Figure's robot demos criticized: 237/420 task success rate undercuts claims
ChongZzZhang · x · 2026-09-18
Citing tonyzzhao's critique, ChongZzZhang points out that Figure's widely praised robot demos actually achieved only a 56% success rate (237/420 tasks) — failing half the time is not "doing real useful work."
He adds two observations: the robot keeps repeating the same 3 tasks across different scenes, essentially data islands where human data may improve robustness only via strong inductive priors; and Figure's earlier impressive demos had roughly 10% success rates.
The takeaway: useful work requires generalization plus reliability, not edited highlight reels.
More from Embodied
- Mask R-CNN author Georgia Gkioxari argues closed-loop self-improving perception is AI's next leap — CSProfKGD · 2026-09-18
- World's first robot vs human fight set for tomorrow night in SF — cixliv · 2026-09-18
- UChicago neuroscientist decodes brain signals to restore touch and movement via brain-machine interfaces — plopesresearch · 2026-09-18
- Camera-free AI glasses are easy for Meta; shaking the 'pervert glasses' reputation is the hard part — Glittering_Royal6799 · 2026-09-18
- Figure's Home Robot Math: $400-600/Month Lease With a 56% Task Completion Rate — yongqianme · 2026-09-18
- Figure May Have Found a Scaling Law for Robot Pretraining, Predicting 8x-Data Runs — rohanpaul_ai · 2026-09-18