Chet Bhateja (@ChetBhateja) header
Chet Bhateja (@ChetBhateja) profile picture

Research engineer @DynaRobotics, prev @1x_tech

Joined July 2019
Photo by Chet Bhateja (@ChetBhateja)Photo by Chet Bhateja (@ChetBhateja)
Our laundry demo is the most ambitious workflow I've ever seen a robot perform, requiring innovations across hardware, controls, modeling, and orchestration. 😮 Incredibly proud of the entire team for coming together to deliver Dyna-2.1!
We are releasing Dyna-2.1, the first Physical Agent that achieves reliable super long-horizon whole-body autonomy. It combines our brand-new semi-humanoid hardware with an agentic system built around Dyna-2 to handle ultra-long real-world workflows. Here is an uncut footage of Dyna-2.1 completing an entire hour-long laundry room workflow, just like a human does.
3
3
27
916
Dyna is building across the full robotics stack. Stay tuned!
Made for the details and the demands of real work. Built for craft and mastery. A first look at our new robot. Coming soon.
1
27
1,857
Excited to share Dyna-2! Enjoyed collaborating with the team to push the boundaries of cross-embodiment and instruction following. We're just getting started!
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 🧵
4
3
39
3,036
Chet Bhateja retweeted
As imitation learning policies continue to scale, deciding how to weigh different robot datasets will become even more difficult. To address this problem we introduce ReMix, a method for automatically curating large RT-X scale imitation learning datasets. 🧵(1/5)
3
36
4
190
39,647