LeCun has been arguing for years that predicting text isn’t enough for real intelligence. This startup is basically his attempt to prove that world models — not

LeCun has been arguing for years that predicting text isn’t enough for real intelligence. This startup is basically his attempt to prove that world models — not bigger LLMs — are the path to AI that can reason about the physical world.