36B Open-Source Robot Model Isaac 0.5 Runs Real-Time Under 50ms Without Local GPU
AkshatS07 · x · 2026-10-03
- Perceptron's Isaac 0.5 is one of the largest open-source robotics foundation models at 36B total parameters; the poster reports running it real-time with <50ms overhead without a local GPU.
- It offers evidence for scaling laws in physical AI: more egocentric video pre-training data reduces the amount of teleoperation data needed to reach the same loss. Frontier robotics labs like DynaRobotics and GeneralistAI are likewise pushing pre-training data from thousands to millions of hours.
- Dreamscale Labs (YC F26) argues data and compute don't scale in isolation—model size must grow too, pointing toward trillion-parameter robotics models.
More from Research
- Light-powered AI detects deepfakes with nearly 98% accuracy — ai-edition · 2026-10-03
- Animation Bench draws buzz: researchers from six labs say the animation capability gap was long overlooked — himanshustwts · 2026-10-03
- Chalmers builds closed-loop AI scientist that proposes and runs its own biology experiments — Brighter-Side-News · 2026-10-03
- Creative chess puzzle generation with diffusion models: new RL recipe, open weights — TZahavy · 2026-10-03
- Discrete diffusion delivers provably lossless LLM inference speedups, drop-in for training — Cohere · 2026-10-03
- When hand tracking misses the plug-in moment: how should robot imitation data be evaluated? — Klutzy_Cap8492 · 2026-10-03