Anyone skeptical about AI’s ability to perform diverse physical tasks should watch this. Physical Intelligence's π₀ model in action: 18 minutes of bimanual robots autonomously handling complex, dexterous chores.

Finally an AI that does laundry!🤯 Demo by Physical Intelligence Would you buy a robot like this?

This is why SpaceX hasn’t developed weapons systems

Excited for the team to share FAST - our new robot action tokenizer - pulling off 5x improvements in VLA training speed. Congrats Karl Pertsch and team Physical Intelligence

Traditional tokenization methods for robotic actions struggle with high-frequency, dexterous tasks due to redundancy and inefficiency. Inspired by JPEG compression, Physical Intelligence has developed a compressed action representation that accelerates VLA model training 5x.

Disappointed with your ICLR paper being rejected? Ten years ago today, Sergey and I finished training some of the first end-to-end neutral nets for robot control 🤖 We submitted the paper to RSS on January 23, 2015. It was rejected for being "incremental" and "unlikely to have

Introducing "Hi Robot": we developed a way to let robots 'think through' complex tasks, breaking them into basic steps, modifying tasks on the fly, and adapting to user feedback. A major step in robot intelligence. Great work Lucy Shi and team at Physical Intelligence! Learn more

Happy π day!

Don't get too excited about videos of humanoid robots dancing. Manipulation is the hard problem we need to solve to make humanoid robots useful, not locomotion. One video of a humanoid slowly performing tasks with its hands is worth a dozen clips of choreographed routines.

We asked π₀.₅ — trained mostly on data from other robots and the web — to “load the dishes” and “make the bed” in homes it had never seen. It parsed the instructions, planned, and executed end‑to‑end. This isn’t just a research milestone; it’s a real step toward

The Amusement Park for Engineers An exclusive peek behind the curtain of Anduril's product engineering machine Featuring first-ever photos from inside Anduril Industries's R&D facilities. By former SVP of Engineering and Physical Intelligence co-founder Adnan Esmail Only in Colossus Review

Great look behind the curtain at Anduril Industries from Adnan Esmail on Anduril’s product engineering machine. joincolossus.com/article/the-am…

Anduril has become one of the most interesting company stories of this generation Adnan Esmail, who ran product engineering at Anduril, brought us inside his experience there—writing about how they build new things And, allowing us to snap photos behind the scenes never seen

Proud and honored to share our incredible journey Anduril Industries

We’re releasing the RoboArena today!🤖🦾 Fair & scalable evaluation is a major bottleneck for research on generalist policies. We’re hoping that RoboArena can help! We provide data, model code & sim evals for debugging! Submit your policies today and join the leaderboard! :) 🧵

Evaluating robotic foundation models is really hard — everyone has different robots, tasks, etc. We are releasing RoboArena as a step toward a global network of decentralized evaluations, where policies can compete head to head on evals in the real world at many institutions!