


Shuo Yang
56 posts

@ShuoYangAIR
CTO & Co-founder @ Mondo Robotics/ Ex Tesla | CMU PhD | Ex DJI




Mondo Robotics should seriously sell these - like now!





We’re excited to share DiT4DiT, an end-to-end Video-Action Model for robot learning that unifies a video Diffusion Transformer and an action Diffusion Transformer in a single cascaded framework. By leveraging the rich spatiotemporal and physical dynamics learned through video generation, rather than static image-text priors, DiT4DiT achieves state-of-the-art results on LIBERO (98.6%) and RoboCasa GR1 (50.8%) with far less training data, delivering over 10× better sample efficiency and up to 7× faster convergence. Real-world deployment on a humanoid robot further shows robust generalization. We believe this is a step toward making video generation a powerful backbone for robot policy learning. This work builds upon the brilliant foundations laid by Nvidia's GR00T and Cosmos. Project: dit4dit.github.io Paper: arxiv.org/abs/2603.10448 Code: Coming soon. In the meantime, you can ask your coding agent to reproduce the method based on GR00T/Cosmos.


We’re excited to share DiT4DiT, an end-to-end Video-Action Model for robot learning that unifies a video Diffusion Transformer and an action Diffusion Transformer in a single cascaded framework. By leveraging the rich spatiotemporal and physical dynamics learned through video generation, rather than static image-text priors, DiT4DiT achieves state-of-the-art results on LIBERO (98.6%) and RoboCasa GR1 (50.8%) with far less training data, delivering over 10× better sample efficiency and up to 7× faster convergence. Real-world deployment on a humanoid robot further shows robust generalization. We believe this is a step toward making video generation a powerful backbone for robot policy learning. This work builds upon the brilliant foundations laid by Nvidia's GR00T and Cosmos. Project: dit4dit.github.io Paper: arxiv.org/abs/2603.10448 Code: Coming soon. In the meantime, you can ask your coding agent to reproduce the method based on GR00T/Cosmos.













20 months since... Can we come up with a new handshake? @tonyzzhao x.com/chichengcc/sta…


This is where you'll shine Use custom prompts in Grok to direct your videos more precisely Try it and show me what you've got:

Been seriously thinking about building a vetted execution network across Asia, contract manufacturers, DFM, prototyping, QA, automation, and more, purpose-built for early-stage robotics startups in the Bay. Huge unlock if done right.

The number of startups trying to sell data to general purpose + humanoid robots companies seems to exceeds the number of actual robot startups. This is ... concerning. Do the hard thing. Hard things filter out competition. Optimize for the world you want to see, not your IRR

And that’s a wrap! My last day at @maticrobots came a little sooner than expected, as our family gets ready for a big move abroad for my husband's robotics startup adventure. What a ride it's been. Joining @maticrobots meant diving headfirst into a challenge that felt both familiar and impossible, a bit of déjà vu from my early days at @DJIGlobal. We weren't just launching a product; we were building trust and delight in a skeptical market, and creating desire for something truly new. We did it. With state-of-the-art tech, a deep belief in the mission, and a surprising number of googly eyes 👀, stickers, and LEGO kits, we brought Matic to life. In just over a year and with a lean budget, we shipped thousands of robots, organically grew a community to over 30,000 strong, and earned a 10/10 from WIRED. Seeing some of the most respected minds in consumer tech and robotics validate not just what we built, but why and how we built it, has been incredible. But for me, the real magic wasn’t in the accolades. It was in the stories of toddlers naming it their favorite robot friend, the photos of pets curled up next to Matic, and the quiet satisfaction of knowing thousands of homes feel calmer and cleaner because of what we created. I’m incredibly proud to have helped transform complex, cutting-edge robotics into something warm, useful, and loved. The work we’ve done is just the beginning, and I can’t wait to see where the team takes it next. I’ll be around the Bay Area for another month and would love to catch up. My calendar’s open for good coffee and good conversation. As for what’s next for me... more on that soon!