name='more'></a></span><div style="text-align: center;"><br /></div></b></span></div></div></div></span></div><div style="text-align: left;"><span style="background-color: white;"><div><b style="color: #1a222a; font-family: verdana; font-size: large;">Background ...</b></div><div><span style="color: #1a222a; font-family: verdana;">Readers of this blog may recall that "The AI Big Bang" was the name we gave to the long, tumultuous weekend that began on Friday evening 11/17/23 when the OpenAI board abruptly notified its CEO Sam Altman that he was fired. It ended on Tuesday evening 11/21/23, when Altman returned to OpenAI as CEO</span></div><div><span style="color: #1a222a; font-family: verdana; font-size: medium;"><br /></span></div><div><span style="color: #1a222a; font-family: verdana;">So who won this brief high tech battle? Microsoft and its CEO Satya Nadella won handily. Microsoft's stock hit an all-time high after Altman's return was announced. </span></div><div><span style="color: #1a222a; font-family: verdana; font-size: medium;"><br /></span></div><div><span style="color: #1a222a; font-family: verdana; font-size: medium;">Microsofts supremacy was quickly challenged by its leading Big Tech AI competitors -- Amazon, Google, Facebook, and Apple. Their complex challenges were under preparation long before the Big Bang, but the competitors were pressured to release their challenges before they were fully completed, presumably to provide their investors with timely reassurance that their AI programs were still viable. </span></div><div><span style="color: #1a222a; font-family: verdana; font-size: medium;"><br /></span></div><div><span style="color: #1a222a; font-family: verdana; font-size: medium;">Most of the challenges were incremental tremors. Only one challenge was an aftershock whose innovative magnitude was</span><span style="font-size: medium;"><span style="color: #1a222a; font-family: verdana;"> </span><span style="color: #1a222a; font-family: verdana;">comparable to Microsoft's Big Bang</span></span></div><div><span style="color: #1a222a; font-family: verdana; font-size: medium;"><br /></span></div><div><span style="color: #1a222a; font-family: verdana; font-size: medium;">To be specific, Amazon's announcements at its "re:Invent" conference last week was a tremor; and as our reader will soon see, the challenges from Facebook and Apple this week were also tremors. Only Google's challenge was an innovative aftershock.</span></div></span></div><div style="text-align: left;"><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;"><div style="color: black; font-family: Times; font-size: medium; font-weight: 400;"><span style="font-family: verdana;"><div class="separator" style="clear: both; text-align: center;"><div style="font-family: Verdana, Geneva, sans-serif; text-align: right;"><div class="separator" style="clear: both; text-align: left;"><span style="font-size: medium;"><b><div><b><i><span style="color: #cc0000;"><br /></span></i></b></div></b></span></div></div></div></span></div></b></div><div style="text-align: left;"><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;">1) Google announced "Gemini"... its new LLM</b></div><div style="text-align: left;"><span style="background-color: white; font-family: verdana;"><div style="color: #1a222a;">Gemini comes in three versions: Gemini <b>Nano</b>, Gemini <b>Pro</b>, and Gemini <b>Ultra</b></div><div><ul style="text-align: left;"><li style="color: #1a222a;"><b>Nano,</b> the lightest version, will eventually run on all Android devices, but currently only runs on Google's Pixel smartphones<br /><br /></li><li style="color: #1a222a;"><b>Pro</b> is more powerful than Nano, and currently runs Bard on all devices. Bard can be accessed at <a href="http://bard.google.com"><b><i>bard.google.com</i></b></a><br />-- Pro also runs Google's NotebookLM application, an innovative AI driven app. NotebookLM can be accessed at <b><i><a href="http://notebooklm.google.com">notebooklm.google.com</a><br /></i></b>-- Pro will be available to developers and enterprise customers in mid-December via Google AI Studio<br /><br /></li><li style="color: #1a222a;"><b>Ultra</b> is the most powerful version, but Google has not yet announced a release date<br /><br /></li><li style="color: #1a222a;">At present Nano and Pro accept text input, then deliver text output, but all versions will eventually be <b>multimodal</b>, accepting input as text, images, video, audio, and code.<br /><br /></li><li style="color: #1a222a;">All models were trained on Google's own chips (Tensor Processing Units) and will run in data centers on those chips<br /><br /></li><li style="color: #1a222a;">According to <b><i><a href="https://www.theverge.com/2023/12/6/23990466/google-gemini-llm-ai-model">The Verge</a></i></b>, "Hassabis says. 'There’s still things like action, and touch — more like robotics-type things.' Over time, he says, Gemini will get more senses, become more aware, and become more accurate and grounded in the process. 'These models just sort of understand better about the world around them.' These models still hallucinate, of course, and they still have biases and other problems. But the more they know, Hassabis says, the better they’ll get."<br /><br /></li><li style="color: #1a222a;"><b>Benchmarks</b><br /><b><i><a href="https://deepmind.google/technologies/gemini/#capabilities">Google's official announcement</a> </i></b>presented the results of 18 benchmarks tests that compared Gemini Ultra with GPT-4. Gemini performed better than GPT-4 by small margins on all but one test.<br /><br /></li><li><span style="color: #1a222a;"><b>Stock market </b>... the most important benchmark???</span><br /><span style="color: #1a222a;">"Alphabet soars as Wall Street cheers arrival of AI model Gemini", Aditya Soni, </span><b><i><a href="https://t.co/tKAFY9a7zp"><span>Reuters</span></a></i></b><span style="color: black;">, 12/7/23</span><br /><br /></li></ul></div></span></div><div style="text-align: left;"><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;">2) Meta, IBM, and others announce the "AI alliance"</b></div><div style="text-align: left;"><span style="color: #1a222a; font-family: verdana; font-size: medium;"><span style="background-color: white;">According to the mission statement of the AI Alliance on its <b><i><a href="https://thealliance.ai/">website</a></i></b>: </span></span></div><div style="text-align: left;"><span style="color: #1a222a; font-family: verdana; font-size: medium;"><blockquote><span style="background-color: white;">"</span>The AI Alliance is focused on accelerating and disseminating open innovation across the AI technology landscape to improve foundational capabilities, safety, security and trust in AI, and to responsibly maximize benefits to people and society everywhere. </blockquote><blockquote>The AI Alliance brings together a critical mass of compute, data, tools, and talent to accelerate open innovation in AI.<span style="background-color: white;">"</span></blockquote></span></div><div style="text-align: left;"><span style="color: #1a222a; font-family: verdana; font-size: medium;"><span style="background-color: white;">There are more than fifty initial members in the alliance: corporations, universities, non-profits, and government agencies ... in the U.S., Germany, Italy, Japan, Israel, India, UAE, UK, Australia </span></span></div><div style="text-align: left;"><span style="color: #1a222a; font-family: verdana; font-size: medium;"><span style="background-color: white;"><br /></span></span></div><div style="text-align: left;"><span style="color: #1a222a; font-family: verdana; font-size: medium;"><span style="background-color: white;">Its mission sounds uncomfortably like the mission of OpenAI when it was founded back in 2014. But as we all know, OpenAI soon became Microsoft's well-funded R&D subsidiary. So will such a large alliance prove to be too unwieldy to be effective? If so, might it be then captured by its corporate sponsors? </span></span></div><div style="text-align: left;"><span style="color: #1a222a; font-family: verdana; font-size: medium;"><span style="background-color: white;"><br /></span></span></div><div style="text-align: left;"><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;">3) Apple announced a machine learnning framework running on its own chips</b></div><div style="text-align: left;"><span style="color: #1a222a; font-family: verdana; font-size: medium;"><span style="background-color: white;">Whereas Microsoft and Google are producing AI tools that can be used today by non-techs, Apple's MLX machine learning framework is a tool that machine learning techs can use to generate models that will provide AI services to non-techs tomorrow ... but probably no time soon.</span></span></div><div style="text-align: left;"><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;"><br /></b></div><div style="text-align: left;"><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;"><br /></b></div><div style="text-align: left;"><b style="background-color: white; color: #1a222a; font-family: verdana; font-size: large;">B. </b><b style="font-family: verdana; font-size: large;">Top 3 stories in past week ...</b></div><div id="Overview"><ol><li><span style="font-family: verdana;"><a href="#Google" style="font-weight: 700; text-align: center;"><span style="color: #cc0000; font-size: x-small;">Google</span></a><br /><span style="font-size: medium;"><span>"Google DeepMind's Demis Hassabis Says Gemini Is a New Breed of AI", Will Knight, <i><b><a href="https://www.wired.com/story/google-deepmind-demis-hassabis-gemini-ai/"><span style="color: #cc0000;">Wired</span></a></b></i>, 12/6/23 <b><span style="color: #cc0000;">***</span></b> <br />-- This story also covered by <b><i><a href="https://www.theverge.com/2023/12/6/23990466/google-gemini-llm-ai-model"><span style="color: #cc0000;">The Verge</span></a></i></b>, <b><i><a href="https://www.technologyreview.com/2023/12/06/1084471/google-deepminds-new-gemini-model-looks-amazing-but-could-signal-peak-ai-hype/"><span style="color: #cc0000;">MIT Tech Review</span></a></i></b>, <b><i><a href="https://www.nytimes.com/2023/12/06/technology/google-ai-bard-chatbot-gemini.html"><span style="color: #cc0000;">NY Tmes</span></a></i></b>, <b><i><a href="https://arstechnica.com/information-technology/2023/12/google-launches-gemini-a-powerful-ai-model-it-says-can-surpass-gpt-4/"><span style="color: #cc0000;">Ars Technica</span></a></i></b> ... and </span><a href="https://deepmind.google/technologies/gemini/" style="font-style: italic; font-weight: bold;"><span style="color: #cc0000;">Google DeepMind</span></a><br /><br />-- Fake demos in Google's announcements: "Google just launched a new AI and has already admitted at least one demo wasn’t real", Emilia David, <b><i><a href="https://www.theverge.com/2023/12/7/23992737/google-gemini-misrepresentation-ai-accusation"><span style="color: #cc0000;">The Verge</span></a></i></b>, 12/7/23 ... Also reported by <b><i><a href="https://t.co/anasViMrTX"><span style="color: #cc0000;">TechCrunch</span></a></i></b><b style="font-style: italic;">, </b><a href="https://t.co/P1P2Y7b493" style="font-style: italic; font-weight: bold;"><span style="color: #cc0000;">Gizmodo</span></a><b style="font-style: italic;">, </b><a href="https://t.co/O44Xajtevq"><span style="color: #cc0000;"><b><i>BBC</i></b></span></a><br /><br />-- Users access Gemini Pro via Bard: "How to Use Google’s Gemini AI Right Now in Its Bard Chatbot", Reece Rogers, <b><i><a href="https://t.co/Hzh2xDShSb"><span style="color: #cc0000;">Wired</span></a></i></b>, 12/6/23<br /><br />-- Google's new Gemini powered notebook: "Google’s AI note-taking app is now available to users in the US", Emma Roth, <b><i><a href="https://www.theverge.com/2023/12/8/23993441/google-notebooklm-available-gemini-pro"><span style="color: #cc0000;">The Verge</span></a></i></b>, 12/8/23 ... Also reported by <b><i><a href="https://twitter.com/engadget/status/1733153365255733551/photo/1"><span style="color: #cc0000;">Engadget</span></a></i></b>, <b><i><a href="https://twitter.com/Gizmodo/status/1733159646368878701/photo/1"><span style="color: #cc0000;">Gizmodo</span></a></i></b>, <br /><br />-- Announcements boosted Google/Alphabet's stock: "Alphabet soars as Wall Street cheers arrival of AI model Gemini", Aditya Soni, <b><i><a href="https://t.co/tKAFY9a7zp"><span style="color: #cc0000;">Reuters</span></a></i></b>, 12/7/23<br /><br /></span></span></li><li><span style="font-family: verdana;"><a href="#LLMNews" style="font-size: small; font-weight: 700; text-align: center;"><span style="color: #cc0000;">LLM News</span></a><br /><span style="font-size: medium;">"Meta and IBM Launch AI Alliance, Promise to Be ‘Open’", Maxwell Zeff, <b><i><a href="https://t.co/EToVwS33jH"><span style="color: #cc0000;">Gizmodo</span></a></i></b>, 12/6/23 <b><span style="color: #cc0000;">***</span></b> <br />-- This story also covered by </span><i><span style="font-size: medium;"><a href="https://t.co/UMXlqc9xZ8" style="font-weight: bold;"><span style="color: #cc0000;">Computerworld</span></a><b>, <a href="https://techcrunch.com/2023/12/05/meta-and-ibm-form-an-ai-alliance-but-to-what-end/"><span style="color: #cc0000;">TechCrunch</span></a>, </b><a href="https://fortune.com/2023/12/05/ibm-meta-ai-alliance-open-source-closed-development-ai/" style="font-weight: bold;"><span style="color: #cc0000;">Fortune</span></a><b>, </b><a href="https://newsroom.ibm.com/AI-Alliance-Launches-as-an-International-Community-of-Leading-Technology-Developers,-Researchers,-and-Adopters-Collaborating-Together-to-Advance-Open,-Safe,-Responsible-AI" style="font-weight: bold;"><span style="color: #cc0000;">Ars Technica</span></a></span><b><span style="font-size: medium;">, <a href="https://www.bloomberg.com/news/articles/2023-12-05/meta-ibm-create-industrywide-ai-alliance-to-share-technology"><span style="color: #cc0000;">Bloomberg</span></a>, <br /><br /></span></b></i></span></li><li><span style="font-family: verdana;"><span style="color: #1a222a;"><a href="#NonLLMNews" style="font-size: small; font-weight: 700; text-align: center;"><span style="color: #cc0000;">Non-LLM News</span></a><br /><span style="font-size: medium;">"Apple Scurries Onto the AI Stage With Too Little, Too Late", Maxwell Zeff, </span></span><span style="font-size: medium;"><b><i><a href="https://t.co/361n2v4GfI"><span style="color: #cc0000;">Gizmodo</span></a></i></b><span><span style="color: #1a222a;">, 12/7/23 ... the MLX machine learning framework that runs on Apple's own chips </span><b><span style="color: #cc0000;">***</span></b><span style="color: #1a222a;"> </span></span><br /><span style="color: #1a222a;">-- This story also covered by </span><b><i><a href="https://t.co/6ZCIut7Jqv"><span style="color: #cc0000;">The Verge</span></a></i></b><span style="color: #1a222a;">, </span><b><i><a href="https://t.co/N5SciOlbXA"><span style="color: #cc0000;">Computerworld</span></a></i></b></span></span></li></ol></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b><br /></b></span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b>C. Who is Demis Hassabis, </b></span><span style="background-color: white; color: #1a222a; font-family: verdana;"><span style="font-size: medium;"><b>CEO of Google DeepMind</b></span></span><b style="font-family: verdana; font-size: large;">? ... Two interviews</b></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;"><span style="background-color: white; color: #1a222a;">"</span><span style="background-color: white; color: #1a222a;">Inside Google’s big AI shuffle — and how it plans to stay competitive, with Google DeepMind CEO Demis Hassabis", Nilay Patel, <b><i><a href="https://www.theverge.com/23778745/demis-hassabis-google-deepmind-ai-alphafold-risks" style="color: #336699; text-decoration-line: none;">The Verge</a></i></b>, 7/10/23 ... audio and transcript of interview <br /></span><span style="background-color: white; color: #1a222a;"><br /></span></span></li><li><span style="font-family: verdana;"><span style="background-color: white; color: #1a222a;">"A.I. Could Solve Some of Humanity's Hardest Problems. It Already Has.". Guest = Demis Hassabis, <a href="https://www.nytimes.com/2023/07/11/opinion/ezra-klein-podcast-demis-hassabis.html" style="color: #336699; text-decoration-line: none;"><b><i>The Ezra Klein Show</i></b></a></span><span style="background-color: white; color: #1a222a;"> (podcast with transcript), 7/11/23 </span></span></li></ul></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><div><div><div style="background-color: white; color: #1a222a; font-size: 13.86px;"><span style="font-family: verdana;"><span style="font-size: medium;"><b>D. Four breakthroughs in science/technology using DeepMind's AI/Neural Networks</b></span></span></div><div style="background-color: white;"><span style="font-family: verdana;"><div style="font-family: Verdana, Geneva, sans-serif;"><ul style="line-height: 1.4; list-style-image: initial; list-style-position: initial; margin: 0.5em 0px; padding: 0px 2.5em;"><li style="border: none; color: #1a222a; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;">"One of the Biggest Problems in Biology Has Finally Been Solved", Tanya Lewis, <b><i><a href="https://www.scientificamerican.com/article/one-of-the-biggest-problems-in-biology-has-finally-been-solved/" style="color: #336699; text-decoration-line: none;">Scientific American</a></i></b>, 10/31/22 ... protein foldimg problem ...<i>DeepMind</i><br />-- This story also covered by <b><i><a href="https://venturebeat.com/ai/a-year-ago-deepminds-alphafold-ai-changed-the-shape-of-science-but-there-is-more-work-to-do/" style="color: #336699; text-decoration-line: none;">VentureBeat</a></i></b>, <b><i><a href="https://www.cnet.com/science/biology/googles-deepmind-ai-predicts-3d-structure-of-nearly-every-protein-known-to-science/" style="color: #336699; text-decoration-line: none;">CNET</a></i></b>, <b><i><a href="https://www.science.org/content/article/google-s-deepmind-aces-protein-folding" style="color: #336699; text-decoration-line: none;">Science</a></i></b>, <br /><br /></span></li><li style="border: none; color: #1a222a; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;">"Courtesy of AI: Weather forecasts for the hour, the week and the century", Devin Coldewey, <b><i><a href="https://t.co/7szMYJTtMu" style="color: #336699; text-decoration-line: none;">TechCrunch</a></i></b>, 11/14/23 ... <i style="color: black;">DeepMind</i><br />-- This story also covered by <b><i><a href="https://t.co/ZVUWEP6aUh" style="color: #336699; text-decoration-line: none;">Gizmodo</a></i></b>, <b><i><a href="https://www.ft.com/content/ca5d655f-d684-4dec-8daa-1c58b0674be1" style="color: #336699; text-decoration-line: none;">Financial Times</a></i></b>, <b><i><a href="https://www.technologyreview.com/2023/11/14/1083366/google-deepminds-weather-ai-can-forecast-extreme-weather-quicker-and-more-accurately/" style="color: #336699; text-decoration-line: none;">MIT Tech Review</a></i></b>, <b><i><a href="https://arstechnica.com/science/2023/11/ai-outperforms-conventional-weather-forecasting-for-the-first-time-google-study/" style="color: #336699; text-decoration-line: none;">Ars Technica</a></i></b>, <b><i><a href="https://www.wired.com/story/google-deepmind-ai-weather-forecast/" style="color: #336699; text-decoration-line: none;">Wired</a></i></b>, <b><i><a href="https://t.co/2P7gFHA3Ut" style="color: #336699; text-decoration-line: none;">ZDNet</a></i></b>, <br />-- These articles describe research published in <b><i><a href="https://www.nature.com/articles/d41586-023-03552-y" style="color: #336699; text-decoration-line: none;">Nature</a><b> </b><span style="font-style: normal; font-weight: 400;">and</span><b> <a href="https://www.science.org/doi/10.1126/science.adi2336" style="color: #336699; text-decoration-line: none;">Science</a><br /><br /></b></i></b></span></li><li style="border: none; color: #1a222a; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;">"AI Tool Pinpoints Genetic Mutations That Cause Disease", Even Callaway & Nature Magazine, <b><i><a href="https://www.scientificamerican.com/article/ai-tool-pinpoints-genetic-mutations-that-cause-disease/" style="color: #336699; text-decoration-line: none;"><span style="color: black;">Scientific American</span><span style="color: #cc0000;">,</span></a></i></b> 11/21/23 ... <i style="color: black;">DeepMind</i><br />-- This story also covered by<span style="color: black;"> </span><span style="color: black;"><a href="https://www.wsj.com/podcasts/tech-news-briefing/how-googles-deepmind-is-using-ai-to-uncover-the-causes-of-diseases/a0a6e750-b1f0-4265-a675-d00129059e7a#:~:text=Trying%20to%20determine%20who%20might,mutations%20that%20could%20cause%20illness" style="color: #336699; font-style: italic; font-weight: bold; text-decoration-line: none;">Wall Street Journal</a><b style="font-style: italic;">, </b><a href="https://www.wired.com/story/deepmind-ai-alphamissense-genetics-rare-diseases/" style="color: #336699; text-decoration-line: none;"><b><i>Wired</i></b></a><b style="font-style: italic;">, <br /><br /></b></span></span></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;">"Google’s DeepMind finds 2.2M crystal structures in materials science win", Michael Peel, <a href="https://arstechnica.com/ai/2023/11/googles-deepmind-finds-2-2m-crystal-structures-in-materials-science-win/" style="text-decoration-line: none;"><b><i>Ars Technica</i></b></a><span style="color: #1a222a;">, 11/29/23 ... "The trove of theoretically stable but experimentally unrealized combinations identified using an AI tool known as GNoME is more than 45 times larger than the number of such substances unearthed in the history of science, according to a paper published in Nature on Wednesday."</span><br style="color: #1a222a;" /><span style="color: #1a222a;">-- This story also covered by</span><b><i> </i></b><b><i><a href="https://www.wired.com/story/an-ai-dreamed-up-380000-new-materials-the-next-challenge-is-making-them/" style="text-decoration-line: none;">Wired</a>, <a href="https://venturebeat.com/ai/google-deepminds-materials-ai-has-already-discovered-2-2-million-new-crystals/" style="text-decoration-line: none;">VentureBeat</a>, <a href="https://www.nature.com/articles/d41586-023-03745-5" style="text-decoration-line: none;">Nature</a>, <a href="https://www.science.org/content/article/materials-predicting-ai-deepmind-could-revolutionize-electronics-batteries-and-solar" style="text-decoration-line: none;">Science</a>,</i></b></span></li></ul></div></span></div></div></div></div><div style="text-align: left;"><span style="font-family: verdana;"><div style="font-family: Times;"><div><span style="font-family: verdana;"><b><br /></b></span></div><div><span style="font-family: verdana;"><b>D. Links to ChatGPT's responses to five basic AI questions:</b> </span></div><div><span style="font-family: verdana;">ChatGPT's answer to each question provides the framework for its response to the next question. So most readers should start with Question 1.</span></div><div><span style="font-family: verdana;"><div style="color: #1a222a; font-size: 13.86px;"><ul style="line-height: 1.4; list-style-image: initial; list-style-position: initial; margin: 0.5em 0px; padding: 0px 2.5em;"><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><b>Question 1: <a href="https://neoskeptics.blogspot.com/2023/11/what-is-artificial-intelligence.html" style="color: #336699; text-decoration-line: none;">What is Artificial Intelligence?</a><br /><br /></b></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><b><h3 class="post-title entry-title" itemprop="name" style="font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-feature-settings: normal; font-kerning: auto; font-optical-sizing: auto; font-size: 18px; font-stretch: normal; font-variant-alternates: normal; font-variant-east-asian: normal; font-variant-numeric: normal; font-variant-position: normal; font-variation-settings: normal; line-height: normal; margin: 0px; position: relative;">Question 2: <a href="https://neoskeptics.blogspot.com/2023/11/question-2-what-is-machine-learning.html" style="color: #336699; text-decoration-line: none;">What is Machine Learning?</a></h3> </b></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-size: medium;"><b><h3 class="post-title entry-title" itemprop="name" style="font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-feature-settings: normal; font-kerning: auto; font-optical-sizing: auto; font-size: 18px; font-stretch: normal; font-variant-alternates: normal; font-variant-east-asian: normal; font-variant-numeric: normal; font-variant-position: normal; font-variation-settings: normal; line-height: normal; margin: 0px; position: relative;">Question 3: <a href="https://neoskeptics.blogspot.com/2023/11/question-4-what-are-neural-networks.html" style="color: #336699; text-decoration-line: none;">What are Neural Networks?</a> </h3><br /></b></span></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><h3 class="post-title entry-title" itemprop="name" style="font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-feature-settings: normal; font-kerning: auto; font-optical-sizing: auto; font-size: 18px; font-stretch: normal; font-variant-alternates: normal; font-variant-east-asian: normal; font-variant-numeric: normal; font-variant-position: normal; font-variation-settings: normal; line-height: normal; margin: 0px; position: relative;">Question 4: <a href="https://neoskeptics.blogspot.com/2023/11/question-5-what-are-gpt-models-chatgpts.html" style="color: #336699; text-decoration-line: none;">What are GPT models? </a><br /><br /></h3></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-size: medium;"><h3 class="post-title entry-title" itemprop="name" style="font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-feature-settings: normal; font-kerning: auto; font-optical-sizing: auto; font-size: 18px; font-stretch: normal; font-variant-alternates: normal; font-variant-east-asian: normal; font-variant-numeric: normal; font-variant-position: normal; font-variation-settings: normal; line-height: normal; margin: 0px; position: relative;">Question 5: <a href="https://neoskeptics.blogspot.com/2023/11/question-5-have-models-other-than-gpt.html" style="color: #336699; text-decoration-line: none;">Have models other than GPT made substantial progress towards AGI?</a> <span style="font-weight: normal;">... ChatGPT's response limited to its knowledge of models in 2022.</span></h3></span></li></ul></div></span></div><div><span style="font-family: verdana;"> </span></div></div></span></div>
Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-37672963341076341632023-12-03T17:20:00.057-05:002024-03-27T19:23:42.491-04:00Amazon's re:Invent, DeepMind's breakthrough, and Microsoft as OpenAI's partner ...TL;DR and Podcast 3Dec23<div style="text-align: right;"><span style="font-family: verdana; font-size: x-small;"><i>Last update: Monday 12/4/23 </i></span></div><div style="text-align: left;"><span style="font-family: verdana;"><div class="separator" style="clear: both; text-align: center;"><div style="text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><span style="font-size: medium;"><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgunIckrnH7yBR56aG_B9wAszTDXTSTmyWcpLhHPSTcgviu6TAvqIc3AHZAqgSSlX60mSCg6Lzt2cjiaZnQYnaOiOW2hw8d6E7EXFZRTBYXls3xFLn-gxlEoIl5Z3hxG1vCA-sU0xHRnGoiNVR3_SEpR_IGtPq_MMmLvveMoIWLirhJC5d_lAXun2bzKCI/s800/LabRobotDeepMindCrystalStructures.jpeg" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="450" data-original-width="800" height="180" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgunIckrnH7yBR56aG_B9wAszTDXTSTmyWcpLhHPSTcgviu6TAvqIc3AHZAqgSSlX60mSCg6Lzt2cjiaZnQYnaOiOW2hw8d6E7EXFZRTBYXls3xFLn-gxlEoIl5Z3hxG1vCA-sU0xHRnGoiNVR3_SEpR_IGtPq_MMmLvveMoIWLirhJC5d_lAXun2bzKCI/s320/LabRobotDeepMindCrystalStructures.jpeg" vspace="10" width="320" /></a></div>Welcome to our TL;DR summary and podcast about the past week's top 3 AI stories on our "<b><i><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html" style="color: #336699; text-decoration-line: none;">Useful AI News</a></i></b>" </span></span></span></span><span style="font-size: medium;"><span><span style="background-color: white; color: #1a222a;">page </span></span><b style="background-color: white; color: #1a222a;">➡</b><span style="background-color: white; color: #1a222a;"> </span><span style="background-color: white; color: #1a222a;">1) Amazon's "re;Invent 2023" customer conference, (2) Google DeepMind's latest scientific breakthrough, and (3) A closer look at Microsoft's side of its partnership with OpenAI.<br /><br /></span></span></div><div style="text-align: left;"><span><span><span style="background-color: white; color: #1a222a;">... </span></span><span style="background-color: white; color: #1a222a;"><span>Click the "start" button in the audio control (below) to hear the podcast</span><span>. If audio fails to start, or gets stuck, try reloading the page.</span></span></span></div><div style="background-color: white; color: #1a222a; font-family: Verdana, Geneva, sans-serif; font-size: 13.86px; text-align: left;"><span style="font-family: verdana; font-size: medium;"></span></div><div style="background-color: white; color: #1a222a; font-family: Verdana, Geneva, sans-serif; text-align: right;"><span style="font-family: verdana; font-size: medium;">
<iframe allow="autoplay" height="50" src="https://drive.google.com/file/d/1KdB-z0WEa2mWF0koXFodcpbcvRK7zRwE/preview?usp=sharing" width="400"></iframe></span><div class="separator" style="clear: both; text-align: left;"><span face="Verdana, Geneva, sans-serif" style="font-family: verdana; font-size: medium;">TL;DR link<b> </b></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-family: verdana; font-size: large; text-decoration-line: none;"><span style="color: #1a222a;">➡</span></a><b style="color: black;"><span style="font-size: medium;"><span style="color: #1a222a;"> </span><i><a href="https://neoskeptics.blogspot.com/2023/12/new-tldr-podcast-3dec23-tldr-summary.html#more"><span style="color: #cc0000;">HERE</span></a></i></span></b></div></div></div></span></div><span><span style="font-family: verdana;"><a name='more'></a></span></span><div style="text-align: left;"><span style="font-family: verdana;"><b style="font-size: large;">1) </b><span style="font-size: medium;"> </span><span><b><span style="font-size: medium;">Amazon's "re:Invent 2023" customer conference </span><span style="background-color: white; color: #1a222a;">(Nov 27-Dec 1, Las Vegas, NV)</span> </b></span></span></div><div style="text-align: left;"><span style="font-family: verdana;">Amazon is playing catchup with regards to the deployment of generative AI technology; more specifically, Amazon is playing catchup to Microsoft. </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Generative Ai's large language models (LLMs) must be run in the cloud because of their massive demand for graphics processing units (GPUs), the kind of chips made by Nvidia. Amazon invented cloud services and still has a much larger share of the global market for cloud services than Microsoft. </span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">Amazon's AWS share = 33 percent </span></li><li><span style="font-family: verdana;">Microsoft's Azure share = 23 percent</span></li></ul><div style="text-align: right;"><i>Note: Percentages from Google search for market share Amazon AWS, Microsoft Azure</i> </div><div><span style="font-family: verdana;">How did Microsoft overcome this disadvantage?</span><span style="font-family: verdana;"> Microsoft's Office dominates the global office productivity market with a 49 percent share.</span><span style="font-family: verdana;"> So Microsoft packaged its generative AI services as productivity enhancing "copilots" for the components of its office productivity suite. Enterprise subscribers to Microsoft Office will purchase additional subscriptions for copilots. </span></div><div><span style="font-family: verdana;"><br /></span></div><div><span style="font-family: verdana;">Moreover, at its recent developers conference, Microsoft announced its intention to manufacture its own GPUs, thereby reducing the dependency of its Azure cloud services on Nvidia's limited supply.</span></div></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Given this context, few were surprised by Amazon's announcements at its "re:Invent" conference of an array of generative AI services that it would provide for enterprise customers of its AWS cloud services. Its two most memorable announcement were for its chatbot and its image generator:</span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">Amazon's chatbot is called "Q"</span></li><li><span style="font-family: verdana;">Its image generator is called "Titan Image Generator"</span></li></ul><div><span style="font-family: verdana;">Note that "Titan" is the name of Amazon's LLM. Also note that Amazon had previously announced its intention to produce its own GPU chips called "Trainium". At re:Invent, it announced the development of more powerful "Trainium" chips </span></div><div><span style="font-family: verdana;"><br /><br /></span></div></div><div style="text-align: left;"><b><span style="font-family: verdana; font-size: medium;">2)</span><span style="font-size: medium;"><span style="font-family: verdana;"> </span><span style="background-color: white; color: #1a222a; font-family: verdana;">Google's DeepMind</span></span></b><span><span style="font-size: medium;"><b style="font-family: verdana;"><span style="background-color: white; color: #1a222a;"> AI program called "</span></b><b style="color: #1a1a1a; font-family: BreveText, helvetica, sans-serif; letter-spacing: 0.108px;">GNoME</b></span><span style="font-family: verdana;"><span style="background-color: white;"><span><span style="color: #1a222a; font-size: medium; font-weight: bold;">" identified 2.2 million stable crystal structures </span><br />The second paragraph from the Ars Technica article listed in the "<b>A. Top Stories</b>" section (below) provides a succinct summary of this achievement: </span></span></span></span></div><div style="text-align: left;"><span><span style="font-family: verdana;"><span style="background-color: white;"><blockquote><span>"The trove of theoretically stable but experimentally unrealized combinations identified using an AI tool known as GNoME is more than 45 times larger than the number of such substances unearthed in the history of science, according to a paper published in Nature on Wednesday."</span></blockquote></span></span></span></div><div style="text-align: left;"><span style="font-family: verdana;">If this were the first major breakthrough in science/technology that had been achieved by Google's DeepMind, one would wonder why so many people seem to think that ChatGPT is the most powerful AI system in the world and therefore the most likely forbear of Artificial General Intelligence (AGI) ... but this isn't DeepMind's first breakthrough nor its greatest. Three other recent major breakthroughs are noted below in section <b>"B. Three other breakthroughs .</b>.. ". </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">To date, neither ChatGPT nor any competitor LLM systems have created significant new knowledge about any science or about any technology. DeepMind's success is not a coincidence. Its founding purpose was to do just that: to create AI tools that would greatly assist in the resolution of the world's biggest problems.</span></div><div style="text-align: left;"><br /></div><div style="text-align: left;"><br /></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b>3) </b></span><span style="background-color: white; color: #1a222a; font-family: verdana;"><span style="font-size: medium;"><b>"The Inside Story of Microsoft’s Partnership with OpenAI"</b></span></span></div><div><div id="Overview" style="background-color: white; color: #1a222a; text-align: left;"><span style="font-family: verdana;"><span>This long investigative piece in the New Yorker is a "must read", because it provides a plausible description of Microsoft's side of its partnership with OpenAI.<br /></span><br />Microsoft's management of this partnership reflected the views of <b>Kevin Scott</b>, Microsoft's chief technology officer, the person most responsible for forging the OpenAI partnership. The following quotation from the New Yorker captures the essential features of his vision:</span></div><div id="Overview" style="background-color: white; text-align: left;"><blockquote><span style="font-family: verdana;"><span style="color: black;">"</span>The discourse around A.I., he believed, had been strangely focussed on science-fiction scenarios—computers destroying humanity—and had largely ignored the technology’s potential to “level the playing field,” as Scott put it, for people who knew what they wanted computers to do but lacked the training to make it happen. He felt that A.I., with its ability to converse with users in plain language, could be a transformative, equalizing force—if it was built with enough caution and introduced with sufficient patience. </span></blockquote><blockquote><span style="font-family: verdana;">Scott and his partners at OpenAI had decided to release A.I. products slowly but consistently, experimenting in public in a way that enlisted vast numbers of nonexperts as both lab rats and scientists: Microsoft would observe how untutored users interacted with the technology, and users would educate themselves about its strengths and limitations. By releasing admittedly imperfect A.I. software and eliciting frank feedback from customers, Microsoft had found a formula for both improving the technology and cultivating a skeptical pragmatism among users. The best way to manage the dangers of A.I., Scott believed, was to be as transparent as possible with as many people as possible, and to let the technology gradually permeate our lives—starting with humdrum uses. And what better way to teach humanity to use A.I. than through something as unsexy as a word processor?<span style="color: #1a222a;">"</span></span></blockquote></div><div><span style="font-family: verdana;"><div></div></span></div></div><div style="text-align: left;"><span style="font-family: verdana;">Whereas DeepMind strives to create AI tools that will enable the world's greatest scientists and technologists to address the world's biggest problems, Microsoft's efforts are grounded in AI's potential to level the playing field by their accessibility to all users in plain language.</span></div><div style="text-align: left;"><br /></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div id="Overview" style="background-color: white; color: #1a222a; font-family: Verdana, Geneva, sans-serif; font-size: 13.86px;"><span style="font-family: verdana;"><b style="font-size: large;">A. Top 3 stories in past week ...</b></span></div><div id="Overview" style="background-color: white; color: #1a222a; font-family: Verdana, Geneva, sans-serif; font-size: 13.86px;"><ol><li style="margin: 0px 0px 0.25em; padding: 0px;"><span style="font-size: medium;"><span style="font-family: verdana;"><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#OtherLLMs" style="color: #336699; font-size: small; font-weight: 700; text-align: center; text-decoration-line: none;"><span style="color: #cc0000;">Other LLMs</span></a><br />Amazon's re:Invent 2023 conference (Nov 27-Dec 1, Las Vegas, NV) <b><span style="color: #cc0000;">***</span></b> <br />-- Overview "Here’s everything Amazon Web Services announced at AWS re:Invent", Christine Hall, <b><i><a href="https://techcrunch.com/2023/11/29/heres-everything-aws-reinvent-2023-so-far/" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">TechCrunch</span></a></i></b>, 11/29/23<br /><br />-- Amazon Q, an AI-powered chatbot for AWS customers ..."Amazon Introduces Q, an A.I. Chatbot for Companies", Karen Weise, <b><i><a href="https://www.nytimes.com/2023/11/28/technology/amazon-ai-chatbot-q.html?smid=nytcore-android-share" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">NY Times</span></a></i></b>, 11/28/23 ... This story also covered by <b><i><a href="https://t.co/e0A5hemJCe" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Bloomberg</span></a></i></b>, <b><i><a href="https://t.co/mZte0D6THp" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">VentureBeat</span></a></i></b>, <b><i><a href="https://twitter.com/Gizmodo/status/1729609880909353016/photo/1" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Gizmodo</span></a></i></b>, <br /><br />-- Image generator "Amazon joins AI image creation fray with new model", Emilia David, <b><i><a href="https://t.co/Z3rkZmVtdm" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">The Verge</span></a></i></b>, 11/29/23 </span>...</span><span style="font-family: verdana;"><span style="font-size: medium;"> This story also covered by <a href="https://news.bloomberglaw.com/ip-law/amazon-has-developed-its-own-version-of-an-ai-image-generator" style="color: #336699; font-style: italic; font-weight: bold; text-decoration-line: none;"><span style="color: #cc0000;">Bloomberg</span></a><b style="font-style: italic;">, </b><b><i><a href="https://t.co/SKtNBenFfr" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Engadget</span></a></i></b><b style="font-style: italic;">, <a href="https://t.co/5tJTiNexUJ" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Gizmodo</span></a>, </b></span><br /><br /></span></li><li style="margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana; font-size: medium;"><span style="font-size: small;"> </span><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#NonLLMNews" style="color: #336699; font-size: small; font-weight: 700; text-align: center; text-decoration-line: none;"><span style="color: #cc0000;">Non-LLM News</span></a><br />"Google’s DeepMind finds 2.2M crystal structures in materials science win", Michael Peel, <b><i><a href="https://arstechnica.com/ai/2023/11/googles-deepmind-finds-2-2m-crystal-structures-in-materials-science-win/" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Ars Technica</span></a></i></b>, 11/29/23 ... "The trove of theoretically stable but experimentally unrealized combinations identified using an AI tool known as GNoME is more than 45 times larger than the number of such substances unearthed in the history of science, according to a paper published in Nature on Wednesday."<br />-- This story also covered by <b><i><a href="https://www.wired.com/story/an-ai-dreamed-up-380000-new-materials-the-next-challenge-is-making-them/" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Wired</span></a></i></b>, <b><i><a href="https://venturebeat.com/ai/google-deepminds-materials-ai-has-already-discovered-2-2-million-new-crystals/" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">VentureBeat</span></a></i></b>, <b><i><a href="https://www.nature.com/articles/d41586-023-03745-5" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Nature</span></a></i></b>, <b><i><a href="https://www.science.org/content/article/materials-predicting-ai-deepmind-could-revolutionize-electronics-batteries-and-solar" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Science</span></a></i></b>,<br /><br /></span></li><li style="margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana; font-size: medium;"><b style="font-size: small; text-align: center;"><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#OpEds" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">OpEds</span></a></b><br />"The Inside Story of Microsoft’s Partnership with OpenAI", Charles Duhigg, <b><i><a href="https://www.newyorker.com/magazine/2023/12/11/the-inside-story-of-microsofts-partnership-with-openai" style="color: #ff8832;"><span style="color: #cc0000;">The New Yorker</span></a></i></b>, 12/1/23 <b><i><span style="color: #cc0000;">***</span></i></b> <i><br /></i><br /></span></li></ol></div><div style="text-align: left;"><span style="font-family: verdana;"><div style="font-family: Times;"><div><span style="font-family: verdana;"><span style="font-size: medium;"><b>B. Three other breakthroughs in science/technology using DeepMind's AI/Neural Networks</b></span></span></div><div><span style="font-family: verdana;"><span><div style="background-color: white; color: #1a222a; font-family: Verdana, Geneva, sans-serif;"><ul style="line-height: 1.4; list-style-image: initial; list-style-position: initial; margin: 0.5em 0px; padding: 0px 2.5em;"><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;">"One of the Biggest Problems in Biology Has Finally Been Solved", Tanya Lewis, <b><i><a href="https://www.scientificamerican.com/article/one-of-the-biggest-problems-in-biology-has-finally-been-solved/" style="color: #336699; text-decoration-line: none;">Scientific American</a></i></b>, 10/31/22 ... protein foldimg problem ...<i>DeepMind</i><br />-- This story also covered by <b><i><a href="https://venturebeat.com/ai/a-year-ago-deepminds-alphafold-ai-changed-the-shape-of-science-but-there-is-more-work-to-do/" style="color: #336699; text-decoration-line: none;">VentureBeat</a></i></b>, <b><i><a href="https://www.cnet.com/science/biology/googles-deepmind-ai-predicts-3d-structure-of-nearly-every-protein-known-to-science/" style="color: #336699; text-decoration-line: none;">CNET</a></i></b>, <b><i><a href="https://www.science.org/content/article/google-s-deepmind-aces-protein-folding" style="color: #336699; text-decoration-line: none;">Science</a></i></b>, <br /><br /></span></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;">"Courtesy of AI: Weather forecasts for the hour, the week and the century", Devin Coldewey, <b><i><a href="https://t.co/7szMYJTtMu" style="color: #336699; text-decoration-line: none;">TechCrunch</a></i></b>, 11/14/23 ... <i style="color: black;">DeepMind</i><br />-- This story also covered by <b><i><a href="https://t.co/ZVUWEP6aUh" style="color: #336699; text-decoration-line: none;">Gizmodo</a></i></b>, <b><i><a href="https://www.ft.com/content/ca5d655f-d684-4dec-8daa-1c58b0674be1" style="color: #336699; text-decoration-line: none;">Financial Times</a></i></b>, <b><i><a href="https://www.technologyreview.com/2023/11/14/1083366/google-deepminds-weather-ai-can-forecast-extreme-weather-quicker-and-more-accurately/" style="color: #336699; text-decoration-line: none;">MIT Tech Review</a></i></b>, <b><i><a href="https://arstechnica.com/science/2023/11/ai-outperforms-conventional-weather-forecasting-for-the-first-time-google-study/" style="color: #336699; text-decoration-line: none;">Ars Technica</a></i></b>, <b><i><a href="https://www.wired.com/story/google-deepmind-ai-weather-forecast/" style="color: #336699; text-decoration-line: none;">Wired</a></i></b>, <b><i><a href="https://t.co/2P7gFHA3Ut" style="color: #336699; text-decoration-line: none;">ZDNet</a></i></b>, <br />-- These articles describe research published in <b><i><a href="https://www.nature.com/articles/d41586-023-03552-y" style="color: #336699; text-decoration-line: none;">Nature</a><b> </b><span style="font-style: normal; font-weight: 400;">and</span><b> <a href="https://www.science.org/doi/10.1126/science.adi2336" style="color: #336699; text-decoration-line: none;">Science</a><br /><br /></b></i></b></span></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;">"AI Tool Pinpoints Genetic Mutations That Cause Disease", Even Callaway & Nature Magazine, <b><i><a href="https://www.scientificamerican.com/article/ai-tool-pinpoints-genetic-mutations-that-cause-disease/" style="color: #336699; text-decoration-line: none;"><span style="color: black;">Scientific American</span><span style="color: #cc0000;">,</span></a></i></b> 11/21/23 ... <i style="color: black;">DeepMind</i><br />-- This story also covered by<span style="color: black;"> </span><span style="color: black;"><a href="https://www.wsj.com/podcasts/tech-news-briefing/how-googles-deepmind-is-using-ai-to-uncover-the-causes-of-diseases/a0a6e750-b1f0-4265-a675-d00129059e7a#:~:text=Trying%20to%20determine%20who%20might,mutations%20that%20could%20cause%20illness" style="color: #336699; font-style: italic; font-weight: bold; text-decoration-line: none;">Wall Street Journal</a><b style="font-style: italic;">, </b><a href="https://www.wired.com/story/deepmind-ai-alphamissense-genetics-rare-diseases/" style="color: #336699; text-decoration-line: none;"><b><i>Wired</i></b></a><b style="font-style: italic;">, </b></span><span style="color: black;"><br /></span></span></li></ul></div></span></span></div><div><span style="font-family: verdana;"><span style="font-size: medium;"><b><br /></b></span></span></div><div><span style="font-family: verdana;"><span style="font-size: medium;"><b>C. Links to ChatGPT's responses to five basic AI questions:</b> </span> </span></div><div><span style="font-family: verdana;">ChatGPT's answer to each question provides the framework for its response to the next question. So most readers should start with Question 1.</span></div><div><span style="font-family: verdana;"><div style="color: #1a222a; font-size: 13.86px;"><ul style="line-height: 1.4; list-style-image: initial; list-style-position: initial; margin: 0.5em 0px; padding: 0px 2.5em;"><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><b>Question 1: <a href="https://neoskeptics.blogspot.com/2023/11/what-is-artificial-intelligence.html" style="color: #336699; text-decoration-line: none;">What is Artificial Intelligence?</a><br /><br /></b></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><b><h3 class="post-title entry-title" itemprop="name" style="font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-feature-settings: normal; font-kerning: auto; font-optical-sizing: auto; font-size: 18px; font-stretch: normal; font-variant-alternates: normal; font-variant-east-asian: normal; font-variant-numeric: normal; font-variant-position: normal; font-variation-settings: normal; line-height: normal; margin: 0px; position: relative;">Question 2: <a href="https://neoskeptics.blogspot.com/2023/11/question-2-what-is-machine-learning.html" style="color: #336699; text-decoration-line: none;">What is Machine Learning?</a></h3> </b></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-size: medium;"><b><h3 class="post-title entry-title" itemprop="name" style="font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-feature-settings: normal; font-kerning: auto; font-optical-sizing: auto; font-size: 18px; font-stretch: normal; font-variant-alternates: normal; font-variant-east-asian: normal; font-variant-numeric: normal; font-variant-position: normal; font-variation-settings: normal; line-height: normal; margin: 0px; position: relative;">Question 3: <a href="https://neoskeptics.blogspot.com/2023/11/question-4-what-are-neural-networks.html" style="color: #336699; text-decoration-line: none;">What are Neural Networks?</a> </h3><br /></b></span></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><h3 class="post-title entry-title" itemprop="name" style="font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-feature-settings: normal; font-kerning: auto; font-optical-sizing: auto; font-size: 18px; font-stretch: normal; font-variant-alternates: normal; font-variant-east-asian: normal; font-variant-numeric: normal; font-variant-position: normal; font-variation-settings: normal; line-height: normal; margin: 0px; position: relative;">Question 4: <a href="https://neoskeptics.blogspot.com/2023/11/question-5-what-are-gpt-models-chatgpts.html" style="color: #336699; text-decoration-line: none;">What are GPT models? </a><br /><br /></h3></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-size: medium;"><h3 class="post-title entry-title" itemprop="name" style="font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-feature-settings: normal; font-kerning: auto; font-optical-sizing: auto; font-size: 18px; font-stretch: normal; font-variant-alternates: normal; font-variant-east-asian: normal; font-variant-numeric: normal; font-variant-position: normal; font-variation-settings: normal; line-height: normal; margin: 0px; position: relative;">Question 5: <a href="https://neoskeptics.blogspot.com/2023/11/question-5-have-models-other-than-gpt.html" style="color: #336699; text-decoration-line: none;">Have models other than GPT made substantial progress towards AGI?</a> </h3></span></li></ul></div></span></div><div><span style="font-family: verdana;"> </span></div></div></span></div>
Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-39717236307487881432023-11-26T13:34:00.015-05:002024-03-27T19:28:48.313-04:00"The AI Big Bang" a/k/a The fall and rise of Sam Altman ... 26Nov23 TL;DR and podcast<div style="text-align: right;"><span style="font-family: verdana; font-size: x-small;"><i> Last update: Sunday 11/26/23 </i></span></div><div style="text-align: left;"><span style="font-family: verdana;"><div class="separator" style="clear: both; text-align: center;"><div style="text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><div class="separator" style="clear: both; font-size: x-large; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhzW7Ky_mXqMeYSpwV_v1apIIJw6KfAyLc1z1mVHjS0XgyrYjqxSdAEKa-xx3W2pdX8vJRqO6vh661MPlukO_mLF5d9BFntXfzLPqI1mvFlDL0CMojxFiuAl9XQR_jHm598AKDHJMRu3V5qM13jKAq4rU45dTVnd0381Xrvle10B9hu2HpwyEQENXZe0sU/s2508/BigBangStarsExpandingGalaxy.jpeg" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="1672" data-original-width="2508" height="213" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhzW7Ky_mXqMeYSpwV_v1apIIJw6KfAyLc1z1mVHjS0XgyrYjqxSdAEKa-xx3W2pdX8vJRqO6vh661MPlukO_mLF5d9BFntXfzLPqI1mvFlDL0CMojxFiuAl9XQR_jHm598AKDHJMRu3V5qM13jKAq4rU45dTVnd0381Xrvle10B9hu2HpwyEQENXZe0sU/s320/BigBangStarsExpandingGalaxy.jpeg" vspace="10" width="320" /></a></div><span style="font-size: medium;"><span><div style="text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><span style="font-size: medium;"><span><div>Welcome to our TL;DR summary of and podcast about the past week's top two AI stories on our "<b><i><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html">Useful AI News</a></i></b>" page ...</div><div>(1) What happened? (2) Why did it happen?</div></span></span></span></span></span></div></span></span></span></span></span></div><div style="text-align: left;"><span><span style="font-size: large;"><span style="background-color: white; color: #1a222a;">... </span></span><span style="background-color: white; color: #1a222a;"><span style="font-size: medium;">Click the "start" button in the audio control (below) to hear the podcast</span><span><span style="font-size: large;">. </span></span></span></span></div><div style="background-color: white; color: #1a222a; font-size: 13.86px; text-align: left;"><span style="font-family: verdana; font-size: medium;"></span></div><div style="background-color: white; text-align: right;"><span style="color: #1a222a; font-family: verdana; font-size: medium;"> <iframe height="50" src="https://drive.google.com/file/d/1KNbYcoKoAGnmsEWGAP_qUr_6nzQlAho0/preview?usp=sharing" width="400"></iframe></span><div class="separator" style="clear: both; text-align: left;"><span face="Verdana, Geneva, sans-serif" style="color: #1a222a; font-size: medium;">TL;DR link<b> </b></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-size: large; text-decoration-line: none;"><span style="color: #1a222a;">➡</span></a><span style="color: #1a222a;"> </span><b><i><a href="https://neoskeptics.blogspot.com/2023/11/the-ai-big-bang-aka-fall-and-rise-of.html#more"><span style="color: #cc0000;">HERE</span></a></i></b></div></div></div></span></div><span><span style="font-family: verdana;"><a name='more'></a></span></span><div style="text-align: left;"><p><span style="background-color: white; color: #1a222a; font-family: verdana;"><b>1) What happened?</b><br /></span></p><p></p><ul style="text-align: left;"><li><span style="background-color: white; color: #1a222a; font-family: verdana;">The tumultuous long weekend began on Friday evening 11/17/23 when the OpenAI board abruptly notified Sam Altman that he was fired.<br /><br /></span></li><li><span style="background-color: white; color: #1a222a; font-family: verdana;">The board then appointed an outsider to become CEO<br /><br /></span></li><li><span style="background-color: white; color: #1a222a; font-family: verdana;">Microsoft's CEO Satya Nadella announced that he was going to hire Altman and other OpenAI staff<br /><br /></span></li><li><span style="background-color: white; color: #1a222a; font-family: verdana;">Hundreds of OpenAI staff announced that they would quit OpenAi and go to work for Microsoft unless Altman was rehired as CEO and the board was fired<br /><br /></span></li><li><span style="background-color: white; color: #1a222a; font-family: verdana;">On Monday 11/20/23, Nadella announced that Altman might return to OpenAI<br /><br /></span></li><li><span style="background-color: white; color: #1a222a; font-family: verdana;">On Tuesday evening 11/21/23, Altman returned to OpenAI as CEO; three out of the four members of the board were dismissed; and two new outsiders were appointed to the board </span></li></ul><p><span style="font-family: verdana;"><span style="background-color: white; color: #1a222a;">2) <b>Why did it happen</b>?<br /></span><span style="background-color: white; color: #1a222a;">Altman's abrupt dismissal embodied an inevitable seismic collision between altruistic "AI Doomers" on OpenAi's board (called "Doomers" because they feared that the pursuit of artificial general intelligence (AGI) might end in a catastrophic doomsday) vs. Sam Altman, the ultimate Silicon Valley, </span><span style="background-color: white; color: #1a222a;">profit driven,</span><span style="background-color: white; color: #1a222a;"> optimistic "AI Boomer". In its dismissal notice the OpenAI board said that it no longer trusted Sam Altman, but it did not say why.</span></span></p><p><span style="background-color: white; color: #1a222a; font-family: verdana;">However, various observers have identified three issues that might have caused the board to become dissatisfied with Mr. Altman's performance, given the board's primary obligation to ensure that OpenAI developed artificial general intelligence (AGI) in a safe manner and in a manner that did not cede control of its AGI to private corporations.</span></p><p></p><ul style="text-align: left;"><li><span style="font-family: verdana;"><span>The </span><b><i><a href="https://www.washingtonpost.com/technology/2023/11/22/sam-altman-fired-y-combinator-paul-graham/"><span style="color: #cc0000;">Washington Post</span></a> </i></b><span>reported that "Four years ago, one of Altman’s mentors, Y Combinator founder Paul Graham, flew from the United Kingdom to San Francisco to give his protégé the boot, according to three people familiar with the incident, which has not been previously reported."<br /><br />"Graham had surprised the tech world in 2014 by tapping Altman, then in his 20s, to lead the vaunted Silicon Valley incubator. Five years later, he flew across the Atlantic with concerns that the company’s president put his own interests ahead of the organization — worries that would be echoed by OpenAI’s board."<br /></span><span><br /></span></span></li><li><span style="font-family: verdana;"><span><a href="https://venturebeat.com/ai/sam-altmans-return-to-openai-highlights-urgent-need-for-trust-and-diversity/" style="font-style: italic; font-weight: bold;"><span style="color: #cc0000;">VentureBeat</span></a><b style="font-style: italic;"> </b>reported that</span><span> Altman had criticized an academic paper published by board member Helen Toner </span><span>in October 2023<br /></span><i><br /></i> "In the paper, Toner, a director of strategy at Georgetown University’s Center for Security and Emerging Technology, included a three-page section that was a detailed and earnest account of the way OpenAI and a major competitor Anthropic approached the release of their latest large language models (LLMs) in March of 2023. OpenAI chose to release its model, in contrast with Anthropic, which chose to delay its model, called Claude, because of concerns about safety ..."<br /><br />"Senior OpenAI leaders then discussed whether Toner should be removed, but co-founder Ilya Sutskever, who was deeply concerned about the risks of AI technology, sided with other board members to oust Altman for not being “consistently candid in his communications with the board."<br /><br />"All of this came after some previous board frustrations with Altman about his moving too quickly on the product side, with other accounts suggesting that the company’s recent DevDay was also a major frustration for the board."<br /><span style="background-color: white; color: #1a222a;"><br /></span></span></li><li><span style="font-family: verdana;"><span style="background-color: white;"> <span style="color: #1a222a;"><a href="https://t.co/S79jedzvDY" style="color: #336699; font-style: italic; font-weight: bold; text-decoration-line: none;"><span style="color: #cc0000;">Reuters</span></a><b style="font-style: italic;"> </b>reported that</span></span><span style="background-color: white; color: #1a222a;"> OpenAi researchers had recently achieved an as yet unannounced major breakthrough in their quest for AGI. If so, then the terms of Microsoft's $13 billion investment in OpenAI that had been negotiated by Altman might give it immediate access to this new technology in a direct violation of OpenAI's defining mission.</span></span></li></ul><p></p></div><div style="text-align: left;"><span style="font-family: verdana;"><b>3) Who won? Who lost?</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">Satya Nadella and Microsoft won. (Indeed, Microsoft's stock hit an all-time high after Altman's return was announced.) Sam Altman and the OpenAI staff won. The three board members who were dismissed lost, and Open AI lost. OpenAI lost because it failed to pursue its mission. OpenAI is now a new organization that is a profit-oriented operation beholden to Microsoft and to other investors. In short, the pessimistic, altruistic "AI Doomers" lost, and the optimistic, profit-driven "AI boomers" won.</span></div><div style="text-align: left;"><br /></div><div style="text-align: left;"><div id="Overview" style="background-color: white; color: #1a222a; font-family: Verdana, Geneva, sans-serif;"><span style="font-family: verdana;"><b>Top 2 stories in past week ...</b></span></div><div id="Overview" style="background-color: white; color: #1a222a; font-family: Verdana, Geneva, sans-serif;"><ol><li style="margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;"><b><b style="text-align: center;"><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#OpenAI" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000; font-size: x-small;">OpenAI</span></a></b><br /><span>AI Big Bang 1 -- What happened? </span></b></span><span><b style="font-family: verdana;"> </b><span style="color: #cc0000; font-family: verdana;">***</span><b style="font-family: verdana;"> </b><br /><span style="font-family: verdana;">"OpenAI Ousts CEO Sam Altman", Will Knight, <b><i><a href="https://t.co/s4RBLzVqjj" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Wired</span></a></i></b>, 11/17/23 </span><span style="font-family: verdana;"><br /><br />"Sam Altman’s Sudden Exit Sends Shockwaves Through OpenAI and Beyond", Will Knight, <a href="https://t.co/zPM24WiY5P" style="color: #336699; text-decoration-line: none;"><b><i><span style="color: #cc0000;">Wired</span></i></b></a>, 11/18/23</span><br /><span style="color: #cc0000; font-family: verdana;"><b><i><br style="color: black; font-family: Times; font-style: normal; font-weight: 400;" /><span style="color: black; font-style: normal; font-weight: 400;"><span style="color: #1a222a;">"Meet Emmett Shear, OpenAI’s ‘Highly Intelligent, Socially Awkward’ Interim CEO", Anit katwala and Will Bedingfield, </span><b><i><a href="https://t.co/uCtM2G3IZo" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Wired</span></a></i></b><span style="color: #1a222a;">, 11/20/23</span></span><br /><br /></i></b></span><span style="font-family: verdana;"> "Microsoft Hires Sam Altman Hours After OpenAI Rejects His Return", Mike Isaac, Kevin Roose and Cade Metz, <b><i><a href="https://www.nytimes.com/2023/11/20/technology/openai-altman-ceo-not-returning.html?smid=nytcore-android-share" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">NY Times</span></a></i></b>, 11/20/23<br /></span><br /><b style="font-family: verdana;"> </b><span style="font-family: verdana;">"Hundreds of OpenAI employees threaten to resign and join Microsoft"</span><b style="font-family: verdana;">, </b><span style="font-family: verdana;">Jacob Kastrenakes and Tom Warren, <b><i><a href="https://www.theverge.com/2023/11/20/23968988/openai-employees-resignation-letter-microsoft-sam-altman" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">The Verge</span></a></i></b>, 11/20/23</span><span style="font-family: verdana;"> </span><span style="font-family: verdana;"><br /></span><br /><span style="font-family: verdana;">"Microsoft CEO Satya Nadella suggests that Sam Altman might return to OpenAI", Kyle Wiggers, <b><i><a href="https://techcrunch.com/2023/11/20/microsoft-ceo-satya-nadella-suggests-that-sam-altman-might-return-to-openai/" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">TechCrunch</span></a></i></b>, 11/20/23</span><br /><br /></span><span style="font-family: verdana;"><span>"Sam Altman Is Reinstated as OpenAI’s Chief Executive", Cade Metz, Mike Isaac, Tripp Mickle, Karen Weise and Kevin Roose, <b><i><a href="https://www.nytimes.com/2023/11/22/technology/openai-sam-altman-returns.html?smid=nytcore-android-share" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">NY Times</span></a></i></b>, 11/22/23<br /></span><br /></span></li><li style="margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;"><b><b style="text-align: center;"><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#OpenAI" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000; font-size: x-small;">OpenAI</span></a></b><br /><span>AI Big Bang 2 -- Why did it happen? </span></b><span><b> </b><span style="color: #cc0000;">***</span><b> <br /></b>Various observers have identified three issues that might have caused the board to become dissatisfied with Mr. Altman's performance, given the board's primary obligation to ensure that OpenAI developed artificial general intelligence (AGI) in a safe manner and in a manner that did not cede control of the development of AGI to a private corporation.<br /><br />-- "Altman’s polarizing past hints at OpenAI board’s reason for firing him", Elizabeth Dwoskin and Nitasha Tiku, <b><i><a href="https://www.washingtonpost.com/technology/2023/11/22/sam-altman-fired-y-combinator-paul-graham/" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Washington Post</span></a></i></b>, 11/22/23 ... Key insight = Altman was fired from Y Combinator because he out his own interest ahead of the company's interest,<br /><br />-- "Sam Altman’s return to OpenAI highlights urgent need for trust and diversity", Matt Marshall, <b><i><a href="https://venturebeat.com/ai/sam-altmans-return-to-openai-highlights-urgent-need-for-trust-and-diversity/" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">VentureBeat</span></a></i></b>, 11/22/23 ... Key insight = Altman’s criticism of board member Helen Toner’s work on AI safety <i>[in October 2023]<br /></i><br />-- "OpenAI researchers warned board of AI breakthrough ahead of CEO ouster, sources say", Anna Tong, Jeffrey Dastin and Krystal Hu, <b><i><a href="https://t.co/S79jedzvDY" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">Reuters</span></a></i></b>, 11/23/23... Key insight = OpenAi researchers had recently achieved an as yet unannounced breakthrough in their quest for AGI. If so, then the terms of Microsoft's investments negotiated by Altman might give it immediate access to this new technology, thereby giving Microsoft an unchallengeable advantage ... a direct violation of OpenAI's defining mission.</span></span></li></ol></div><div id="Overview" style="background-color: white; color: #1a222a; font-family: Verdana, Geneva, sans-serif; text-align: center;"></div></div><div style="text-align: left;"><span style="font-family: verdana;">____________________________________</span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b>Links to related notes on this blog: </b></span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana; font-size: medium;"><b><br /></b></span></li></ul></div>Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-69256833390755026182023-11-19T14:12:00.019-05:002024-03-27T19:30:49.018-04:00Microsoft's dev con, Google's GraphCast, and Sam Altman's farewell ... TL;DR + Podcast 19Nov23<div style="text-align: right;"><span style="font-family: verdana; font-size: x-small;"><i>Last update: <span style="color: #cc0000;"><b>Monday 11/20/23</b></span> </i></span></div><div style="text-align: left;"><span style="font-family: verdana;"><div class="separator" style="clear: both; text-align: center;"><div style="text-align: left;"><span><span><span style="background-color: white; color: #1a222a;"><span style="font-size: medium;"><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj4iAC9DH4R2uRI-boa5TtXtBIFURER4gQFxhe16pmYvcNYBy_FXjpfa0vTSBeDLkGDLzawQvXJkxOxiEZBMgsgJ0qW73GLI_HTYjrmAmEqNvDGzHyIp6V-fPi_fr0G67bxMoR43vRDd-tcpUwK117dqoAy_9ZOMMfTTXbK3R1NhyrsrNJ2T3_V3q4KrYc/s1000/CEONadellaAtMicrosoftIgniteDevCon23.webp" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="667" data-original-width="1000" height="213" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj4iAC9DH4R2uRI-boa5TtXtBIFURER4gQFxhe16pmYvcNYBy_FXjpfa0vTSBeDLkGDLzawQvXJkxOxiEZBMgsgJ0qW73GLI_HTYjrmAmEqNvDGzHyIp6V-fPi_fr0G67bxMoR43vRDd-tcpUwK117dqoAy_9ZOMMfTTXbK3R1NhyrsrNJ2T3_V3q4KrYc/s320/CEONadellaAtMicrosoftIgniteDevCon23.webp" vspace="10" width="320" /></a></div>Welcome to our </span></span></span></span><span style="font-size: medium;"><span style="background-color: white; color: #1a222a;">19Nov23</span><span style="background-color: white; color: #1a222a;"> </span><span style="background-color: white; color: #1a222a;">TL;DR summary + podcast about the past week's top 3 AI stories on our "</span><b style="color: #1a222a;"><i><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html">Useful AI News</a></i></b><span style="background-color: white; color: #1a222a;">" page:<br /></span></span><span style="font-size: medium;"><span style="background-color: white; color: #1a222a;">1) Microsoft's conference for developers and IT professionals, (2) Google Deep Mind's awesome AI weather forecasts, and (3) the sudden seismic firing of Sam Altman by OpenAI's board </span></span><span style="font-size: large;"><span style="background-color: white; color: #1a222a;">... </span></span><span style="background-color: white; color: #1a222a;"><span style="font-size: medium;">Click the "start" button in the audio control (below) to hear the podcast</span><span style="font-size: large;">.</span></span></div><div style="background-color: white; color: #1a222a; font-family: Verdana, Geneva, sans-serif; font-size: 13.86px; text-align: left;"><span style="font-family: verdana; font-size: medium;"></span></div><div style="background-color: white; color: #1a222a; font-family: Verdana, Geneva, sans-serif; text-align: right;"><span style="font-family: verdana; font-size: medium;">
<iframe height="50" src="https://drive.google.com/file/d/1KDNwg6bW9T3gr7wxYu7H7uVQgs8JD2qN/preview?usp=sharing" width="400"></iframe></span><div class="separator" style="clear: both; text-align: left;"><div style="color: black; font-family: Times;"><br /></div></div><div class="separator" style="clear: both; text-align: left;"><span face="Verdana, Geneva, sans-serif" style="font-family: verdana; font-size: medium;">TL;DR link<b> </b></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-family: verdana; font-size: large; text-decoration-line: none;"><span style="color: #1a222a;">➡</span></a><span style="color: #1a222a; font-family: verdana;"> </span><b style="color: black; font-family: verdana;"><i><a href="https://neoskeptics.blogspot.com/2023/11/microsofts-dev-con-googles-graphcast.html#more"><span style="color: #cc0000;">HERE</span></a></i></b></div><div class="separator" style="clear: both; text-align: left;"><b style="color: black; font-family: verdana;"><br /></b></div></div></div></span></div><span><span style="font-family: verdana;"><a name='more'></a></span></span><div style="text-align: left;"><div><span style="font-family: verdana;">Large language models (LLMs) have dominated the headlines and revitalized the tech sector in recent months, as per our first story about Microsoft's conference</span></div><div><span style="font-family: verdana;"><br /></span></div><div><span style="font-family: verdana;">However, our second story reminds us that AI is much broader than chatbots that produce useful, but buggy first drafts of solutions. Another kind of AI, more specifically, another kind of neural network provides tools that yield accurate solutions to the world's most complex scientific problems.</span></div><div><span style="font-family: verdana;"><br /></span></div><div><span style="font-family: verdana;">Our third story relays the breathless Friday afternoon breaking news headlines that OpenAI's board of directors had abruptly fired its superstar CEO, Sam Altman, because the board had determined that it could no longer trust him</span></div></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">1. <b>Microsoft's most important announcements at its conference for developers and IT professionals </b></span><span style="font-family: verdana;">(Nov 14-17, Seattle, WA)</span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="background-color: white; color: #1a222a; font-family: verdana;">Rename "Bing Chat" to "Microsoft Copilot"</span></li><li><span style="background-color: white; color: #1a222a; font-family: verdana;">Build its own Ai chips rather than buy AI chips from Nvidia</span></li><li><span style="background-color: white; color: #1a222a; font-family: verdana;">Enable Microsoft's Azure cloud to support open source LLMs, e.g., Facebook's LLaMa</span></li></ul><div><span style="color: #1a222a; font-family: verdana;">These three initiatives provide Microsoft with strategic independence. The first removes allusions to OpenAi's ChatGPT and facilitates the possibility that Microsoft might replace Chat GPT's LLM with its own LLM; the second guarantees Microsoft with AI chips in a market where its competitors will be competing for a limited supply from Nvidia; and the third enables Microsoft to earn income from competitors who use open source LLMs. </span></div></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><b>2. Google DeepMind's awesome AI weather forcaster</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">The best weather forecasts have traditionally been produced by complex physics-based models that ran on supercomputers. Google DeepMind's new machine learning "GraphCast" model uses current weather patterns to predict future weather patterns. What's amazing is that </span><span style="font-family: verdana;">GraphCast</span><span style="font-family: verdana;"> runs on desktop computers, but produces forecasts that are more accurate than the forecasts made by the traditional models running on supercomputers ... :-)</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><b>3. Sam Altman's farewell</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">On Friday 17 November 2023, </span><span style="font-family: verdana;">OpenAI's board issued a terse statement that it had conducted “a deliberative review process” and that Altman “was not consistently candid in his communications with the board, hindering its ability to exercise its responsibilities. The board no longer has confidence in his ability to continue leading OpenAI” </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">The board provided no specifics, so this is an ongoing story; hence most of the many news reports cited in our "Top stories" section should be regarded as unfounded speculations. </span></div><div><span style="font-family: verdana;"><br /></span></div><div><span style="font-family: verdana;">However, on Sunday, November 19th, a senior tech reporter for the New York Times conveyed rumors from well connected sources that Altman and OpenAI's board were currently involved in discussions about his possible return. </span><span style="font-family: verdana;">So this is an ongoing story.</span><span style="font-family: verdana;"> </span></div><div><span style="font-family: verdana;"><br /></span></div><div><span style="font-family: verdana;"><span>On Monday morning (11/20/23), the NY Times announced that Microsoft had hired Altman <br /> </span></span></div><div><span style="font-family: verdana;"><span>On Monday morning (11/20/23), </span></span><span style="font-family: verdana;">Hundreds of OpenAI employees threatened to resign and join Microsoft</span></div><div><span style="font-family: verdana;"><span style="font-size: large;"><br /></span></span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><div id="Overview"><span><span style="font-family: verdana;"><b style="font-size: large;">Top 3 stories in past week ...</b><ol><li><span style="font-family: verdana;"><span style="font-family: verdana;"><b style="font-size: small; text-align: center;"><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#OpenAI"><span style="color: #cc0000;">OpenAI/Microsoft</span></a></b><br /><span style="font-size: medium;">Microsoft's Ignite 2023 conference for developers and IT professionals (Nov 14-17, Seattle, WA)</span><b><span style="color: #cc0000;">***</span></b><span style="font-size: medium;"> </span><br />-- Overview: "Microsoft Ignite 2023: Copilot AI expansions, custom chips and all the other announcements", <b><i><a href="https://t.co/8Rq4pok5eF"><span style="color: #cc0000;">TechCrunch</span></a></i></b><span style="font-size: medium;">, 11/15 & 16/23 ... This page contains TechCrunch summaries of </span><span style="font-size: medium;">every</span><span style="font-size: medium;"> announcement made by Microsoft. Reports about the most important announcements were also published by:</span><br /><span style="font-size: medium;">-- Rename "Bing Chat" to "Microsoft Copilot": </span><b><i><a href="https://t.co/KsgiaoVBSx"><span style="color: #cc0000;">The Verge</span></a>, <a href="https://arstechnica.com/information-technology/2023/11/bing-chat-is-now-microsoft-copilot-in-potentially-confusing-rebranding-move/"><span style="color: #cc0000;">Ars Technica</span></a>, <a href="https://www.zdnet.com/article/bing-chat-now-goes-by-copilot-and-feels-a-lot-more-like-chatgpt/"><span style="color: #cc0000;">ZDNet</span></a>, </i></b><b><i> <span style="font-size: medium;"><a href="https://t.co/p85TSzODMM"><span style="color: #cc0000;">Engadget</span></a>,</span></i></b><br /><span style="font-size: medium;">-- Custom Ai chips: </span><b><i><a href="https://t.co/opxWOeGciL"><span style="color: #cc0000;">The Verge</span></a></i></b><span style="font-size: medium;">, </span><b><i><a href="https://fortune.com/2023/11/16/microsoft-satya-nadella-semiconductor-chips-nvidia-chatgpt-openai/"><span style="color: #cc0000;">Fortune</span></a></i></b><span style="font-size: medium;">, </span><b><i><a href="https://www.cnbc.com/2023/11/15/microsoft-reveals-maia-ai-processor-and-cobalt-arm-based-chip.html"><span style="color: #cc0000;">CNBC</span></a></i></b><span style="font-size: medium;">, </span><b><i><a href="https://www.cnbc.com/2023/11/15/microsoft-reveals-maia-ai-processor-and-cobalt-arm-based-chip.html"><span style="color: #cc0000;">Financial Times</span></a></i></b><span style="font-size: medium;">, </span><b><i><a href="https://www.zdnet.com/article/microsoft-unveils-first-ai-chip-maia-100-and-cobalt-cpu/"><span style="color: #cc0000;">ZDNet</span></a></i></b><span style="font-size: medium;">, </span><b><i><a href="https://www.bloomberg.com/news/articles/2023-11-15/microsoft-unveils-its-first-custom-designed-ai-cloud-chips"><span style="color: #cc0000;">Bloomber<span style="font-size: medium;">g</span></span></a></i></b><span style="font-size: medium;"><span>, </span><a href="https://t.co/0DEgtbGhUh" style="font-style: italic; font-weight: bold;"><span style="color: #cc0000;">Gizmodo</span></a><b><i>, </i><a href="https://t.co/OIYJOiuyQR"><i><span style="color: #cc0000;">VentureBeat</span></i></a><i>, </i></b></span><br /><span style="font-size: medium;">-- AI Studio/Custom Ai copilots: </span><b><i><a href="https://t.co/CERb8RiV0L"><span style="color: #cc0000;">The Verge</span></a></i></b><span style="font-size: medium;">, </span><b><i><a href="https://www.zdnet.com/article/azure-ai-studio-takes-the-stage-at-ignite-2023-unlock-the-potential-of-azure-ai-studios-ai-toolkit/"><span style="color: #cc0000;">ZDNet</span></a></i></b><span style="font-size: medium;">, </span><br /><span style="font-size: medium;">-- Copilot for Windows 10: </span><b><i><a href="https://www.theverge.com/2023/11/16/23963576/microsoft-copilot-windows-10-ai-features-release-preview"><span style="color: #cc0000;">The Verge</span></a></i></b><span style="font-size: medium;">, </span><b><i><a href="https://www.cnbc.com/2023/11/16/microsoft-is-bringing-its-copilot-chatbot-to-windows-10.html"><span style="color: #cc0000;">CNBC</span></a></i></b><span style="font-size: medium;">, </span><b><i><a href="https://arstechnica.com/gadgets/2023/11/microsoft-is-revisiting-windows-10-by-backporting-the-copilot-ai-assistant/"><span style="color: #cc0000;">Ars Technica</span></a></i></b><span style="font-size: medium;">, </span><b><i><a href="https://www.engadget.com/microsofts-copilot-ai-is-officially-coming-to-windows-10-180020032.html"><span style="color: #cc0000;">Engadget</span></a></i></b><span style="font-size: medium;">, </span><br /><span style="font-size: medium;">-- Teams Copilot: </span><b><i><a href="https://www.cnbc.com/2023/11/15/microsoft-announces-new-copilot-features-in-outlook-powerpoint-teams.html"><span style="color: #cc0000;">CNBC</span></a></i></b><span style="font-size: medium;">, </span><b><i><a href="https://www.theverge.com/2023/7/18/23798816/microsoft-teams-365-copilot-ai-chat"><span style="color: #cc0000;">The Verge</span></a></i></b><span style="font-size: medium;">, <br />-- Support for open source LLMs: <b><i><a href="https://t.co/LTwMQwD4sH"><span style="color: #cc0000;">VentureBeat</span></a></i></b><br /><br /></span></span></span></li><li><span style="font-family: verdana;"><span style="font-family: verdana;"><span style="font-size: medium;"><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#OtherNews" style="font-size: small; font-weight: 700; text-align: center;"><span style="color: #cc0000;">OtherNews</span></a><br />"Courtesy of <i>[Google DeepMind]</i> AI: Weather forecasts for the hour, the week and the century", Devin Coldewey, <b><i><a href="https://t.co/7szMYJTtMu"><span style="color: #cc0000;">TechCrunch</span></a></i></b>, 11/14/23 <b><span style="color: #cc0000;">***</span></b> <br /><span style="font-family: verdana;">-- </span><span style="font-family: verdana;">This story also covered by <b><i><a href="https://t.co/ZVUWEP6aUh"><span style="color: #cc0000;">Gizmodo</span></a></i></b>, <b><i><a href="https://www.ft.com/content/ca5d655f-d684-4dec-8daa-1c58b0674be1"><span style="color: #cc0000;">Financial Times</span></a></i></b>, <b><i><a href="https://www.technologyreview.com/2023/11/14/1083366/google-deepminds-weather-ai-can-forecast-extreme-weather-quicker-and-more-accurately/"><span style="color: #cc0000;">MIT Tech Review</span></a></i></b>, <b><i><a href="https://arstechnica.com/science/2023/11/ai-outperforms-conventional-weather-forecasting-for-the-first-time-google-study/"><span style="color: #cc0000;">Ars Technica</span></a></i></b>, <b><i><a href="https://www.wired.com/story/google-deepmind-ai-weather-forecast/"><span style="color: #cc0000;">Wired</span></a></i></b>, <b><i><a href="https://t.co/2P7gFHA3Ut"><span style="color: #cc0000;">ZDNet</span></a></i></b>, <br />-- These articles describe research published in </span></span></span></span><span><span style="font-size: medium;"><a href="https://www.nature.com/articles/d41586-023-03552-y" style="font-style: italic; font-weight: bold;"><span style="color: #cc0000;">Nature</span></a><b style="font-style: italic;"> </b>and<b style="font-style: italic;"> <a href="https://www.science.org/doi/10.1126/science.adi2336"><span style="color: #cc0000;">Science</span></a></b></span><br /><br /></span></li><li><span style="font-family: verdana;"><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#OtherNews" style="background-color: white; color: #ff8832; font-weight: 700; text-align: center;"><span style="color: #cc0000;"><span style="font-size: x-small;">OtherNews</span></span></a><br class="Apple-interchange-newline" /><span style="font-size: medium;">"OpenAI Ousts CEO Sam Altman", Will Knight, <b><i><a href="https://t.co/s4RBLzVqjj"><span style="color: #cc0000;">Wired</span></a></i></b>, 11/17/23 <b><span style="color: #cc0000;">***</span></b> <br /><span style="font-family: verdana;">-- </span><span style="font-family: verdana;">This story also covered by <b><i><a href="https://www.nytimes.com/2023/11/17/technology/openai-sam-altman-ousted.html?smid=nytcore-android-share"><span style="color: #cc0000;">NY Times</span></a></i></b>, <b><i><a href="https://t.co/P3BpF1LuES"><span style="color: #cc0000;">Mashable</span></a></i></b>, <b><i><a href="https://t.co/k9AfmVRKIa"><span style="color: #cc0000;">Computerworld</span></a></i></b>, <b><i><a href="https://twitter.com/ReutersTech/status/1725652738199916934/photo/1"><span style="color: #cc0000;">Reuters</span></a></i></b>, <b><i><a href="https://t.co/0NuXD3j7OD"><span style="color: #cc0000;">VentureBeat</span></a></i></b>, <b><i><a href="https://t.co/V5MQCgjlWT"><span style="color: #cc0000;">Blomberg</span></a></i></b>, <b><i><a href="https://t.co/iL11wgtKZI"><span style="color: #cc0000;">Engadget</span></a></i></b>, <b><i><a href="https://www.wsj.com/tech/sam-altman-departs-open-ai-mira-murati-interim-ceo-41f6d51e?mod=e2twd"><span style="color: #cc0000;">Wall Street Journal</span></a></i></b>, <b><i><a href="https://go.forbes.com/c/h8C5"><span style="color: #cc0000;">Forbes</span></a></i></b>, <b><i><a href="https://t.co/tosMzoAxrG"><span style="color: #cc0000;">BBC</span></a></i></b>, </span></span></span><span style="font-size: medium;"><br /><br />-- On Monday morning (11/20/23), the NY Times announced that Microsoft had hired Altman </span><b style="font-size: large;">➡ </b><span style="font-size: large;">"Microsoft Hires Sam Altman Hours After OpenAI Rejects His Return", Mike Isaac, Kevin Roose and Cade Metz, <b><i><a href="https://www.nytimes.com/2023/11/20/technology/openai-altman-ceo-not-returning.html?smid=nytcore-android-share"><span style="color: #cc0000;">NY Times</span></a></i></b>, 11/20/23<br /><br />--<b> </b><span>"Hundreds of OpenAI employees threaten to resign and join Microsoft"</span><b>, </b><span>Jacob Kastrenakes and Tom Warren, <b><i><a href="https://www.theverge.com/2023/11/20/23968988/openai-employees-resignation-letter-microsoft-sam-altman"><span style="color: #cc0000;">The Verge</span></a></i></b>, 11/20/23</span> </span></li></ol><div><span style="font-size: medium;"><b><br />Relayed stories worth reading/re-reading </b><b>➡</b> What drove George Hinton and Demis Hassabis to become preeminent AI experts?<b> </b></span></div><div><ol style="text-align: left;"><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-size: medium;">"Why the Godfather of A.I. Fears What He’s Built", Joshua Rothman, <b><i><a href="https://www.newyorker.com/magazine/2023/11/20/geoffrey-hinton-profile-ai" style="color: #336699; text-decoration-line: none;"><span style="color: #cc0000;">The New Yorker</span></a></i></b>, 11/13/23 ... a rambling, but insightful piece based on a series of interviews with the man whose neural network research provided the foundation for large language models (LLMs).<br /><br /></span></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-size: medium;">"A.I. Could Solve Some of Humanity's Hardest Problems. It Already Has.". Guest = Demis Hassabis, <b style="background-color: white;"><i><a href="https://www.nytimes.com/2023/07/11/opinion/ezra-klein-podcast-demis-hassabis.html" style="text-decoration-line: none;"><span style="color: #cc0000;">The Ezra Klein Show</span></a></i></b><span style="background-color: white; color: #1a222a;"> (podcast with transcript), 7/11/23 ... a focused, insightful interview with the man whose neural network research provided the foundation for neural networks that solve complex problems in science.</span></span><br /><br /></li></ol></div></span></span></div><div id="Overview" style="text-align: center;"><span><div id="ChatGPT"><b><b style="background-color: white;"><br /></b></b></div></span></div></div><div style="text-align: left;"><span style="font-family: verdana;">____________________________________</span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b>Links to related notes on this blog: </b></span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana; font-size: medium;"><b><br /></b></span></li></ul></div>Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-37223937703861427872023-11-11T19:56:00.144-05:002024-03-27T19:32:21.173-04:00OpenAI's Dev Con, Elon's Grok, and Humane's AI pin ... TL;DR + Podcast 12Nov23 <div style="text-align: right;"><span style="font-family: verdana;"><i><span style="font-size: x-small;">Last update: </span><b><span style="color: #cc0000;">11/13/23 </span></b></i></span></div><div style="text-align: left;"><span style="font-family: verdana;"><div class="separator" style="clear: both; text-align: center;"><span style="margin-left: 1em; margin-right: 1em;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjRyH2pBWj4rWsb5gHn-aNALuEshSRArYu4XnWFF-xS9C2mPrDlyPdhhFXkogSvAR3i9WHWRMlNvAxEC4rcryp7iaEZRbIReuVkXZVxtUrqgzK0T-YtIra80WFcOvudbgrRZcrim7Ffat1vAczO9oGVtVc0FUlofjwieVabCKRIskAO5SDnfZy1VWIh3EM/s1280/OpenAIDevDay6Nov23.webp" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="960" data-original-width="1280" height="240" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjRyH2pBWj4rWsb5gHn-aNALuEshSRArYu4XnWFF-xS9C2mPrDlyPdhhFXkogSvAR3i9WHWRMlNvAxEC4rcryp7iaEZRbIReuVkXZVxtUrqgzK0T-YtIra80WFcOvudbgrRZcrim7Ffat1vAczO9oGVtVc0FUlofjwieVabCKRIskAO5SDnfZy1VWIh3EM/s320/OpenAIDevDay6Nov23.webp" vspace="10" width="320" /></a></span><div style="text-align: left;"><span><span style="font-size: large;"><span style="background-color: white; color: #1a222a;">Welcome to our TL;DR summary + podcast about the past week's top AI stories on our "<b><i><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html">Useful AI News</a></i></b>" page: 1) Open AI's first-ever developers conference, (2) "Grok" (Elon Musk's sardonic chatbot), and (3) Humane's new AI pin.<br />... </span></span><span style="background-color: white; color: #1a222a;"><span style="font-size: medium;">Click the "start" button in the audio control (below) to hear the podcast</span><span style="font-size: large;">.</span></span></span></div><div style="background-color: white; color: #1a222a; font-family: Verdana, Geneva, sans-serif; font-size: 13.86px; text-align: left;"><span style="font-family: verdana; font-size: medium;"></span></div><div style="background-color: white; color: #1a222a; font-family: Verdana, Geneva, sans-serif; text-align: right;"><span style="font-family: verdana; font-size: medium;"> <iframe allow="autoplay" height="50" src="https://drive.google.com/file/d/1KCsyKLs0D4-oLS6p56OjOubxuIOynJD7/preview?usp=sharing" width="400"></iframe></span><div class="separator" style="clear: both; text-align: left;"><span face="Verdana, Geneva, sans-serif" style="font-family: verdana; font-size: medium;">TL;DR link<b> </b></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-family: verdana; font-size: large; text-decoration-line: none;"><span style="color: #1a222a;">➡</span></a><span style="color: #1a222a; font-family: verdana;"> </span><b style="color: black; font-family: verdana;"><i><a href="https://neoskeptics.blogspot.com/2023/11/openai-dev-con-elons-grok-ai-badge-top.html#more"><span style="color: #cc0000;">HERE</span></a></i></b></div><div class="separator" style="clear: both; text-align: left;"><b><span style="font-size: medium;"><br /></span></b></div></div></div></span></div><span><span style="font-family: verdana;"><a name='more'></a></span></span><div style="text-align: left;"><span style="font-family: verdana;">This blog's editor had three reactions to the top three stories this week ➡ shrug ... shrug ... WOW!!!. </span></div><div style="text-align: left;"><div class="separator" style="clear: both; text-align: left;"><div class="separator" style="clear: both;"><span style="font-family: verdana;"><br /></span></div><div class="separator" style="clear: both;"><span style="font-family: verdana;">He shrugged off the first two stories because they were predictable and did not change the course of recent developments; but the AI pin is a head turner that might radically change the way that many of us interact physically with the all-encompassing Internet.</span></div><div class="separator" style="clear: both;"><span style="font-family: verdana;"> </span></div><div class="separator" style="clear: both;"><span style="font-family: verdana;">1. First, let's consider the biggest announcements at OpenAI's Dev Con:</span></div><div class="separator" style="clear: both;"><span style="font-family: verdana;"><br /></span></div><div class="separator" style="clear: both;"><span style="font-family: verdana;">a) A new LLM called "GPT-4 Turbo" that is more powerful than GPT-4 ... Of course. OpenAI has to protect its claim to having the most powerful LLM on the market.</span></div><div class="separator" style="clear: both;"><span style="font-family: verdana;"><br /></span></div><div class="separator" style="clear: both;"><span style="font-family: verdana;">b) A GPT store for apps, just like Apple's App Store. Like Apple, OpenAI intends to make lots of money by charging substantial fees for helping professional developers sell their apps through its store if they meet OpenAI's standards. <i>(OpenAI has standards???)</i></span></div><div class="separator" style="clear: both;"><span style="font-family: verdana;"><br /></span></div><div class="separator" style="clear: both;"><span style="font-family: verdana;">c) A new service called "GPTs" that allows individuals and small businesses to become amateur developers of customized versions of ChatGPT that they can instantly share with friends and associates. One has to assume that amateur versions of ChatGPT will also hallucinate, display bias against women and minority groups, violate other people's copyrights, be easily hacked to produce misinformation, etc, etc, etc. <i>(If they don't, they won't meet OpenAI's standards.)</i></span></div><div class="separator" style="clear: both;"><span style="font-family: verdana;"><br /></span></div><div class="separator" style="clear: both;"><span style="font-family: verdana;">2. Moving on to Elon Musk's "Grok", it's a new chatbot that Elon proudly announced has an uncensured potty mouth. Of course.</span></div><div class="separator" style="clear: both;"><span style="font-family: verdana;"><br /></span></div><div class="separator" style="clear: both;"><span style="font-family: verdana;">3. Now for the "WOW!!!" </span></div><div class="separator" style="clear: both;"><span style="font-family: verdana;">Humane has developed an "AI pin" (that looks like and is worn like a badge) with funding from Microsoft, OpenAI, Salesforce, and other wealthy Silicon Valley patrons. It might replace smart phones as our primary user interface with the Internet. Instead, the AI pin uses the palms of our hands with a little help from our fingertips. </span></div><div class="separator" style="clear: both;"><span style="font-family: verdana;"><br /></span></div><div class="separator" style="clear: both;"><span style="font-family: verdana;">Please, please, please watch the demo videos embedded in the New York Times article so that you can see what we're talking about ➡ <b><i><a href="https://www.nytimes.com/2023/11/09/technology/silicon-valleys-big-bold-sci-fi-bet-on-the-device-that-comes-after-the-smartphone.html?smid=nytcore-android-share"><span style="color: #cc0000;">https://www.nytimes.com/2023/11/09/technology/silicon-valleys-big-bold-sci-fi-bet-on-the-device-that-comes-after-the-smartphone.html?smid=nytcore-android-share</span></a></i></b></span></div></div><div class="separator" style="clear: both; text-align: left;"><span style="font-family: verdana;"><br /></span></div><div id="Overview"><b><span style="font-size: medium;"><span style="font-family: verdana;">Top stories in past week ...</span></span></b></div><div id="Overview"><ol><li><span style="font-family: verdana;"><b style="font-size: small; text-align: center;"><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#OpenAI"><span style="color: #cc0000;">OpenAI/Microsoft</span></a></b><br class="Apple-interchange-newline" /><span style="font-size: medium;">OpenAI's First-ever Developer Day Conference (Nov 6, San Francisco, CA) </span></span><span style="font-size: medium;"><b style="font-family: verdana;"><span style="color: #cc0000;">**</span></b><span style="color: #cc0000; font-family: verdana; font-size: medium;">*</span><br /></span><span style="font-family: verdana;"><span style="font-size: medium;">-- <i>Overview</i>: "Biggest Moments From OpenAI's Dev Day", Maxwell Zeff, <b><i><a href="https://t.co/egfz6gO7m9"><span style="color: #cc0000;">Gizmodo</span></a></i></b>, 11/6/23 ... Overview also provided by <b><i><a href="https://t.co/uHiufSmFHj"><span style="color: #cc0000;">ZDNet</span></a></i></b><br />-- <i>GPT-4 Turbo</i>: "GPT-4 Turbo is OpenAI’s most powerful large language model yet", Pranav Dixit, <b><i><a href="https://t.co/zKOfrpqhdH"><span style="color: #cc0000;">Engadget</span></a></i></b>, 11/6/23 ... GPT-4 Turbo also covered by <b><i><a href="https://t.co/U3St2lPyXS"><span style="color: #cc0000;">VentureBeat</span></a>, <a href="https://t.co/Q2KTqwCDR6"><span style="color: #cc0000;">TechCrunch</span></a></i></b><br />-- <i>App store:</i> "App Store for AI: OpenAI’s GPT Store lets you build (and monetize) your own GPT", Devin Coldewey, <b><i><a href="https://t.co/jNtc9LzG3n"><span style="color: #cc0000;">TechCrunch</span></a></i></b>, 11/6/23<br /></span><span style="font-size: medium;">-- Build your own Chatbot: "OpenAI Lets Mom-and-Pop Shops Customize ChatGPT", Cade Metz, <b><i><a href="https://www.nytimes.com/2023/11/06/technology/openai-custom-chatgpt.html?smid=nytcore-android-share"><span style="color: #cc0000;">NY Times</span></a></i></b>, 11/9/23 ... BYO chatbots also covered by <b><i><a href="https://www.zdnet.com/article/heres-how-to-create-your-own-custom-chatbots-using-chatgpt/"><span style="color: #cc0000;">ZDNet</span></a></i></b>, <b><i><a href="https://tunein.com/podcasts/Arts--Culture-Podcasts/Sway-p1365583/?topicId=347328526"><span style="color: #cc0000;">HardFork podcast</span></a></i></b>,</span><br /><br /></span></li><li><span style="font-family: verdana; font-size: medium;"><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#OtherNews" style="font-size: small; font-weight: 700; text-align: center;"><span style="color: #cc0000;">OtherNews</span></a><br />"Elon Musk Announces Grok, a ‘Rebellious’ AI With Few Guardrails", Will Knight, <b><i><a href="https://t.co/1XrDh80OLv"><span style="color: #cc0000;">Wired</span></a></i></b>, 11/6/23 <b><span style="color: #cc0000;">***</span></b> <br />-- This story also covered by <b><i><a href="https://t.co/lmF9VRKwKZ"><span style="color: #cc0000;">ZDNet</span></a></i></b>, <a href="https://t.co/TC9Xo0HmWd"><span style="color: #cc0000;"><b><i>Mashable</i></b></span></a>, <b><i><a href="https://www.cnbc.com/2023/11/05/elon-musk-debuts-grok-ai-bot-to-rival-chatgpt-others-.html"><span style="color: #cc0000;">CNBC</span></a></i></b>, <b><i><a href="https://www.fastcompany.com/90978698/elon-musk-xai-grok-bias-misinformation"><span style="color: #cc0000;">FastCompany</span></a><br /><br /></i></b></span></li><li><span style="font-family: verdana; font-size: medium;"><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#OtherNews" style="font-size: small; font-weight: bold; text-align: center;"><span style="color: #cc0000;">OtherNews</span></a><br /><span style="font-family: verdana;">"Silicon Valley’s Big, Bold Sci-Fi Bet on the Device That Comes After the Smartphone", Erin Griffith, Tripp Mickle and Photographs and </span><span style="font-family: verdana; font-weight: bold;">Videos</span><span style="font-family: verdana;"> by Kelsey McClellan, <b><i><a href="https://www.nytimes.com/2023/11/09/technology/silicon-valleys-big-bold-sci-fi-bet-on-the-device-that-comes-after-the-smartphone.html?smid=nytcore-android-share"><span style="color: #cc0000;">NY Times</span></a></i></b>, 11/10/23 <b><span style="color: #cc0000;">***</span></b> ... Funded by Microsoft, OpenAI, SalesForce and others<br />-- </span><span style="font-family: verdana;">This story also covered by <b><i><a href="https://t.co/DGlwnuaOhl"><span style="color: #cc0000;">ZDNet</span></a>,</i></b> </span><a href="https://t.co/5MZaNEHyW7"><b><i><span style="color: #cc0000;">TechCrunch</span></i></b></a><span style="font-family: Times;">, </span><span style="font-family: verdana;"><span><a href="https://www.wired.com/story/humane-ai-pin-700-dollar-smartphone-alternative-wearable/" style="color: #cc0000; font-style: italic; font-weight: bold;"><span style="color: #cc0000;">Wired</span></a><span style="color: #cc0000; font-style: italic; font-weight: bold;">, </span><a href="https://t.co/To46AQ5cNd" style="color: #cc0000; font-style: italic; font-weight: bold;"><span style="color: #cc0000;">Forbes Tech</span></a><span style="color: #cc0000; font-style: italic; font-weight: bold;">, </span><span><a href="https://t.co/q5aDcnaGhx" style="font-style: italic; font-weight: bold;"><span style="color: #cc0000;">The Verge</span></a><br />-- The videos in the Times' article provide the best descriptions of this extraordinary device</span></span></span></span></li></ol></div><div id="Overview" style="text-align: center;"><b><span></span></b></div></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">____________________________________</span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b>Links to related notes on this blog: </b></span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana; font-size: medium;"><b><br /></b></span></li></ul></div>Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-74850272517162258342023-11-09T19:36:00.017-05:002024-03-27T19:34:23.171-04:00A podcast about our podcasts ... Coming soon!!!<div style="text-align: right;"><span style="font-family: verdana; font-size: x-small;"><i>Last update: 11/9/23 </i></span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhpKzRPe4vWL-toUG3bV8X2qx9tx1xKtK4uG-Oozm-QKtoQBKGcCk2nWT3UQlkk6tFhi0qR-zR4UXzjvtBwhwxNeUFRUlrAhtj5801jMCdvaUIBE1lTNKiWmjNtz9vKsGH5RwJj7lypnWIcHo_YVVDqpxMHn0-ZjjaNAUEvulp8ep_d_NYYUU1FjYnY6wM/s4480/HeadphonesForPodcasts.jpeg" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="2560" data-original-width="4480" height="183" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhpKzRPe4vWL-toUG3bV8X2qx9tx1xKtK4uG-Oozm-QKtoQBKGcCk2nWT3UQlkk6tFhi0qR-zR4UXzjvtBwhwxNeUFRUlrAhtj5801jMCdvaUIBE1lTNKiWmjNtz9vKsGH5RwJj7lypnWIcHo_YVVDqpxMHn0-ZjjaNAUEvulp8ep_d_NYYUU1FjYnY6wM/s320/HeadphonesForPodcasts.jpeg" vspace="10" width="320" /></a></div>Welcome to a short podcast about our forthcoming weekly podcasts ...<br /><br /></span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"></span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"> <iframe allow="autoplay" height="80" src="https://drive.google.com/file/d/1JbqBIUUhnqFqADBZx7Yqphs4OxDSP4n4/preview?usp=sharing" width="240"> </iframe></span><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg6C0kYM8xFBxdO4IC-vNpmXWkgUaA0UV6TBs-D8uRHqYtZE7EU_HJTzSA6V8t1rCbqQSdarZIAectFDk3BKVfjRXDzDGu8NePAYjFXB6R3-uXE8wLaNP_iLtCFyUEQ7Udt9nNXfM5WWXbju_SfpFJuZdn7B0zj8arx-u9aaH5SEkKSoeMvkhNBhspyDNE/s73/dad10_small_bigger.jpg" style="margin-left: 1em; margin-right: 1em;"><img align="right" border="0" data-original-height="73" data-original-width="73" height="73" hspace="1" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg6C0kYM8xFBxdO4IC-vNpmXWkgUaA0UV6TBs-D8uRHqYtZE7EU_HJTzSA6V8t1rCbqQSdarZIAectFDk3BKVfjRXDzDGu8NePAYjFXB6R3-uXE8wLaNP_iLtCFyUEQ7Udt9nNXfM5WWXbju_SfpFJuZdn7B0zj8arx-u9aaH5SEkKSoeMvkhNBhspyDNE/s1600/dad10_small_bigger.jpg" vspace="1" width="73" /></a></div><br /></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><br /></span></div><div style="text-align: left;"><div class="separator" style="clear: both; color: #1a222a; font-family: Verdana, Geneva, sans-serif;"><span face="Verdana, Geneva, sans-serif" style="font-family: verdana; font-size: medium;">Read more<b> </b></span><a href="https://neoskeptics.blogspot.com/2023/12/first-came-ai-big-bang-now-comes.html" style="color: #336699; font-family: verdana; font-size: large; text-decoration-line: none;"><span style="color: #1a222a;">➡</span></a><span style="color: #1a222a; font-family: verdana;"> </span><b style="color: black; font-family: verdana;"><i><a href="https://neoskeptics.blogspot.com/2023/11/a-podcast-about-our-podcasts-coming-soon.html#more"><span style="color: #cc0000;">HERE</span></a></i></b></div><div><br /></div></div><span><span style="font-family: verdana;"><a name='more'></a></span></span><div style="text-align: left;"><span style="font-family: verdana;"><span style="font-size: medium;"><b>TRANSCRIPT ...</b></span></span></div><div style="text-align: left;"><div><span style="font-family: verdana;"> </span></div><div><span style="font-family: verdana;">This is a podcast about our weekly podcasts that will be coming soon!!!</span></div><div><span style="font-family: verdana;"><br /></span></div><div><span style="font-family: verdana;">It's from me, Roy Beasley, editor of the Neoskeptics Blog.</span></div><div><span style="font-family: verdana;"><br /></span></div><div><span style="font-family: verdana;">Every day, with the help of a few friends and associates, I try to identify the most useful AI related news stories in about 20 tech media. </span></div><div><span style="font-family: verdana;"><br /></span></div><div><span style="font-family: verdana;">I post links to these stories on our blog's "Useful Ai News" page in various categories.</span></div><div><span style="font-family: verdana;"><br /></span></div><div><span style="font-family: verdana;">I'm retired, so I have the time to read all of these articles, but I assume that most of our readers don't.</span></div><div><span style="font-family: verdana;"> </span></div><div><span style="font-family: verdana;">So on Fridays, I identify the top 2, 3, 4 or 5 articles that were added to the AI news page during that week.</span></div><div><span style="font-family: verdana;"><br /></span></div><div><span style="font-family: verdana;">I copy the links to those articles into the "Top stories" headline section at the top of the news page.</span></div><div><span style="font-family: verdana;"><br /></span></div><div><span style="font-family: verdana;">Unfortunately, most of our readers don't have the time to read all of these top stories.</span></div><div><span style="font-family: verdana;"><br /></span></div><div><span style="font-family: verdana;">That's why I recently added a short note above the headline section, a TL;DR -- "Too long, did not read" -- i.e.. an executive summary of that week's top stories that only takes a few minutes to read. To date, I have published TL;DR summary notes on the blog each week for the last three weeks,</span></div><div><span style="font-family: verdana;"><br /></span></div><div><span style="font-family: verdana;">Unfortunately, our busiest readers don't even have the time to sit down at their computer screens or to look at their mobile phones long enough to read the TL;DR summaries ... but they might listen to a podcast,</span></div><div><span style="font-family: verdana;"><br /></span></div><div><span style="font-family: verdana;">Our busiest readers could listen to a podcast while doing something else ... while getting dressed in the morning, while eating breakfast, while driving to work. They would not need to look at a screen while listening to a podcast. </span></div><div><span style="font-family: verdana;"><br /></span></div><div><span style="font-family: verdana;">Our podcast audio will convey the same summary information as the printed TL;DR section, but will be phrased in a conversational format. If readers find podcast audio comments about a top story of sufficient interest, they will be able to click the links in the TL;DR section that will load the associated news story into their browsers.</span></div><div><span style="font-family: verdana;"><br /></span></div><div><span style="font-family: verdana;">Going forward, the weekly TL;DR page will also contain audio controls that will play the weekly podcast file. The TL;DR plus podcast will be published on this blog on Friday evenings. Announcements of thepublication will be posted on the @Neoskeptics Twitter account and on my LinkedIn personal account on Friday evenings and on Saturday mornings.</span></div><div><span style="font-family: verdana;"><br /></span></div><div><span style="font-family: verdana;">Thank you.</span></div><div><span style="font-family: verdana;"><br /></span></div><div><span style="font-family: verdana;">11/9/23 @ 6:54 pm</span></div><div><br /></div><div><br /></div><div><br /></div><div><br /></div><div><br /></div></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">____________________________________</span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b>Links to related notes on this blog: </b></span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana; font-size: medium;"><b><br /></b></span></li></ul></div>Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-66943041327371768122023-11-03T20:06:00.015-04:002023-11-03T21:14:53.735-04:00Civil war in the AI community ➡ Open vs. closed AI platforms ... Top AI stories TL;DR 3Nov23<div style="text-align: right;"><span style="font-family: verdana; font-size: x-small;"><i>Last update: 11/3/23 </i></span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiVRBLpx_MG6DpvofJa2oodLDdnJa2Kaj-QIdIOjHpIBuLjcRbxDKZYxVBbJbbzbepYaAibU_DqbJWbk1XSwQUmjI9E6NRpyIkqO_M33-wp8FySFiPJ1q0dig3KrvBSvtGvng9bD2tmkBY2LpfqoQaTI8fZGdgAF0fC3GHdl1vePRefDWoFVIWN0QV60Mw/s750/MegaTweetThurs2Nov23WorldCupAiPolicy.jpg" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="429" data-original-width="750" height="183" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiVRBLpx_MG6DpvofJa2oodLDdnJa2Kaj-QIdIOjHpIBuLjcRbxDKZYxVBbJbbzbepYaAibU_DqbJWbk1XSwQUmjI9E6NRpyIkqO_M33-wp8FySFiPJ1q0dig3KrvBSvtGvng9bD2tmkBY2LpfqoQaTI8fZGdgAF0fC3GHdl1vePRefDWoFVIWN0QV60Mw/s320/MegaTweetThurs2Nov23WorldCupAiPolicy.jpg" vspace="10" width="320" /></a></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;">This week's top stories covered two related developments. The first played out onstage in three of the grandest global venues; whereas the second unfolded backstage ... <i>Note: The current edition of "Useful AI News" can be found </i></span><i><span style="background-color: white; color: #1a222a; font-family: verdana;">➡</span><span style="font-family: verdana; font-size: large;"> </span><b><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html">HERE</a></b></i></div></span></div><span><span style="font-family: verdana;"><a name='more'></a></span></span><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><br /></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">The onstage global developments featured ultra polite, high sounding speeches and broad non-binding agreements among policy leaders who strove mightily to convince their audiences in the media that they really understood the meaning of their speeches and agreements about why and how Ai should be regulated. </span></div><div style="text-align: left;"><div id="Overview"><span><span style="font-family: verdana;"><div id="Overview"><span><br />The backstage developments featured world class experts in AI who were not polite to one another. Indeed, their no-holds-barred exchanges resounded like cannon volleys in a civil war. The pessimistic advocates of closed AI platforms urged immediate government regulation of AI, such as was being proposed by the onstage policy actors, because of their certainty that the dangers of the newest AI technologies are potentially catastrophic.</span></div><div id="Overview"><span><br /></span></div><div id="Overview"><span>By contrast, the optimistic advocates of open platforms opposed regulation at this time because they think that the newest AI tech is so new that we cannot make credible predictions about its future development. Some open advocates accused closed advocates from Big Tech companies of "fearmongering" designed to consolidate Big Tech's current positions as the world's AI leaderss.</span></div><div id="Overview"><span><br /></span></div><div id="Overview"><span>The editor of this blog recommends that his readers merely skim the headlines for the first story, but somehow find the time to read the second story's articles that cover the arguments among the experts. Given that the doomsayers have received and probably will continue to receive far more coverage in the media, the editor also recommends that his readers attempt to correct this imbalance by reading the following long piece by another ardent advocate for open AI platforms: "Attenuating Innovation (AI)", Ben Thompson, <b><i><a href="https://stratechery.com/2023/attenuating-innovation-ai/"><span style="color: #cc0000;">STRATCHERY</span></a></i></b>, 11/1/23</span></div><div id="Overview"><span style="font-size: medium;"><br /></span></div></span></span></div><div id="Overview"><b><span style="font-size: medium;"><span style="font-family: verdana;">Top 2 stories this week ...</span></span></b></div><div id="Overview"><ol><li><span style="font-family: verdana;"><span style="color: #cc0000; font-size: x-small;"><b>Policy</b></span><br />"<span style="font-size: medium;">The ‘World Cup’ of AI policy: will USA win?", Sharon Goldman, <b><i><a href="https://t.co/Ga8ZtFBjuX"><span style="color: #cc0000;">VentureBeat</span></a></i></b>, 10/31/23 ... This article applies a soccer championship metaphor to three overlapping AI "Big Policy" events <b><span style="color: #cc0000;">***</span></b> <br /><br />a. </span></span><span style="font-size: medium;"><span style="font-family: verdana;">"Joe Biden’s Big AI Plan Sounds Scary—but Lacks Bite", Matt Laslo, <b><i><a href="https://www.wired.com/story/joe-bidens-big-ai-plan-sounds-scary-lacks-bite/?mbid=social_twitter&utm_brand=wired&utm_source=twitter&utm_medium=social&utm_social-type=owned"><span style="color: #cc0000;">Wired</span></a></i></b>, 10/31/23 ... This event also coveed by <b><i><a href="https://t.co/u0aiqdRM21"><span style="color: #cc0000;">NY Times,</span></a></i></b> <b><i><a href="https://t.co/xg0MsiIMHh"><span style="color: #cc0000;">VentureBeat</span></a></i></b>, <b><i><a href="https://www.forbes.com/sites/davidjeans/2023/10/30/bidens-new-executive-order-will-regulate-ai-models-that-could-threaten-national-security/?utm_source=newsletter&utm_medium=email&utm_campaign=dailydozen&cdlcid=623cf7856e1a1d1211df6bb6&section=tech&sh=1a1db49c1c5b"><span style="color: #cc0000;">Forbes</span></a></i></b>, <br /><br />b. "World leaders are gathering at the U.K.’s AI Summit. Doom is on the agenda.", Cat Zakrzewski, Anthony Faiola and Gerrit De Vynck, <b><i><a href="https://www.washingtonpost.com/technology/2023/10/31/uk-ai-safety-summit-rishi-sunak-elon-musk/?utm_campaign=wp_todays_headlines&utm_medium=email&utm_source=newsletter&wpisrc=nl_headlines"><span style="color: #cc0000;">Washington Post,</span></a></i></b> 10/31/23 </span>... <span style="font-family: verdana;">The "Bletchley Declaration" that emerged from this meeting was discussed by</span><span style="font-family: verdana;"> <b><i><a href="https://t.co/ePTtN8m4kS"><span style="color: #cc0000;">ZDNet</span></a></i></b>, <b><i><a href="https://t.co/b6tB8n9ZiR"><span style="color: #cc0000;">Wired</span></a></i></b></span><br /><br /><span style="font-family: verdana;">c. "G7 to agree AI code of conduct for companies", Foo Yun Chee, <b><i><a href="https://www.reuters.com/technology/g7-agree-ai-code-conduct-companies-g7-document-2023-10-29/"><span style="color: #cc0000;">Reuters</span></a></i></b>, 10/29/23 </span> ... <span style="font-family: verdana;"> This event also covered by </span><b style="font-family: verdana;"><i><a href="https://www.forbes.com/sites/cindygordon/2023/10/31/g7-leaders-release-ai-governance-code-same-day-as-usa-president-signs-an-ai-executive-order/?sh=797bc0e54b44"><span style="color: #cc0000;">Forbes</span></a></i></b><span style="font-family: verdana;">, </span><a href="https://venturebeat.com/ai/to-promote-safe-secure-trustworthy-ai-g7-introduces-voluntary-code-of-conduct/" style="font-family: verdana; font-style: italic; font-weight: bold;"><span style="color: #cc0000;">VentureBeat</span></a></span><span style="font-family: verdana;"><span style="font-size: medium;"><br /></span><br /></span></li><li><span style="font-family: verdana;"><span style="color: #cc0000; font-size: x-small;"><b>OpEds</b></span><br /><span style="font-size: medium;"><span>Civil war breaks out among AI experts ... open vs. closed AI platforms</span><br /><br /><span>a. </span><a href="https://twitter.com/ylecun/status/1718670073391378694" style="font-style: italic; font-weight: bold;"><span style="color: #cc0000;">Yann LeCun's manifesto</span></a><b style="font-style: italic;">, </b><span>megapost on X/Twitter, 10/29/23</span><br /><br /><span>b. "Google DeepMind boss hits back at Meta AI chief </span><i>[Yann LeCun] </i><span>over ‘fearmongering’ claim", Ryan Browne, </span><b><i><a href="https://t.co/J3I2ccj8oK"><span style="color: #cc0000;">CNBC</span></a></i></b><span>, 10/31/23</span></span></span></li></ol></div></div><div style="text-align: left;"><span style="font-family: verdana;">____________________________________</span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b>Links to related notes on this blog: </b></span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana; font-size: medium;"><b><br /></b></span></li></ul></div>Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-68507123567990351562023-10-28T19:37:00.008-04:002023-11-03T20:27:19.271-04:00Google vs. Microsoft... Top AI stories TL;DR 27OCT23<div style="text-align: right;"><span style="font-family: verdana; font-size: x-small;"><i>Last update: 10/27/23</i></span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhbROJTHO5ZTXNn1zE57lc59SwXT28cuLz0v7VmwTXZUSliHt-Sd8kJs8psB7S4XLOghyphenhyphenLN8nL9kSZve0n3qMxJvZFscUmACJ7W30S4cV2PRLeqWisYKLWC4t6tE7IJXdM2x_CzNFbHShp8BiLrdPwMZjSFykoIeRzklOwRcLA1X-NhCyLiEt2DiacO_kY/s1440/MegaTweetFri27Oct23MicrosoftCEONadallaSmiling.jpg" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="1071" data-original-width="1440" height="238" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhbROJTHO5ZTXNn1zE57lc59SwXT28cuLz0v7VmwTXZUSliHt-Sd8kJs8psB7S4XLOghyphenhyphenLN8nL9kSZve0n3qMxJvZFscUmACJ7W30S4cV2PRLeqWisYKLWC4t6tE7IJXdM2x_CzNFbHShp8BiLrdPwMZjSFykoIeRzklOwRcLA1X-NhCyLiEt2DiacO_kY/s320/MegaTweetFri27Oct23MicrosoftCEONadallaSmiling.jpg" vspace="10" width="320" /></a></div>Consider the image for this blog note. Question: Why is that man smiling? Answer: He is smiling because he is Satya Nadella, the very well paid CEO of Microsoft. As described in the articles reporting our first top story, Nadella's huge investment bets in AI have yielded substantial quarterly earnings, earnings that have convinced Wall Street that AI will soon become a highly productive and highly profitable money maker for Microsoft. Microsoft’s growth in the most recent quarter stemmed from the business unit the company labels “Azure and other cloud services,” which houses its investments in AI ... <span style="font-family: verdana; font-size: medium;"><i>Note: The current edition of "Useful AI News" can be found </i></span><i style="font-family: Times; font-size: medium;"><span style="background-color: white; color: #1a222a; font-family: verdana;">➡</span><span style="font-family: verdana; font-size: large;"> </span><b style="font-family: verdana; font-size: large;"><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html">HERE</a></b></i><br /></span><span style="font-family: verdana; font-size: large;"> </span></div><span><span style="font-family: verdana;"><a name='more'></a></span></span><div style="text-align: left;"><div id="Overview"><span><span style="font-family: verdana;"><div id="Overview">Indeed, within a few days Microsoft will formally roll out "Copilots" for all of its products and services, "Copilot" being the name that Microsoft is calling generative AI extensions to each of its products and services. Enterprise clients will pay per user monthly subscription fees for access to Copilots. Here's a link to Microsoft's official description of its Copilot initiative ➡ <a href="https://blogs.microsoft.com/blog/2023/09/21/announcing-microsoft-copilot-your-everyday-ai-companion/"><b><i>https://blogs.microsoft.com/blog/2023/09/21/announcing-microsoft-copilot-your-everyday-ai-companion/</i></b></a></div><div id="Overview"> </div><div id="Overview">By contrast, our second top story relates to Google's quarterly report (Alphabet's report) that documents its failure to make substantial gains in the cloud. Google's failure to substantially increase its share of the market for cloud services indicated that its higher profits were not related to larger purchases of generative AI services that are inherently cloud-based operations. So Google's Bard chatbot may have caught up to Microsoft's ChatGPT, but the race between Google vs Microsoft is just beginning. At this point Microsoft is definitely "winning".</div></span></span></div><div id="Overview"><b><span><span style="font-family: verdana;"><br /></span></span></b></div><div id="Overview"><b><span><span style="font-family: verdana;">Top 2 stories this week ...</span></span></b></div><div id="Overview"><ol><li><span style="font-family: verdana;"><span style="color: #cc0000; font-size: x-small;"><b>OtherNews</b></span><br /> "Wall Street is demanding financial results that support all the AI hype. Microsoft’s latest earnings finally delivered", Rachyl Jones, <b><i><a href="https://fortune.com/2023/10/24/microsoft-earnings-q1-2024-ai-cloud-financial-results-wall-street/"><span style="color: #cc0000;">Fortune</span></a></i></b> <span style="font-family: verdana;"><span style="color: #cc0000; font-style: italic; font-weight: bold;">***</span><b style="font-style: italic;"> <br /></b></span><span style="font-family: verdana;">-- This story also covered by <b><i><a href="https://www.microsoft.com/en-us/Investor/earnings/default.aspx"><span style="color: #cc0000;">Microsoft</span></a></i></b>, <b><i><a href="https://t.co/xsAXfmpV7k"><span style="color: #cc0000;">NY Times</span></a></i></b>, <b><i><a href="https://www.cnn.com/2023/10/24/tech/microsoft-google-snap-tech-earnings/index.html"><span style="color: #cc0000;">CNN</span></a></i></b>, <b><i><a href="https://www.forbes.com/sites/dereksaul/2023/07/25/microsoft-earnings-ai-lifts-redmond-to-highest-quarterly-sales-ever/"><span style="color: #cc0000;">Forbes</span></a></i></b>, </span><br /><br /></span></li><li><span style="color: #cc0000; font-family: verdana; font-size: x-small;"><b>OtherNews</b></span><br /><span style="font-family: verdana;"> "Alphabet [Google] shares drop as cloud miss overshadows better-than-expected overall results", Jennifer Elias, <b><i><a href="https://www.cnbc.com/2023/10/24/alphabet-googl-earnings-q3-2023.html"><span style="color: #cc0000;">CNBC</span></a></i></b>, 10/25/23 <b><span style="color: #cc0000;">***</span></b> <br />-- This story also covered by <b><i><a href="https://t.co/NKn7Gf9fhB"><span style="color: #cc0000;">NY Times</span></a></i></b>,</span><span style="font-family: verdana;"> </span></li></ol></div></div><div style="text-align: left;"><span style="font-family: verdana;">____________________________________</span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b>Links to related notes on this blog: </b></span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana; font-size: medium;"><b><br /></b></span></li></ul></div>Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-87814171625329670152023-10-20T15:08:00.010-04:002023-11-03T20:28:01.507-04:00Bard vs. ChatGPT ... Top AI stories TL;DR 20Oct23<p style="text-align: right;"><span style="font-family: verdana;"><i>Last update: 10/20/23</i></span></p><p><span style="font-family: verdana;"></span></p><div class="separator" style="clear: both; text-align: center;"><span style="font-family: verdana;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiRF-UNC3-m3Nb8NibdUW5c6PWoZK2LTFQR4P7SI3vNGZ8-eT6BuQERCOq9Dsh9zfJcpcGP1neyGB292sfDAsVxiEILK02OuA4jqQlWO0IFC68AqhH-SaL0AliKgdMDzTo3vB7OmBriXdnLG2w3ELl6TdbY3SxouTw-CRm37HR_9tpAGrZHr8GlpJ0hED8/s2476/MegaTweetFri20Oct23TwoMotorcycleRacers.jpg" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="1694" data-original-width="2476" height="219" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiRF-UNC3-m3Nb8NibdUW5c6PWoZK2LTFQR4P7SI3vNGZ8-eT6BuQERCOq9Dsh9zfJcpcGP1neyGB292sfDAsVxiEILK02OuA4jqQlWO0IFC68AqhH-SaL0AliKgdMDzTo3vB7OmBriXdnLG2w3ELl6TdbY3SxouTw-CRm37HR_9tpAGrZHr8GlpJ0hED8/s320/MegaTweetFri20Oct23TwoMotorcycleRacers.jpg" vspace="10" width="320" /></a></span></div><span style="font-family: verdana;"><div><span style="font-size: medium;"><span>When OpenAI introduced ChatGPT driven by a comprehensive LLM called "GPT-3.5" back in late 2022, it caught Google off-guard. Google was even more befuddled when OpenAI replaced GPT-3.5 with the more powerful model GPT-4 in early 2023; then OpenAI released an incessant stream of new capabilities for ChatGPT in the months that followed ... </span><span style="font-family: verdana;"><i>Note: The current edition of "Useful AI News" can be found </i></span><i><span style="background-color: white; color: #1a222a;">➡</span> <b><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html">HERE</a></b></i></span></div></span><p></p><p><span style="font-family: verdana;"><span></span></span></p><a name='more'></a><span style="font-family: verdana;">Google eventually introduced its Bard chatbot in March 2023 driven by a comprehensive LLM called "LaMDA". This was, at best, a weak response. However, Google subsequently released its own incessant stream of new capabilities for Bard in the months that followed. On 9/19/23, Google replaced Bard's original LLM with a more powerful model called "PaLM 2". </span><p></p><p><span style="font-family: verdana;">Now that Google has introduced so many improvements, do you think that Bard has finally caught up to ChatGPT? If you don't, how much more time do you think Google will need: three months, six months, another year?</span></p><p><span style="font-family: verdana;">The two articles featured this week provide the kind of data that readers will need in order to answer these questions for themselves. The first is Google's official timeline for its enhancements to Bard; the second is the TechCrunch timeline for OpenAI's enhancements to ChatGPT.</span></p><p><span style="font-family: verdana;">Note that Bard is free; users only need a Google account (also free); its URL = <a href="https://bard.google.com"><b><i>https://bard.google.com</i></b></a></span></p><p></p><ol style="text-align: left;"><li><span style="font-family: verdana;"> "Bard Experiment -- Updates", Google, 9/27/23 <b style="font-size: large;">➡</b> <b><i><a href="https://bard.google.com/updates">https://bard.google.com/updates</a></i></b><br />-- Google is currently testing Med-PaLM 2, a limited LLM that focuses on healthcare. A brief overview is presented in "Google’s Medical AI Chatbot: The Future of Healthcare Conversations", Walter Shields, LinkedIn, 7/15/23 <b style="font-size: large;">➡</b> <b><i><a href="https://www.linkedin.com/pulse/googles-medical-ai-chatbot-future-healthcare-walter-shields/">https://www.linkedin.com/pulse/googles-medical-ai-chatbot-future-healthcare-walter-shields/</a></i></b><br /><br />For more info <b><i><b style="font-size: large; font-style: normal;">➡</b><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#Other"> https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#Other</a></i></b><br /><br /></span></li><li><span style="font-family: verdana;"> "ChatGPT: Everything you need to know about the AI-powered chatbot", Alyssa Stringer and Kyle Wiggers, TechCrunch, 10/17/23 <b style="font-size: large;">➡</b> <b><i><a href="https://t.co/N9P9zfOIll">https://t.co/N9P9zfOIll</a></i></b><br /><br />For more info <b style="font-size: large;">➡</b> <b><i><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#OpenAI">https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html#OpenAI</a></i></b> </span></li></ol><p></p>Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-6781410263341661952023-05-31T15:01:00.023-04:002023-06-02T08:24:31.832-04:00Ezra Klein, "The Matrix", and Big Tech's current efforts to trivialize LLMs <div style="text-align: right;"><span style="font-family: verdana; font-size: x-small;"><i>Last update: 5/31/23 </i></span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEidoPv_Jk1v_1TJCBuYEMUdeF5_vRnYtduwWX_X1ULm4_9fW8noFg4WrvA-f8JDzUc_v0bvyzoWR8gBlBBSkAoG9tyej2g4mjiEox09JrqGlCXRUERWQPQuPuryaKVsdFTSzE_SYotJcGpfeaPrA3UTh50dfRQv4g6_n-CYmejb2Lx4bQBnUjQevT1N/s2731/TwoFacingHeadsBooksInBrainsPublicIntellectual.jpg" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="1536" data-original-width="2731" height="180" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEidoPv_Jk1v_1TJCBuYEMUdeF5_vRnYtduwWX_X1ULm4_9fW8noFg4WrvA-f8JDzUc_v0bvyzoWR8gBlBBSkAoG9tyej2g4mjiEox09JrqGlCXRUERWQPQuPuryaKVsdFTSzE_SYotJcGpfeaPrA3UTh50dfRQv4g6_n-CYmejb2Lx4bQBnUjQevT1N/s320/TwoFacingHeadsBooksInBrainsPublicIntellectual.jpg" vspace="10" width="320" /></a></div>Once again, the editor of this blog calls his readers' attention to some cogent observations about artificial intelligence (AI) from Ezra Klein, his favorite public intellectual. Mr. Klein's observations appeared in a recent NY Times op ed ==> "Beyond the ‘Matrix’ Theory of the Mind", Ezra Klein,<b><i><a href="https://www.nytimes.com/2023/05/28/opinion/artificial-intelligence-thinking-minds-concentration.html"> NY Times</a></i></b>, 5/28/23. As the editor suggested in a previous note on this blog, "<b><i><a href="https://neoskeptics.blogspot.com/2023/04/ezra-klein-public-intellectual-as-ai.html">Ezra Klein: the public intellectual as AI critic</a></i></b>", Klein is brilliant and well informed; but he is not an AI expert, so his incisive comments illuminate the limits of what can be understood about large language models (LLMs) and other AI technologies by non-experts.</span></div><span><span style="font-family: verdana;"></span></span><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><span></span><span><a name='more'></a></span><b><br /></b></span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b>Klein's critique</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">Klein's op ed doubts that ChatGPT, Bard, and other chatbots based on LLMs will increase the productivity of their users. The first part of his piece reminds readers that the Internet was also hyped as a productivity booster, but such gains were seldom realized. Indeed email, messaging, and other Internet social media have become notorious time sucks. Way back in the 1960s <b><i><a href="https://en.wikipedia.org/wiki/Herbert_A._Simon">Herbert Simon</a></i></b>, one of the founding fathers of AI, observed that when information becomes plentiful, attention becomes the scarce resource.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Turning to the super hyped LLMs, Klein disputes their productivity benefits because they embody flawed notions that gathering more information faster than ever would make us more productive. Readers who have seen any of the "Matrix" movies will recall that the heroes in those features acquired new knowledge instantly via injections. In real life one has to read and re-read and contemplate new material, a time consuming process, in order integrate new information into our personal knowledge bases. Ironically, the energetic Big Tech promoters of chatbots based on complex neural networks ignored the complexity of the real neural networks in the brains of their human users.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b>Dazzling first impressions ... followed by rueful second thoughts</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">Now let's give Klein's ironic screw a few more turns. When ChatGPT running GPT-3.5 was introduced in the fall of 2022, we were impressed; but when the underlying LLM was upgraded to GPT-4, we were dazzled because it seemed to know everything about everything ... until we learned that it sometimes rendered incorrect responses to questions, emitted biased responses, infringed on individual privacy, violated copyrights, and hallucinated from time to time. Beyond these shortfalls, it was easily hacked by spurious prompts. </span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Worst of all we learned that many of the most respected experts in the field thought there was at least a 10 percent chance that LLMs were on a development path that someday, much sooner than anyone had previously expected, we might lose our control of these synthetic minds; indeed, they might pose an existential threat to mankind.</span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="background-color: white; color: #1a222a; font-family: verdana;">"In Sudden Alarm, Tech Doyens Call for a Pause on ChatGPT", Will Knight and Paresh Daveand, <b><i><a href="https://www.wired.com/story/chatgpt-pause-ai-experiments-open-letter/?redirectURL=https%3A%2F%2Fwww.wired.com%2Fstory%2Fchatgpt-pause-ai-experiments-open-letter%2F" style="color: #336699; text-decoration-line: none;">Wired</a></i></b>, 3/29/23 <br />-- "Pause Giant AI Experiments: An Open Letter" ... from Elon Musk and others, <b><i><a href="https://futureoflife.org/open-letter/pause-giant-ai-experiments/" style="color: #336699; text-decoration-line: none;">Future of Life</a></i></b>, 3/29/23<br />-- This story also covered by<b><i><a href="https://www.theverge.com/2023/3/29/23661374/elon-musk-ai-researchers-pause-research-open-letter" style="color: #336699; text-decoration-line: none;"> The Verge</a>, </i></b></span><b style="background-color: white; color: #1a222a; font-family: verdana;"><i><a href="https://gizmodo.com/open-letter-musk-demand-pause-on-advanced-ai-1850275951?utm_source=twitter&utm_medium=SocialMarketing&utm_campaign=dlvrit&utm_content=gizmodo" style="color: #336699; text-decoration-line: none;">Gizmodo</a>, </i></b><b style="background-color: white; color: #1a222a; font-family: verdana;"><i><a href="https://techcrunch.com/2023/03/28/1100-notable-signatories-just-signed-an-open-letter-asking-all-ai-labs-to-immediately-pause-for-at-least-6-months/?tpcc=tcplustwitter" style="color: #336699; text-decoration-line: none;">TechCrunch</a><br /><br /></i></b></li><li><span style="background-color: white; color: #1a222a; font-family: verdana;">"OpenAI’s Altman and other AI giants back warning of advanced AI as ‘extinction’ risk", Natasha Lomas</span><span style="color: #1a222a;"><span style="background-color: white; font-family: verdana;">, </span><b style="background-color: white; font-family: verdana;"><i><a href="https://techcrunch.com/2023/05/30/ai-extiction-risk-statement/?tpcc=tcplustwitter" style="text-decoration-line: none;">TechCrunch</a></i></b><span style="background-color: white; font-family: verdana;">, 5/30 </span><span style="background-color: white; font-family: verdana;"><b>***</b><br />-- This story also covered by the <b><i><a href="https://www.nytimes.com/2023/05/30/technology/ai-threat-warning.html" style="text-decoration-line: none;">NY Times</a></i></b>, <b><i><a href="https://t.co/2oGXKrStwI" style="text-decoration-line: none;">Engadget</a></i></b>, <b><i><a href="https://t.co/iVDkCUkBy2" style="text-decoration-line: none;">The Verge</a></i></b>, <b><i><a href="https://www.wsj.com/articles/ai-threat-is-on-par-with-pandemics-nuclear-war-tech-executives-warn-39105eeb" style="text-decoration-line: none;">Wall Street Journal</a></i></b>, <b><i><a href="https://www.washingtonpost.com/business/2023/05/30/ai-poses-risk-extinction-industry-leaders-warn/" style="text-decoration-line: none;">Washington Post</a></i></b>, <b><i><a href="https://www.wired.com/story/runaway-ai-extinction-statement/?utm_social-type=owned&utm_source=twitter&mbid=social_twitter&utm_brand=wired&utm_medium=social" style="text-decoration-line: none;">Wired</a>, <a href="https://gizmodo.com/ai-chatgpt-extinction-warning-letter-openai-sam-altman-1850486688?utm_source=twitter&utm_medium=SocialMarketing&utm_campaign=dlvrit&utm_content=gizmodo" style="text-decoration-line: none;">Gizmodo</a>, <a href="https://arstechnica.com/information-technology/2023/05/openai-execs-warn-of-risk-of-extinction-from-artificial-intelligence-in-new-open-letter/?utm_brand=arstechnica&utm_source=twitter&utm_social-type=owned&utm_medium=social" style="text-decoration-line: none;">Ars Technica</a>, </i></b></span></span></li></ul></div><div style="text-align: left;"><span style="font-family: verdana;">Across the land, a rising chorus demanded that LLM development be subjected to government regulation. Big Tech seemed to support this demand (sometimes) ... Nevertheless, Big Tech accelerated the pace of its development. Indeed, in May 2023, first Google, then Microsoft rolled out splashy announcements at their annual conferences of upgrades for just about all of their software products and services, upgrades that featured glitzy new capabilities enabled by the insertion of LLM technology. But neither announced significant breakthroughs in resolving the problems that had triggered second thoughts, then strident demands for government regulation</span></div><div style="text-align: left;"><ul style="background-color: white; color: #1a222a; font-family: Verdana, Geneva, sans-serif; line-height: 1.4; list-style-image: initial; list-style-position: initial; margin: 0.5em 0px; padding: 0px 2.5em;"><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;"><b>Google I/O 2023</b> conference on 5/10/23 <br />-- Overviews of all announcements provided by <b><i><a href="https://t.co/Vch4mM3hom" style="color: #336699; text-decoration-line: none;">Gizmodo</a>, <a href="https://t.co/BpBRI2cojX" style="color: #336699; text-decoration-line: none;">TechCrunch</a>, <a href="https://www.wired.com/story/google-io-2023-everything-announced/?utm_brand=wired&mbid=social_twitter&utm_social-type=owned&utm_source=twitter&utm_medium=social" style="color: #336699; text-decoration-line: none;">Wired</a>, <br /></i></b>-- AI announcements were covered by<b style="font-style: italic;"> </b><b><i><a href="https://mashable.com/article/google-io-2023-everything-you-need-to-know-ai?utm_campaign=mash-com-tw-main-link&utm_contnet=tech&utm_medium=twitter&utm_source=social" style="color: #336699; text-decoration-line: none;">Mashable</a>, <a href="https://gizmodo.com/google-ai-chatbot-bard-search-maps-workspace-adobe-fire-1850424758?utm_source=twitter&utm_medium=SocialMarketing&utm_campaign=dlvrit&utm_content=gizmodo" style="color: #336699; text-decoration-line: none;">Gizmodo</a></i></b><br />-- "Google launches PaLM 2, its next-gen large language model", Frederic Lardinois, <b><i><a href="https://techcrunch.com/2023/05/10/google-launches-palm-2-its-next-gen-large-language-model/?tpcc=tcplustwitter" style="color: #336699; text-decoration-line: none;">TechCrunch</a></i></b>, 5/10/23<br />-- "Google jumps into the AI coding assistant fray with Codey and Studio Bot", Sanuel Axon, <b><i><a href="https://arstechnica.com/gadgets/2023/05/google-jumps-into-the-ai-coding-assistant-fray-with-codey-and-studio-bot/?utm_brand=arstechnica&utm_source=twitter&utm_social-type=owned&utm_medium=social" style="color: #336699; text-decoration-line: none;">Ars Technica</a></i></b>, 5/10/23<br /><br /></span></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;"><b>Microsoft Build 2023</b> conference 5/24/34 <b><i> <br /><span style="font-style: normal; font-weight: 400;">-- Keynote address = </span><span style="color: black;"><a href="https://www.engadget.com/watch-the-microsoft-build-2023-keynote-in-under-10-minutes-202055599.html?src=rss" style="color: #336699; text-decoration-line: none;">Engadget</a>, <a href="https://t.co/wO9SHdrS0g" style="color: #336699; text-decoration-line: none;">The Verge</a>,<i> </i><br style="font-style: normal; font-weight: 400;" /><span style="font-style: normal; font-weight: 400;">-- Fabric, a new data and analytics platform = </span><b style="font-style: normal;"><i><a href="https://techcrunch.com/2023/05/23/microsoft-launches-fabric-a-new-end-to-end-data-and-analytics-platform/?tpcc=tcplustwitter" style="color: #336699; text-decoration-line: none;">TechCrunch</a></i></b><span style="font-style: normal; font-weight: 400;">, </span><br style="font-style: normal; font-weight: 400;" /><span style="font-style: normal; font-weight: 400;">-- Plugins for AI apps = </span><b style="font-style: normal;"><i><a href="https://techcrunch.com/2023/05/23/microsoft-goes-all-in-on-plugins-for-ai-apps/?tpcc=tcplustwitter" style="color: #336699; text-decoration-line: none;">TechCrunch</a><br /></i></b><span style="font-style: normal; font-weight: 400;">-- Copilot for Windows 11 = </span><b style="font-style: normal;"><i><a href="https://t.co/ZQ8ECGKvuO" style="color: #336699; text-decoration-line: none;">Ars Technica</a>, <a href="https://www.theverge.com/2023/5/23/23732454/microsoft-ai-windows-11-copilot-build" style="color: #336699; text-decoration-line: none;">The Verge</a> </i></b></span></i></b></span></li></ul></div><div style="text-align: left;"><br /></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b>No disruptive black swans here ... yet</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">So why did Microsoft and Google produce such blatant displays of their strong commitment to LLMs in a context of rising public concerns? The editor of this blog suggests two reasons: their expectations of huge increases in short term profits and their hopes that widespread distribution of mundane upgrades would dampen their users' fears of LLMs and thereby decrease the public's demands for regulation.</span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">Regardless of the profound reservations of the most respected AI experts, </span><span style="font-family: verdana;">the market thinks that AI is the purest possible gold, the strongest indicator being the sudden rise to a trillion dollar valuation of Nvidia, the largest supplier of graphics processing units (GPUs), essential chips in the most powerful AI systems.<br /></span><span style="background-color: white; color: #1a222a; font-family: verdana;">-- See "Welcome to the trillion-dollar club, Nvidia", Alex Wilhelm, </span><b style="background-color: white; font-family: verdana;"><i><a href="https://techcrunch.com/2023/05/30/welcome-to-the-trillion-dollar-club-nvidia/?tpcc=tcplustwitter" style="text-decoration-line: none;">TechCrunch</a></i></b><span style="background-color: white; font-family: verdana;">, 5/30/23</span><span style="background-color: white; font-family: verdana;"><b>. </b></span><span style="background-color: white; font-family: verdana;">This story also covered by </span><b style="background-color: white; font-family: verdana;"><i><a href="https://t.co/O13ptxkNBW" style="text-decoration-line: none;">Gizmodo</a></i></b><span style="background-color: white; font-family: verdana;">, </span><b style="background-color: white; font-family: verdana;"><i><a href="https://www.cnbc.com/2023/05/30/nvidia-on-track-to-hit-1-trillion-market-cap-when-market-opens.html" style="text-decoration-line: none;">CNB</a>C</i></b><span style="font-family: verdana;"><span style="background-color: white;"> </span><span style="background-color: white;"><br /></span></span><br /></li><li><span style="font-family: verdana;">The upgrades were "mundane" in the sense that they were good enough to encourage most users to try them out, but they were not <b><i><a href="https://en.wikipedia.org/wiki/Black_swan_theory">black swans</a></i></b>. Users could implement the upgrades as slowly as they chose without fear of being overwhelmed by competitors who opted for faster adoption schedules. Users who double and triple checked the first drafts produced so quickly by these upgrades could use the edited drafts with confidence that they were not jeopardizing the viability of their operations. This was not the stuff of human extinction; so there was no need to rush towards restrictive regulations.</span></li></ul></div><div style="text-align: left;"><span style="font-family: verdana;">____________________________________</span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b>Links to related notes on this blog: </b></span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">"<b><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html">Useful AI News</a></b>", <i>Updated every day</i></span></li></ul></div>Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-14340757833443411842023-05-17T20:49:00.050-04:002023-06-02T08:25:32.446-04:00Developing reliable specialized large language models (LLMs) for the multi billion dollar online/hybrid education market<div style="text-align: right;"><span style="font-family: verdana; font-size: x-small;"><i>Last update: Friday 5/12/23 </i></span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><div class="separator" style="clear: both; text-align: center;"><span style="margin-left: 1em; margin-right: 1em;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjRCBQhpwXsw_GbCj1hYhpS47jFvC6hO-bgsScwGWMb8MAi0jWXEqQswTPo5i7A786UroQi7rMW5d8ukxtMYrcWRrzynRjvgb7WBpdUOw5q-MISxgVXl9ekDwUfe9HqDcsq3mlLD0Kqa4hVNclbFeNFw-C9aHU0DpjB4bPXQbJlZFlSIpUb15FQkuUL/s2508/DartHittingBullsEye-Specialization.jpg" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="1672" data-original-width="2508" height="213" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjRCBQhpwXsw_GbCj1hYhpS47jFvC6hO-bgsScwGWMb8MAi0jWXEqQswTPo5i7A786UroQi7rMW5d8ukxtMYrcWRrzynRjvgb7WBpdUOw5q-MISxgVXl9ekDwUfe9HqDcsq3mlLD0Kqa4hVNclbFeNFw-C9aHU0DpjB4bPXQbJlZFlSIpUb15FQkuUL/s320/DartHittingBullsEye-Specialization.jpg" vspace="10" width="320" /></a></span></div>Why did OpenAI release ChatGPT with the GPT-3.5 model, then with GPT-4, even though both models were prone to factual errors, bias, and "hallucinations". Open AI claimed that it did so, even at the risk of being banned in some countries (e.g., <b><i><a href="https://techcrunch.com/2023/04/28/chatgpt-resumes-in-italy/">Italy</a></i></b> temporarily), in order to learn from widespread user experience with the flawed models. </span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><span style="font-size: large;"><a name='more'></a></span><b><span style="font-size: medium;">A. Enhanced success of OpenAI's stock offering</span></b></span></div><div style="text-align: left;"><span style="font-family: verdana;">Whatever OpenAI's real motivations, a limited release would have attracted a tiny fraction </span><span style="font-family: verdana;">of the publicity that was generated by the widespread initial distribution, publicity that enabled OpenAi to quickly complete a stock offering that valued the company at $27 billion to</span><span style="font-family: verdana;"> $29 billion:</span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">"OpenAI closes $300M share sale at $27B-29B valuation", Jagmeet Singh, Ingrid Lunden, <b><i><a href="https://techcrunch.com/2023/04/28/openai-funding-valuation-chatgpt/?tpcc=tcplustwitter"><span style="color: #cc0000;">TechCrunch</span></a></i></b>, 4/28/23</span></li></ul></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b><br /></b></span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b>B. Ask me anything</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">The fact that ChatGPT acquired over 100 million users within a few months fueled the success of OpenAi's stock offering. Its users were dazzled by the apparent scope and depth of ChatGPT's knowledge. It seemed to know everything about everything. Yes, it sometimes made statements that were wrong, but so did most people. And yes, it sometimes made assertions that were value judgements with the same certainty as its statements of facts, but so did most people. But how many people could summarize the <b><i><a href="https://www.ushistory.org/us/18a.asp">Bill of Rights</a></i></b> -- the first ten amendments to the U.S. Constitution -- in the form of a Shakespearean sonnet within five seconds? Not many, if any, so ChatGPT's millions of users were dazzled. (Note: Here's a link to a <b><i><a href="https://neoskeptics.blogspot.com/p/chatgpt-summary-of-bill-of-rights.html">ChatGPT summary</a></i></b>)<br /></span></div><br /><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b><br /></b></span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b>C. Second thoughts</b></span></div><div style="text-align: left;"><div class="separator" style="clear: both; text-align: left;"><span style="font-family: verdana;">OpenAI explained that its GPT-4 model yielded better results than GPT-3 because "four" was based on far more data than "three". From this explanation, many users inferred that as the data files underlying subsequent models grew larger and larger, the remaining flaws in GPT-4 would be reduced, perhaps eliminated ... but would they? Sam Altman, OpenAI's CEO, recently rejected this inference:</span></div><div class="separator" style="clear: both; text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">"OpenAI’s CEO Says the Age of Giant AI Models Is Already Over" ... “We'll make them better in other ways", Will Knight, <a href="https://www.wired.com/story/openai-ceo-sam-altman-the-age-of-giant-ai-models-is-already-over/?utm_medium=social&utm_brand=wired&mbid=social_twitter&utm_social-type=owned&utm_source=twitter"><b><i>Wired</i></b></a>, 4/19/23</span></li></ul></div><div class="separator" style="clear: both; text-align: left;"><span style="font-family: verdana;">Unfortunately, Altman did not suggest alternative strategies. The rest of this note will discuss a collection of lucrative use cases that would be well served by smaller, more specialized models. The editor of this blog is not an AI expert, so his options will be sketched at a high, 10,000 foot conceptual level.</span></div><div class="separator" style="clear: both; text-align: left;"><span style="font-family: verdana;"><br /></span></div><div class="separator" style="clear: both; text-align: left;"><span style="font-family: verdana; font-size: medium;"><b><br /></b></span></div><div class="separator" style="clear: both; text-align: left;"><span style="font-family: verdana; font-size: medium;"><b>D. Focus on limited, legally accessible information</b></span></div><div class="separator" style="clear: both; text-align: left;"><span style="font-family: verdana;">Most of the use cases for GPT-4 reported in the media have involved chatbot responses to questions involving school or work-related related issues</span></div><div class="separator" style="clear: both; text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">In a non-work, non-school context, we might be comfortable with chatbots that seemed to know everything about everything, but sometimes uttered false statements, just like our real friends and associates, e.g., providing us with recipes that tasted horrible, suggesting physical workout schedules that weren't sustainable, or providing advice about relationships that was ineffective.<br /><br /></span></li><li><span style="font-family: verdana;">By contrast, our responsibilities as students or professionals in whatever field we studied or were employed would not require that our chatbot assistants know everything about everything; they would only have to know everything about the subjects related to our school assignments or work tasks.<br /><br /></span></li><li><span style="font-family: verdana;">Most users would prefer chatbots that based their assertions on data to which their users had legitimate access. Information that violated someone else's personal rights or copyrights/patents might be useful in the short term, but would entail longer term risks of costly penalties for users who we found themselves entangled in lawsuits or facing criminal charges</span></li></ul></div><div class="separator" style="clear: both; text-align: left;"><span style="font-family: verdana;">In the next section, the editor will propose an alternative cluster of options based on LLMs that would be much smaller than GPT-4's, but would produce more reliable results that were based on information that was acquired through voluntary, binding agreements with the information owners. These options were selected because of their similarities to the most dazzling use cases so far, <b>software engineering</b>, cases that have not been plagued by unreliable results, biases, or hallucinations.</span></div><div class="separator" style="clear: both; text-align: left;"><span style="font-family: verdana;"><br /></span></div><div class="separator" style="clear: both; text-align: left;"><span style="font-family: verdana;">ChatGPT can write reliable original code and debug flawed code submitted for its review. Why? Because its training data only contained a tiny percentage of error and was presented in a context that embodied minimum gender, racial, national, or other kinds of bias. </span></div><div class="separator" style="clear: both; text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;"><i>Note</i>: The editor is not suggesting that LLMs that respond to questions that did not relate to jobs or academic studies will not be developed. However, the biggest drivers of LLM development are hugely profitable ginormous corporations -- Google, Microsoft, Amazon, Baidu, Ali Baba -- all of whom want to use LLMs to earn the biggest earnings from the lowest investment costs while being subjected to the least regulations and incur minimum losses in lawsuits. In other words, the biggest developers will want to grab the most highly profitable low hanging fruit first.</span></li></ul></div><div class="separator" style="clear: both; text-align: left;"><span style="font-family: verdana;"><span><div class="separator" style="clear: both; font-family: Times;"><span style="font-family: verdana; font-size: medium;"><b><br />E. Reliable specialized LLMs</b></span></div><div class="separator" style="clear: both; font-family: Times;"><span style="font-family: verdana;">This discussion proposes that LLMs focus on STEM+M+D, law, and medicine, i.e., on </span><b style="font-family: verdana;"><span style="font-size: medium;">S</span></b><span style="font-family: verdana;">cience, </span><b style="font-family: verdana;"><span style="font-size: medium;">T</span></b><span style="font-family: verdana;">echnology, </span><b style="font-family: verdana;"><span style="font-size: medium;">E</span></b><span style="font-family: verdana;">ngineering, and </span><b style="font-family: verdana;"><span style="font-size: medium;">M</span></b><span style="font-family: verdana;">athematics plus the algorithmic components of </span><b style="font-family: verdana;"><span style="font-size: medium;">M</span></b><span style="font-family: verdana;">anagement science (e.g., supply chain management, risk assessment, project management), </span><b style="font-family: verdana;"><span style="font-size: medium;">D</span></b><span style="font-family: verdana;">ata science/analysis</span><span style="font-family: verdana;">, law, and medicine.</span></div><div class="separator" style="clear: both; font-family: Times;"><span style="font-family: verdana;"><br /></span></div><div class="separator" style="clear: both; font-family: Times;"><span style="font-family: verdana;">Publications from reputable publishers in STEM+M+D, law, and medicine share important qualities with articles and books about the software engineering component of technology, the "T" in STEM+M+D. Therefore LLMs trained on their publications should also be highly reliable.</span></div><div class="separator" style="clear: both; font-family: Times;"><ul><li><span style="font-family: verdana;">High quality publications are accessible via the Internet from reputable publishers for free or via paid subscriptions. Redistribution fees can usually be negotiated with the publisher/owners, so copyrights need not be infringed.<br /><br /></span></li><li><span style="font-family: verdana;">Individuals whose behavior is discussed in these publications are usually not identified, so privacy rights are not violated.<br /><br /></span></li><li><span style="font-family: verdana;">The high quality of the information in the publications from reputable publishers of textbooks and journal articles in STEM+M+ D, law, and medicine has been attained by rigorous peer reviews, editorial guidance, and extensive fact checking.</span></li></ul></div></span></span></div><div class="separator" style="clear: both; text-align: left;"><span style="font-family: verdana;"><span><b><span style="font-size: medium;"><br /></span></b></span></span></div><div class="separator" style="clear: both; text-align: left;"><span style="font-family: verdana;"><span><b><span style="font-size: medium;">F. Low hanging fruit</span> = The multi-billion dollar online/hybrid market </b><b>for academic degrees and professional certificates in STEM+M+D, law, and medicine</b><br />The most lucrative sectors of the online/hybrid education market offer degree and certificate programs for undergraduate students, graduate </span><span>students, and </span><span>practicing professionals. The highest tuition and fees are charged for programs in </span><span>STEM+M+D, law, and medicine; and the lion's share of online academic degree courses and online professional certificate courses are offered in these fields.</span></span></div><div class="separator" style="clear: both; text-align: left;"><ul style="text-align: left;"><li><b style="font-family: verdana;">Online</b><span style="font-family: verdana;"> </span><span style="font-family: verdana;">courses are fully online, i.e., they have no classroom components; whereas</span><span style="font-family: verdana;"> </span><b style="font-family: verdana;">hybrid</b><span style="font-family: verdana;"> </span><span style="font-family: verdana;">courses have substantial online and classroom components<br /><br /></span></li><li><span style="font-family: verdana;">It is widely conceded that the best face-to-face courses in the world are still better than the best online courses; and the best hybrid courses are best of all<br /><br />But the best face-to-face courses and the best hybrid courses are only accessible to students enrolled on the elite campuses wherein these courses are taught; by contrast, the best online courses are accessible to students anywhere in the world. Unfortunately, online courses require more self-discipline, better time management skills, and more effective study habits than face-to-face or hybrid courses<br /><br />However, the addition of reliable, specialized LLMs that provide customized one-on-one tutoring will help students in hybrid courses cover more material in between classroom sessions. Indeed, the deployment of reliable, specialized LLMs will facilitate the conversion of most traditional courses into hybrid formats. Furthermore, these instantly accessible LLM tutors will also make it easier for more students to become proficient learners in solitary online environments.<br /></span></li></ul></div><div class="separator" style="clear: both; text-align: left;"><span style="font-family: verdana;">In summary, we should anticipate that the deployment of reliable, specialized LLMs will greatly expand the market for online and hybrid courses in STEM+M+D, law, and medicine that are components of <span>degree and certificate programs for undergraduate students, graduate </span><span>students, and </span><span>practicing professionals. </span></span></div><div class="separator" style="clear: both; text-align: left;"><span style="font-family: verdana;"><br /></span></div><div class="separator" style="clear: both; text-align: left;"><span style="font-family: verdana;"><b>Question</b> -- How large will the online/hybrid education market be for academic degrees and professional certificates in </span><span style="font-family: verdana;">STEM+M+D, law, and medicine within the next few years</span><span style="font-family: verdana;">?</span></div><div class="separator" style="clear: both; text-align: left;"><span style="font-family: verdana;"><b><br /></b></span></div><div class="separator" style="clear: both; text-align: left;"><span style="font-family: verdana;"><b>Answer #1 </b>(low ball estimate)<b> -- At least $267 billion by 2027 </b></span></div><div class="separator" style="clear: both; text-align: left;"><span style="font-family: verdana;">According to a report posted by <a href="https://www.statista.com/outlook/dmo/eservices/online-education/worldwide" style="font-family: verdana; font-style: italic; font-weight: bold;">Statista</a>, the world-wide market for online courses should reach $167 billion in 2023 and rise by $72 billion, i.e., by 43%, to $239 billion by 2027. </span></div><div class="separator" style="clear: both; text-align: left;"><span style="font-family: verdana;"><br /></span></div><div class="separator" style="clear: both; text-align: left;"><span style="font-family: verdana;">These forecasts were calculated in October 2022, before the initial public distribution of GPT 3.5, followed by the far more impressive GPT 4. Accordingly, they do not embody the potential boost that more reliable specialized LLMs might bestow. If we assumed a modest 17 percent boost from better LLMs, the forecast for the online market would rise by 60%, i.e., $100 billion, to about $267 billion in 2027.</span></div><div class="separator" style="clear: both; text-align: left;"><span style="font-family: verdana;"><div class="separator" style="clear: both;"><br /></div><div class="separator" style="clear: both;"><b>Answer #2 </b>(a higher low ball estimate) <b>-- At least $534 billion by 2027</b></div><div class="separator" style="clear: both;">The previous estimates only covered online courses; they did not cover hybrid courses. At the present time online and hybrid courses occupy small shares of the total market for undergraduate degree, graduate degree, and professional certificate courses; the vast majority of such courses are still face-to-face. In principle, every face-to-face course could be upgraded to hybrid status by adding access to LLM's. But there won't be much movement in that direction with error-prone, biased, hallucinating LLM's like GPT 4.</div><div class="separator" style="clear: both;"><br /></div><div class="separator" style="clear: both;">However the development of reliable, specialized LLMs is likely to trigger tsunamis of upgrades in STEM+M+D, law, and medicine courses. Indeed, students in these fields will demand access to the new LLMs. Colleges, universities, and other educational entities that don't provide such access will be penalized by substantial reductions in their share of their sectors of the education market.</div><div class="separator" style="clear: both;"><br /></div><div class="separator" style="clear: both;">Nevertheless, out of an abundance of caution, we propose another modest, low ball estimate. We suggest that the number of students enrolled in hybrid courses that provide access to reliable specialized LLMs in 2027 will be at least as many as those enrolled in online courses, i.e., a $267 billion market for hybrid courses. Therefore the combined hybrid/online market will be at least $267 + $267 = $534 billion</div><div class="separator" style="clear: both;"><br /></div></span></div></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b><br /></b></span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b>G. "Will I hallucinate?"</b></span></div><div style="text-align: left;"><span style="font-family: verdana;">One of the most troublesome of GPT-4's deficiencies has been its occasional emission of "hallucinations", i.e., wildly invalid statements that seem to come from out of nowhere. There is reason to believe that ordinary errors and bias reflect errors and bias in the data on which GPT-4 was trained. But where do hallucinations come from?</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">Readers who have seen Stanley Kubrick's classic film "2001" will remember the touching scene wherein HAL -- the AI system that supported all of the astronauts' activities -- learned that it will be turned off. HAL asked anxiously, "Will I dream?" Hal's cyber genius developer calmly responded that Hal will dream because all intelligent entities dream. The corresponding question for a reliable, specialized LLM would be </span><span style="font-family: verdana;">"Will I hallucinate?" Only time will tell, time and extensive testing before its responsible developers release it to the public.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">____________________________________</span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b>Links to related notes on this blog: </b></span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">"<b><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html">Useful AI News</a></b>", <i>Updated every day</i></span></li></ul></div>Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0tag:blogger.com,1999:blog-8043580455676201364.post-3073422106241567872023-04-23T16:52:00.014-04:002023-04-24T14:43:43.880-04:00Ezra Klein: the public intellectual as AI critic<div style="text-align: right;"><span style="font-family: verdana; font-size: x-small;"><i>Last update: Sunday 4/23/23<div class="separator" style="clear: both; text-align: center;"><span style="margin-left: 1em; margin-right: 1em;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiKndUQwKA67bOvclBaCRI4ABXiGAoHF4EoI_q_nR4MrcGPGj3g2yC7r1k2K860fJV5g_IvLin-_gNVURG7HaioNdMaQn3Rmju5-VIL2BwODDWJDvfhXUdrzAC-A5KsngvGZdDV7KPLCVjeV18Xx6HbT-1L9NKLjqr0OR_JcK4A2U1KaGowGpHE0_Sl/s2731/TwoFacingHeadsBooksInBrainsPublicIntellectual.jpg" style="margin-left: 1em; margin-right: 1em;"><img align="left" border="0" data-original-height="1536" data-original-width="2731" height="180" hspace="10" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiKndUQwKA67bOvclBaCRI4ABXiGAoHF4EoI_q_nR4MrcGPGj3g2yC7r1k2K860fJV5g_IvLin-_gNVURG7HaioNdMaQn3Rmju5-VIL2BwODDWJDvfhXUdrzAC-A5KsngvGZdDV7KPLCVjeV18Xx6HbT-1L9NKLjqr0OR_JcK4A2U1KaGowGpHE0_Sl/s320/TwoFacingHeadsBooksInBrainsPublicIntellectual.jpg" vspace="10" width="320" /></a></span></div> </i></span></div><div style="text-align: left;"><span style="font-size: medium;"><span style="font-family: verdana;">According to <b><i><a href="https://en.wikipedia.org/wiki/Intellectual#Public_intellectual">Wikipedia</a></i></b>, "The term public intellectual "describes the intellectual participating in the public-affairs discourse of society, in addition to an academic career.</span><span style="font-family: verdana;">" The editor of this blog frequently cites Wikipedia definitions of important concepts, but he was surprised and disappointed by the narrow academic focus of its definition in this case. <b><i><a href="https://en.wikipedia.org/wiki/Ezra_Klein">Ezra Klein</a></i></b> is not a tenured professor in a university and does not have a PhD, the union card usually required for admission to tenure. </span><span style="font-family: verdana;">But neither did </span><b style="font-family: verdana;"><i><a href="https://en.wikipedia.org/wiki/Walter_Lippmann">Walter Lippmann</a></i></b><span style="font-family: verdana;">, one of this nation's first and most eminent public intellectuals </span></span></div><div style="text-align: left;"><span><a name='more'></a></span></div><div style="text-align: left;"><span style="font-family: verdana;"><br />Ezra Klein is a journalist with a bachelors degree from a prestigious university, same as Lippmann. Like Lippmann in the early 20th century, Klein has been an incisive explorer of issues that were or should have been at the top of our public agenda in the early 21st century. And like Lippmann, Klein has striven mightily to explain complex issues to concerned voters in order to enable these voters to make informed decisions when they go to the polls. In recent months, Klein has confronted the complex challenges posed by ChatGPT and related technologies in his columns and podcasts for the New York Times, a few of which are cited in the following bullets:</span></div><div style="text-align: left;"><ul style="background-color: white; color: #1a222a; line-height: 1.4; list-style-image: initial; list-style-position: initial; margin: 0.5em 0px; padding: 0px 2.5em;"><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;">"The Imminent Danger of A.I. Is One We’re Not Talking About", Ezra Klein, <b><i><a href="https://www.nytimes.com/2023/02/26/opinion/microsoft-bing-sydney-artificial-intelligence.html" style="color: #336699; text-decoration-line: none;">NY Times</a>,</i></b> 2/26/23<br /><br /></span></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;">"A.I. is about to get much weirder: Here's what to watch for", Guest = Kelsey Piper (VOX), <b><i><a href="https://www.nytimes.com/2023/03/21/opinion/ezra-klein-podcast-kelsey-piper.html" style="color: #336699; text-decoration-line: none;">The Ezra Klein Show</a></i></b> (podcast with transcript), 3/21/23<br /><br /></span></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;">"Why A.I. Might Not Take Your Job or Supercharge the Economy", Guest = Ezra Klein. Klein answers listeners' questions about ChatGPT, <b><i><a href="https://www.nytimes.com/2023/04/07/opinion/ezra-klein-podcast-ama-april2023.html" style="color: #336699; text-decoration-line: none;">The Ezra Klein Show</a></i></b> (podcast with transcript), 4/7/23 <br /><br /></span></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;">"A.I. Vibe Check With Ezra Klein, and Kevin Tries Phone Positivity", Ezra Klein is the guest, <b><i><a href="https://www.nytimes.com/2023/04/07/podcasts/ai-vibe-check-with-ezra-klein-and-kevin-tries-phone-positivity.html?action=click&module=audio-series-bar&region=header&pgtype=Article" style="color: #336699; text-decoration-line: none;">HardFork</a></i></b> (podcast with transcript), 4/7/23 <br /><br /></span></li><li style="border: none; margin: 0px 0px 0.25em; padding: 0px;"><span style="font-family: verdana;">"What Biden's Top AI Thinker Concluded We Should Do", Guest = Alondra Nelson (former deputy director and acting director of the Biden White House’s Office of Science and Technology Policy), <b><i><a href="https://www.nytimes.com/2023/04/11/opinion/ezra-klein-podcast-alondra-nelson.html" style="color: #336699; text-decoration-line: none;">The Ezra Klein Show</a></i></b> (podcast with transcript), 4/11/23<i><br /></i></span></li></ul></div><div style="text-align: left;"><div class="separator" style="clear: both; text-align: left;"><span style="background-color: white; color: #1a222a; font-family: verdana;">Ezra Klein is brilliant and well informed; but he is not an AI expert, so his incisive comments illuminate the limits of what can be understood about large language models (LLMs) and other AI technologies by non-experts. In other words, he embodies the outer limits of what concerned voters need to know and what concerned voters should expert their public officials to know when they develop policies for regulating these technologies.</span></div><div class="separator" style="clear: both; text-align: left;"><br /></div></div><div style="text-align: left;"><span style="font-family: verdana;">As Klein has pointed out again and again, we cannot let profit-oriented corporations -- like Google, Microsoft, Amazon, and Facebook -- regulate themselves in the development of AI because </span><span style="font-family: verdana;">profit-oriented</span><span style="font-family: verdana;"> </span><span style="font-family: verdana;">corporations will use these powerful technologies to persuade the members of our society to buy products and services that they may not need, that might actually be harmful their users and, far worse, that might have unanticipated dire consequence for our entire society.</span></div><div style="text-align: left;"><span style="font-family: verdana;"><br /></span></div><div style="text-align: left;"><span style="font-family: verdana;">____________________________________</span></div><div style="text-align: left;"><span style="font-family: verdana; font-size: medium;"><b>Links to related notes on this blog: </b></span></div><div style="text-align: left;"><ul style="text-align: left;"><li><span style="font-family: verdana;">"<b><i><a href="https://neoskeptics.blogspot.com/p/last-update-sunday-319-23-this.html">Useful AI News</a></i></b>" ... This twice-weekly news guide tracks <a style="background-color: #fcfcfb; color: #2f6b9a; cursor: pointer; font-family: "Source Sans Pro", "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 16px; outline: none 0px; white-space: pre-wrap;" tabindex="0">#LargeLanguageModels</a><span face=""Source Sans Pro", "Helvetica Neue", Helvetica, Arial, sans-serif" style="background-color: #fcfcfb; color: #241f21; font-size: 16px; white-space: pre-wrap;"> and other useful AI tools</span>, <i>Updated twice per week</i></span></li></ul></div>Roy Bhttp://www.blogger.com/profile/00238618203707732616noreply@blogger.com0