Galaxea Co-founder Hang Zhao's CoRL Keynote: Split Robot Intelligence into System 2/1/0 by Timescale
YuXiang_IRVL · x · 2026-11-13
At CoRL 2026, Galaxea co-founder and Tsinghua assistant professor Hang Zhao argued in an Early Career Keynote that robot intelligence should be split by timescale, not by perception/planning/control.
- Core claim: the physical world doesn't pause while a robot computes — inference latency, not the controller, sets the pace.
- Layered architecture: System 2 for seconds-scale reasoning (Fast-WAM), System 1 policies at tens of Hz (G0.5), System 0 contact control at hundreds of Hz (ActionCodec2).
- Key number: Fast-WAM cuts single-step inference from 800ms to 190ms, a 4x+ speedup.
A clear framework for multi-model, multi-timescale embodied intelligence — recommended for anyone following robot systems.
More from Embodied
- DeepMind unveils Gemini Robotics 2.5: robots tie bows autonomously, 96% pick-and-pack success at Mercedes-Benz — AI_Andrew · 2026-11-13
- CoreWeave Q3 revenue hits $3.66B, up 168%; signs multi-year Wayve training deal — thursdai_pod · 2026-11-13
- NVIDIA's BirdDreams jointly generates perspective and BEV video for driving world models — nvidia · 2026-11-13
- Apparent OpenAI hardware unboxing photo leaks: robot arm plus handheld console — yungcontent · 2026-11-13
- FINGR dexterous hand solves a 2x2 Rubik's Cube with continuous finger tricks, wins Best Paper at CoRL 2026 workshop — Scobleizer · 2026-11-13
- Fei-Fei Li amplifies new Atlas demo: 'It is very hard to explain how cool Atlas is' — drfeifei · 2026-11-13