Richard Sutton – Father of reinforcement learning thinks LLMs are a dead end [video] | Hacker News Reader