Welcome to check our collaboration with biotech/chemistry labs at Philadelphia convention center. We are here Oct 28-Oct 29 at FutureLabs Live
#robot#futurelabslive#physicalAI
As the only team running a live industrial humanoid demo at #RoboBusiness together with our partner @AGIBOT_US , we let the performance speak for itself. No simulations, no empty promises—just real-time, reliable control!
Humanoid demos are everywhere. Humanoids working real factory shifts? Much rarer.
Next week at #RoboBusiness2026 I'll share what it takes: data, sim, on-prem models & safety.
🗓️ Oct 20, 1:15 PM · Humanoids Track, Santa Clara Convention Center.
Congrats Brendan and other coauthors!
This paper explores one-shot multimodal learning with force-constrained Learning from Demonstration (LfD), enabling robots to learn not only how to reproduce demonstrated motions, but also how to proactively avoid excessive forces during execution. Looking forward to seeing the work presented at #IROS26!
How do we run RL with real-time chunking (RTC)? In this work we figured out how to use a small RL policy with a large robot foundation model, where the RL policy observed more recent images (due to faster inference) and steers the policy toward better behaviors!
A fun collaboration with Siemens, led by Brian Zhu, Momen Khalil, Emanuele Poggi from Siemens and @ehharrison4 from Berkeley, with lots of amazing contributors!
Great article but 1) batching / pooling would make the latency problem even worse; 2) Thor is a good choice, but there are other cheaper on-device solutions;
A Brain Too Big to Carry — On-Device vs Datacenter Inference https://t.co/20vgmcJ0l5
We all know Vision-Language-Action (VLA) models are game-changers for robotics. But trying to run a heavy 14B VLA model on edge hardware without lagging or melting the GPU? That's a whole different story.
Our team will be at IROS 2026 in Pittsburgh, September 28–30! If you’re working on Physical AI, robotics, or embodied intelligence, we’d love to connect and share thoughts.
I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.