Training Dyna-2 on 1M hours: Infrastructure challenges in robotics
josh_wills · x · 2026-08-18
DynaRobotics published details on training Dyna-2 on over 1,000,000 hours of egocentric video. They highlight that methods effective at smaller scales failed at this magnitude. Key challenges included capped ingestion throughput, 48-hour delays in building training manifests, and GPU latency from reading petabytes of cloud storage. The post emphasizes that infrastructure and data shoveling are the main bottlenecks in building robotics foundation models.
More from Embodied
- EmbodiedEvo tops RoboDojo as agents inherit reusable skills and lessons — Aiden_Tech_Ai · 2026-10-16
- Sunday Robotics is hiring a Senior Infra PM to scale consumer robot deployments — saranormous · 2026-10-16
- Bond's Submarine Lotus Esprit Spotted at Event, Fueling Steering-Wheel-Free Robotaxi Speculation — PTrubey · 2026-10-16
- Maker wires Grok and Claude into ROS2 robot with RealSense stereo cameras — chrismatthieu · 2026-10-16
- NEURA Robotics launches Neuron AI, a unified Perception-Language-Action model for robots — Distinct-Question-16 · 2026-10-16
- Dev demos open wearable AI passport running multilingual flashcards on tiny memory — ezshine · 2026-10-16