Today we are introducing Dyna-2, a world-action model pre-trained on one million hours of human video. At this scale, for the first time, we discovered several new scaling laws: • world-action models exhibit scaling law on human data across four orders of magnitude, from 1000 to 1,000,000 hours, • this human data scaling law implied a scaling law on never seen robot data, • both data and objective matter; world modeling and scaling on video data are essential for cross-embodiment scaling transfer to emerge 🧵

Path has signed a $600M+ agreement with @wearehii (part of the $900M HII HYPR project). When I started Path eight years ago, I believed AI and robotics could give the world abundant skilled labor, helping people build the things that matter. Stronger infrastructure. A revitalized industrial base. The ships, factories, and critical systems our country depends on. Today is a major step toward that vision. In this agreement we'll deliver not only welding, but also dexterous assembly, enabling raw components to be turning into finished products. Unlocking throughput that was previously impossible, and unthinkable, without physical AI. Path was never just a welding company. The goal was always physical AI for all manufacturing and excited to take that next step at scale. Time to keep building.