AlphaGo Veteran Warns Today's LLMs Can't Actually Reason

A core AlphaGo team member who just left Google DeepMind argues LLMs still run on intuition alone and calls for AlphaGo-style reasoning architectures.