CoRL best paper SuperMimic faces similarity questions; RWM author explains why it resonated
chris_j_paxton · x · 2026-11-14
SuperMimic won CoRL's best paper award, but many noted similarities to Chris Paxton's (Uncertainty-Aware) Robotic World Models (RWM/RWM-U). Asked repeatedly at CoRL 2026, the original author responded publicly.
He argued SuperMimic is not the first to make model-based RL work on real robots, nor are its core ideas or tasks entirely new. But it resonated for good reasons: a compelling motivation and a strong demonstration of lowering the computational barriers to training large models like SONIC — a timely, practically important emphasis on training efficiency that was a much harder direction to pursue earlier.
More from Embodied
- PartCAD: open-source package manager turns hardware specs into CAD, BOMs and AI-agent-readable docs — IgorCarron · 2026-11-14
- TRIMI wearable robot hand at CoRL pulls credit cards from a wallet with ease — chris_j_paxton · 2026-11-14
- Lemonade Expands FSD-Based Tesla Insurance to 7 States With 50% Off Autonomous Miles — XFreeze · 2026-11-14
- Open-source AI robot learning tool wins Japan e-Learning Award's Physical AI category — 4310sy · 2026-11-14
- A rare German film asks what home care looks like once a humanoid robot moves in — CyberRobooo · 2026-11-14
- Walden talk: base models improving fast, mid-training data commoditizing — ZeYanjie · 2026-11-14