Richard Sutton – Father of RL thinks LLMs are a dead end - YouTube

youtube.com · 9,694 words · saved by 1 readers

Reinforcement learning expert Richard Sutton, a Turing Award winner, debates the future of AI. The interview explores Sutton's perspective that LLMs lack the capacity for true on-the-job learning, contrasting their limitations with the adaptable nature of biological intelligence. Expect a lively discussion challenging current AI paradigms. Follow along using the transcript. Reinforcement learning expert Richard Sutton, a Turing Award winner, debates the future of AI. The interview explores Sutton's perspective that LLMs lack the capacity for true on-the-job learning, contrasting their limitations with the adaptable nature of biological intelligence. Expect a lively discussion challenging current AI paradigms. Follow along using the transcript.

Richard Sutton is the father of reinforcement learning, winner of the 2024 Turing Award, and author of The Bitter Lesson. And he thinks LLMs are a dead end. After interviewing him, my steel man of Richard’s position is this: LLMs aren’t capable of learning on-the-job, so no matter how much we scale, we’ll need *some* new architecture to enable continual learning. And once we have it, we won’t need a special training phase — the agent will just learn on-the-fly, like all humans, and indeed, like all animals. This new paradigm will render our current approach with LLMs obsolete. In our…

Richard Sutton – Father of RL thinks LLMs are a dead end - YouTube

related reading