AnalysisAI ModelsSeptember 24, 2026

Manning: world models need causality, not pretty pixels

Read original source →youtube.com

Christopher Manning, Stanford professor and former Stanford AI Lab director now at Moonlake AI, argues world models require causal understanding rather than visual fidelity. He notes Google had a language model trained on 2 trillion tokens in 2007, the same order of magnitude as today's frontier models.

People · Christopher Manning

1 source

More stories today

Open the live feed