I've been compiling papers on Physical AI — the intersection of foundation models and robotics. This covers Vision-Language-Action (VLA) models like RT-2 and π₀, world models (DreamerV3, Genie 2, JEPA), diffusion policies, real-world deployment and latency problems, cross-embodiment transfer, scaling laws, and safety/alignment for robots.

The field has exploded in the past 18 months. We went from "lets try llms on robotics" to having so many dimensions to optimize for. so felt right to maintain a running list of resources.

Organized by: foundations → architectures → action representations → world models → learning paradigms → deployment → applications.

Contributions welcome — especially corrections and missing papers.