Robot Learns Hand Skills From 1M Hours of First-Person Video, Generalizes to Unseen Tasks
Scobleizer · x · 2026-10-14
Scoble highlights a robotics result: a model trained on roughly a million hours of first-person video showing people using their hands in everyday tasks. This lets it handle tasks it was never explicitly trained on — when facing a new object or a new room, it reuses moves it already knows to get the job done, demonstrating notable generalization from human video data.
More from Embodied
- Ukrainian FPV Drone Sim Leaves Early Access, Players Generate Training Data — generativist · 2026-10-14
- Menlo Research teases mystery robotics kit build in new photo — victormustar · 2026-10-14
- China's Hesai launches spatial camera to boost AI-powered robotics — pstAsiatech · 2026-10-14
- Nvidia's TacitVLA matches CoT performance at up to 3.4x lower latency — arankomatsuzaki · 2026-10-14
- $3,500 Flourish 1 Home Robot Uses a Raspberry Pi and Learns Chores From Your Phone — Matt Wolfe · 2026-10-14
- SlashVibe, a physical controller for AI agents, hits Kickstarter October 27 — sujingshen · 2026-10-14