ETHERSON@LiamEtherson
Instead of committing to the first action, imagine an agent testing a few possible outcomes before making its move. That's what makes PAN 2 interesting. It can simulate different actions, compare what happens next, and choose the option most likely to work. A useful direction…
ETHERSON@LiamEtherson
Changing a strawberry into a rubber duck sounds like a simple visual edit, but the interesting part is what happens next. The splash changes because the object and its physical properties are different. PAN 2 is working towards models that understand these interactions, not just…
ETHERSON@LiamEtherson
The interesting part of PAN 2 is how it separates understanding from visualization. The LLM figures out what should happen and in what order, while the diffusion model brings those ideas into a visual simulation. This combination makes it possible for agents to reason through…