Yes, the guy with a PhD in Machine Intelligence, co-author of Reinforcement Learning: An Introduction, which is universally considered the bible of the field, recipient of the AAAI fellowship award and the Turing Award, and the inventor of Temporal Difference Learning doesn't know what he's talking about.

Surprisingly enough, Turing Award winner and father of reinforcement learning Richard Sutton knows perfectly well what he's talking about. The whole talk is about the need to have the ability to test novel outputs against reality and iterate to find ones that are good. This is exactly what Claude Code, the agent framework, adds to Claude, the LLM, to allow it to find novel coding solutions that actually work.