Don’t be fooled—LLMs don’t reason

Current LLM "reasoning" via chain-of-thought is pattern prediction, not genuine deliberation, says ex-AlphaGo architect Thore Graepel.

• Move 37 explained: AlphaGo's famous 2016 move (a 1-in-10,000 chance play by human standards) came from explicit tree search, not intuition alone, author Thore Graepel argues.
• Two-system architecture: AlphaGo paired a fast 'System 1' policy network with a slow 'System 2' search mechanism that tested hunches against future consequences.

Themes, sentiment and reasons are AI-written paraphrases, checked so they never quote an episode. They may contain errors. Not investment advice. Episodes remain the property of their creators.