AW-0 unveiled: first world action model that plans in code, predicts video, controls robots
trevordarrell · x · 2026-10-27
Introducing AW-0, an Agentic World Action Model—the first world action model that plans in code, predicts the future, and controls the robot in one generative model. Built on a Mixture-of-Transformers backbone, it jointly generates explicit skill programs (code), predictive visual dynamics (video), and continuous motor control (actions). Code makes skill calls, object arguments, and control flow explicit; the video expert attends to code and the action expert to both, letting semantic plans shape physical predictions with a path to execution.
More from Embodied
- Wayve CEO argues end-to-end deep learning should power driving, safety and world models — alexgkendall · 2026-10-27
- Mundo turns a few room photos into an explorable, simulatable 3D world — at 1.5 FPS — mark_k · 2026-10-27
- Zurich Physical AI Hackathon puts attendees in a room with $1.5M of humanoids — ___Mufasaa · 2026-10-27
- Apple responds after iPhone Duo inner displays crack along the hinge — mobileraj · 2026-10-27
- VLA models encode task concepts as curved multi-dimensional geometry, 19-author study finds — ninamiolane · 2026-10-27
- Munro Live Posts Full Cybercab Engineering Deep-Dive Video — yunta_tsai · 2026-10-27