What is Generative AI?
& Some Definitions
lecun-20230324-nyuphil.pdf
For those of you interested to hear a different opinion about LLM and to understand why, have a look at this presentation from: Yann Le Cun (Chief AI Scientist at Meta) : he is bold and not popular in his affirmation: "Auto-Regressive LLMs are doomed. They cannot be made factual, non-toxic, etc. They are not controllable Probability e that any produced token takes us outside of the set of correct answers"... He backs it up with a scientific analysis of the models and proposes a Cognitive Architecture
For those of you interested to hear a different opinion about LLM and to understand why, have a look at this presentation from: Yann Le Cun (Chief AI Scientist at Meta) : he is bold and not popular in his affirmation: "Auto-Regressive LLMs are doomed. They cannot be made factual, non-toxic, etc. They are not controllable Probability e that any produced token takes us outside of the set of correct answers"... He backs it up with a scientific analysis of the models and proposes a Cognitive Architecture
capable of reasoning & planning. Here is his public deck presented at NYU. Interesting and understandable demonstration.
capable of reasoning & planning. Here is his public deck presented at NYU. Interesting and understandable demonstration.