Last update: Monday 2/12/24
A. TL;DR ... Top 4 stories in past week ...
1) Google, Apple, Meta, Microsoft, Amazon, and everybody else join US consortium to advance responsible AI
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:
"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,” said Gina Raimondo, Secretary,U.S. Dept. of Commerce
High sounding words, but so generic/nonspecific that we won't know what the consortium will do until it actually does something.
2) The Federal Communications Comission banned A.I.-Generated Robocalls
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:
"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. "
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.
3) Sam Altman Seeks Trillions (or was that bazillions???) of Dollars to Reshape Business of Chips and AI
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.
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. As such, his pronouncements about the future of AI technology should not be taken seriously. But as a super salesman, 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.
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:
- "The Inside Story of Microsoft’s Partnership with OpenAI", Charles Duhigg, New Yorker, 12/1/23
4) Google runs Microsoft's playbook ... very badly
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
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.
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.
Google: 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.
Bystander: 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.
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???
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.
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 Android 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?
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.
B. Top 4 stories in past week ...
- Public Policy
"Google, Apple, Meta and other huge tech companies join US consortium to advance responsible AI", Sharon Goldman, Engadget 2/8/24
-- This story also covered by VentureBeat ... and the U.S. Dept of Commerce - Public Policy
"F.C.C. Bans A.I.-Generated Robocalls", Cecilia Kang, NY Times, 2/8/24 ***
-- This story also covered by CNN, Forbes, NPR ... and the FCC - OpEds
"(Dr. Evil Voice) Sam Altman Wants 7 Trillion Dollars", Michelle Cheng, Gizmodo, 2/9/24 ... *** ... LOL, ROTF, LMFAO ... :-)
-- This story also covered by Wall Street Journal, Ars Technica, Computer World - Google
"Google rebrands Bard chatbot as Gemini, rolls out paid subscription", Jeffrey Dastin, Reuters, 2/8/24 ***
-- This story also covered by VentureBeat, NY Times, ZDNet, Engadget, MIT Tech Review
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.
No comments:
Post a Comment
Your comments will be greatly appreciated ... Or just click the "Like" button above the comments section if you enjoyed this blog note.