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

Name: Richard Sutton – Father of RL thinks LLMs are a dead end
Description: 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.

Dwarkesh Patel

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.

HN commenters cite it as the serious skeptic's case against LLMs reaching general intelligence, usually in debates about whether scaling is enough.

6 Hacker News mentions · 2025–2026

Topics

LLMs & generative AI
Machine learning & deep learning
AI safety, AGI & impact

Watch on YouTube

Explore more videos