Re: [FRIAM] ChatGPT is not very smart...

Thanks for the pointer to this article. I found the new Bing chat engine quite impressive. To a surprising extent it felt like the author was interacting with an actual human being. The following brief exchange doesn't sound or feel like it was generated by stringing together words found during

[FRIAM] Whisper, a speech-to-text prrogram based on GPT-3

I downloaded Whisper and tried it out. I fed it a 20-minute screencast I did about 10 years ago. There are about four levels you can choose which trade accuracy for speed. I tried the recommended level, down one from the highest. After a substantial wait, I got the results. First, the input

Re: [FRIAM] A very good piece about ChatGPT.

I'm curious why it's unanticipated. Is this a (very mild) form of the Nobel Effect: https://en.wikipedia.org/wiki/Nobel_Prize_effect? The idea that one's been successful in some persnickety sub-domain triggers some dopamine and makes them think they'll achieve easier success in some other

Re: [FRIAM] ChatGPT is not very smart...

I don't think it is necessarily the case a very large neural net would need to backtrack. It could devote more and more resources to different hypotheses so long as it was not energy constrained. -Original Message- From: Friam On Behalf Of Santafe Sent: Friday, February 10, 2023 3:11

Re: [FRIAM] ChatGPT is not very smart...

This was laugh out loud funny for me. YMMV. Arguing with AI: My first dispute with Microsoft’s brilliant and boneheaded Bing search engine

Re: [FRIAM] A very good piece about ChatGPT.

Very nice piece by Rich Sutton! Re NanoGPT. Even though I'm retired, I wanted to learn more about how the LLMs work. So I started a project in which a number of students and i work through Karpathy's nanoGPT. We've just started, but I think it will be worth the effort. Karpathy uses non-trivial

Re: [FRIAM] ChatGPT is not very smart...

From the New Yorker article Russ posted: "If a compression program knows that force equals mass times acceleration, it can discard a lot of words when compressing the pages about physics because it will be able to reconstruct them." ... "Perhaps arithmetic is a special case, one for which large

Re: [FRIAM] A very good piece about ChatGPT.

Interesting article. I have found this link to "GPT in 60 Lines of NumPy" on Hacker Newshttps://jaykmody.com/blog/gpt-from-scratch/It is similar to nanoGPT and minGPT from Andrej Karpathy, who has joined OpenAI again recentlyhttps://github.com/karpathy/nanoGPTThe secret is apparently not to use

Re: [FRIAM] A very good piece about ChatGPT.

Thanks for the share. This article explains things very well from aspects I had never considered. On Fri, Feb 10, 2023 at 10:23 PM Russ Abbott wrote: > It's in the New Yorker > > and intended for a

[FRIAM] A very good piece about ChatGPT.

It's in the New Yorker and intended for a non-scientific audience. -. --- - / ...- .- .-.. .. -.. / -- --- .-. ... . / -.-. --- -.. . FRIAM Applied Complexity Group listserv Fridays 9a-12p Friday St. Johns

Re: [FRIAM] ChatGPT is not very smart...

So probably this has all been analyzed to death already, by people who, unlike me, know what they are talking about. But in re-reading it, I feel like the structure of the problem is characterizable. It is as if “facts” that are constraints on the sentences to be composed are ordered in a