Just playing a bit of a Devil's advocate role here: Chollet and LeCun are, undoubtedly, crucial figures in deep learning landscape, but where is their participation? At least GPT models are able to generate coherent structures and change styles, not unlike modern image generation.
The probability formula LeCun came up with, though, is not unlike a bucket of iced water considering its implications.
Given how high up Chollet and LeCun are at their orgs, I wouldn't be surprised if they have been extremely involved with some of the higher profile models. Keep in mind- Meta, Google, and other big Tech was working with LLMs before OpenAI took the scene. OpenAI just got theirs to mass market earlier.
Specifically about their critiques here- I use it more to back up my claim of architectural problems with LLMs. They both state 2 of the issues I had highlighted in my Why ChatGPT Lies article, just in less detail.
Your article is brilliant, actually. It puts things into a very interesting perspective and helps with using LLMs consciously. So far, it's not intelligence; it's a tool - useful but harmful if used incorrectly. It needs proper instructions, though.
Just playing a bit of a Devil's advocate role here: Chollet and LeCun are, undoubtedly, crucial figures in deep learning landscape, but where is their participation? At least GPT models are able to generate coherent structures and change styles, not unlike modern image generation.
The probability formula LeCun came up with, though, is not unlike a bucket of iced water considering its implications.
Given how high up Chollet and LeCun are at their orgs, I wouldn't be surprised if they have been extremely involved with some of the higher profile models. Keep in mind- Meta, Google, and other big Tech was working with LLMs before OpenAI took the scene. OpenAI just got theirs to mass market earlier.
Specifically about their critiques here- I use it more to back up my claim of architectural problems with LLMs. They both state 2 of the issues I had highlighted in my Why ChatGPT Lies article, just in less detail.
Your article is brilliant, actually. It puts things into a very interesting perspective and helps with using LLMs consciously. So far, it's not intelligence; it's a tool - useful but harmful if used incorrectly. It needs proper instructions, though.