Embodied AI Faces Two-Order-of-Magnitude Data Gap as Industry Heads to 10M Hours
GaryMarcus · x · 2026-11-11
Citing an interview with the Memories AI team, CyberRobooo notes Physical AI data collection could reach the 10-million-hour range by next year — up from 1M hours discussed months ago and 1–2M hours recently.
Even that may not be enough. At the World Internet Conference in Wuzhen, Li Shuai of China MIIT's Fifth Electronics Research Institute quantified the gap: embodied AI's core data supply-demand shortfall is roughly two orders of magnitude.
Memories AI's approach turns first-person human work video into searchable Physical AI training data with per-hand action tracking, aiming to improve humanoid dexterity via curated high-quality data.
More from Embodied
- Flashing cheap ESP32 IoT devices with ESPHome for a smart-home upgrade — haydendevs · 2026-11-11
- MessyMem on Compact Long-Term Memory for Agentic Robots Hits CoRL 2026 Main Poster — leto__jean · 2026-11-11
- NYU's Steertuning Lets a Robot Learn a Chore From 5 Minutes of Human Demo — CyberRobooo · 2026-11-11
- Boston Dynamics Atlas demoed with Blender 5.3 — TomLikesRobots · 2026-11-11
- Lamport Robotics posts robot 'wake-up test' demo video — R4nsack · 2026-11-11
- NVIDIA hosts CoRL 2026 meetup on scaling robot training with Isaac Lab — m_wulfmeier · 2026-11-11