limits

Dwarkesh Patel interviews Richard Sutton, reinforcement learning pioneer, Turing Award winner and author of The Bitter Lesson. Sutton argues LLMs are a dead end because they cannot learn on the job, and that continual learning from experience will require a new architecture.

A 1983 BBC interview in which Richard Feynman is asked how magnets work and explains why 'why' questions are hard to answer in everyday terms. He reflects on what counts as an explanation and the limits of analogy.

A short clip of Richard Feynman answering an interviewer's question about why magnets repel, explaining that a 'why' question depends on a shared framework of understanding. It illustrates how hard it is to explain fundamentals to a layperson.

A short clip of Richard Feynman answering an interviewer's question about how magnets work, in which he explains why a why-question has no simple answer without a shared frame of reference.

A clip from the BBC Fun to Imagine series in which Richard Feynman is asked how magnets work and explains why 'why' questions depend on a shared framework of understanding and cannot always be answered in layman's terms.

Dwarkesh Patel interviews OpenAI chief scientist Ilya Sutskever on time to AGI, what comes after generative models, data limits, alignment, and competing with Google. The title centers on whether next-token prediction could surpass human intelligence.