3 days of caffeine, code, and chaos at CalHacks 12.0 🤖 Helped teams build everything from Spectacles-controlled humanoids to burger-making robots. Someone even rewrote the LeRobot SO-101 SDK from scratch. Thanks BitRobot Network for backing the robotics madness!

We trained a World Model for Lunar Landing using Thermal Computing! This was the result from a wild experiment with Francesco Sacco , David Dobáš, Théo at the Extropic hackathon!

We just released results for our newest VLA from Physical Intelligence: π*0.6. This one is trained with RL, and it makes it quite a bit better: often doubles throughput, enables real-world tasks like folding real laundry and making espresso drinks at the office.

How to train a humanoid: yesterday, I reproduced the full pipeline for Unitree G1, together with Emerson Segura 1) Created a motion in LaFan1 dataset format 2) Trained a BeyondMimic policy using mjlab 3) Tested in RoboJuDo simulation environment 4) Deployed on real G1 via

🥉 3rd Place: IRRS (Infinite Realistic Robotic Simulation) IRRS generates realistic 3D environments from prompts, lets robots interact within them, and automatically produces high-quality, annotated training data. David Dobáš Théo

A new robot is coming. Tomorrow, 9am. Innate

Introducing MARS, your first personal AI robot First units shipping this weekend innate.bot