Fei Fei Li: The Race to Build World Models For AI

On the path to spatial intelligence, generating pixels that are truly spatially contextualized and grounded, that is the very hard step that Atlas has taken. We've seen video models as being built on next frame prediction.

It goes beyond the title with direct discussion of like, kind, model, including: We know LLMs are built on next token prediction.

1. 01On the path to spatial intelligence, generating pixels that are truly spatially contextualized and grounded, that is the very hard step that Atlas has taken
2. 02We've seen video models as being built on next frame prediction
3. 03Each time we made the model bigger and each time we trained it for longer, it got significantly better