Developing reliable specialized large language models (LLMs) for the multi billion dollar online/hybrid education market

Last update: Friday 5/12/23 

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., Italy temporarily), in order to learn from widespread user experience with the flawed models. 

Ezra Klein: the public intellectual as AI critic

Last update: Sunday 4/23/23

 

According to Wikipedia, "The term public intellectual "describes the intellectual participating in the public-affairs discourse of society, in addition to an academic career." 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. Ezra Klein is not a tenured professor in a university and does not have a PhD, the union card usually required for admission to tenure. But neither did Walter Lippmann, one of this nation's first and most eminent public intellectuals 

The Neoskeptics editor asked the GPT-4 Oracle: "What's the next big thing after Large Language Models (LLMs)?" -- DRAFT

Last update: Saturday 4/15/23 

Once again, the editor of this blog has asked the GPT-4 oracle to look into its crystal ball. Of course, an LLM can't analyze anything or predict anything; but it should be able to produce a plausible distillation of the consensus of opinions among the small community of experts who have published the most papers on the Internet that addressed the issues raised by his questions. Like millions of other "well informed" people, the editor had been surprised, nay flabbergasted by the sudden appearance of ChatGPT last fall running GPT-3.5, and again a few weeks ago when the new GPT-4 model was introduced. 

ChatGPT's response to our 3rd question: "What do I have to learn in order to become a competent user of applications based on large language models?"

Last update: Monday 4/10/23 

On 4/10/23 the editor of this blog posed the question in the title of this note to ChatGPT running its GPT-4 model.  In the editor's opinion, the chatbot's detailed response lays out a realistic learning program; it is not a "hallucination".

Note that this is the editor's second revision of a question he originally posed to ChatGPT on 4/8/23, revised yet again because this third version is a simpler, more effective "prompt" to ChatGPT to produce guidance for obtaining the knowledge and skills that most non-AI experts really want ... :-)

"Pandemic Survivor" -- Does anybody still watch that show?

Last update: Tuesday 7/26/22 

Is that show still on? Does anybody really watch it anymore? I haven't watched since last Christmas. Who are the main characters now? Remember way back in 2020 when  "Masks" and "Other Mitigations" were the good guys and we all tuned in every week to watch them flatten the curves?... "Masks" and "Other Mitigations" were pushed out in early 2021 by "Vaccines" and "Herd Immunity". When ratings began to sag, CDC Productions replaced them with bad guys, "Delta" and "Breakthroughs". Nobody like them, so CDC tried to bring back "Masks"and "Other Mitigations", which didn't work out. After Christmas they replaced "Delta" with "Omicron". He was supposed to be a good guy because he was less lethal than Delta, yet somehow he killed more people. I remember feeling confused and bored and suddenly noticed that I had gained 15 pounds from watching too much TV; so I decided to go out more and get more exercise and fresh air. That's when I learned that Netflix announced that it would cancel its contract with CDC Productions before the fall 2022 season began. Meanwhile, is anybody still watching the show?