Dyna-2 Shows Scaling Laws in World-Action Models Trained on 1M Hours of Video
DJiafei · x · 2026-08-15
Dyna Robotics published a research blog on Dyna-2, a World-Action Model (WAM) pre-trained on over one million hours of human video data. Key findings include that the model exhibits clear scaling laws on held-out human data. Crucially, it demonstrates the emergence of a human-to-robot transfer scaling law on unseen robot data, proving that scaling human video pre-training directly improves zero-shot robot performance.
Related event: Dyna-2 Model Pretrained on 1M Hours Video(2 posts)→
More from Embodied
- VLA-Δ: Zhejiang University Adds Physical State Transition Supervision to VLA Models — OmniAI-ZJU · 2026-10-16
- PINT: Shared Predictive Intent Tokens Unify Reasoning, World Modeling, and Robot Actions — Xinyu Guo · 2026-10-16
- TouchScale releases 500 hours of unified egocentric vision-tactile data for scaling study — jiqizhixin · 2026-10-16
- Ghost raises $11M for Core, a personal AI computer that runs locally in your home — nick_linck · 2026-10-16
- Robot shows up at Costco to fetch the famous $1.50 hotdog — majidmanzarpour · 2026-10-16
- 15 Leju Kuavo humanoid robots run 2 months at Hangzhou auto-parts line, 98% success rate — CyberRobooo · 2026-10-16