IROS Debate: Generalist Robots Nail Coarse Placement 19/20 but Fail Precision Insertion 2/20
GeorgiaChal · x · 2026-10-01
At IROS 2026, Ken Goldberg and Andrea Bajcsy moderated a generalist-vs-specialist robotics debate, opened by Matt Mason's provocation: "Should animals be humans?"
Key takeaways:
- Harry Xu (pro-generalist): "Don't build Excel. Build GPT" — since tomorrow's needs are unknown, buy generality and amortize it; notably GPT didn't replace Excel, it learned to drive it.
- Georgia Chalvatzaki decomposed "general" into 4 axes — body, scene, contact, task — arguing breadth and reliability must be measured per axis and one doesn't imply the other; "the next task is the test."
- The most honest number of the day: a general model calling robot tools scored 19/20 on coarse block placement but only 2/20 on precision insertion — coarse pick-and-place works, physical composition and precise contact don't yet.
More from Embodied
- Unitree T-series humanoid shows off impressive kip-up move — jonstephens85 · 2026-10-01
- Dyna-2.1 Breakdown: How the Team Built the First Physical Agent With 1-Hour Whole-Body Autonomy — chris_j_paxton · 2026-10-01
- Why RL for Low-Level Robot Control: Dyna-2.1 Author's Three Key Reasons — chris_j_paxton · 2026-10-01
- InSpatio-World 1.5 turns images and videos into real-time 4D worlds, tops WorldScore — Scobleizer · 2026-10-01
- MiTaS: Multi-Resolution Tactile Fusion Hits 80% Success on Contact-Rich Robot Tasks — GeorgiaChal · 2026-10-01
- NVIDIA Researcher Explains Cosmos: A World Foundation Model Is a Foundation Model of World Models — liu_mingyu · 2026-10-01