
@JonMSchwartz Agreed, obvious in hindsight. Jim Fan called it the integral of data types at Actuate and I thought that was a good way to frame it
English
Caleb Appleton
164 posts

@appleton_caleb
Cyclist, coffee roaster, recovering triathlete and engineer. Deep tech investor at Bison Ventures. Always learning. Views are my own.


We discovered an emergent property of VLAs like π0/π0.5/π0.6: as we scale up pre-training, the model learns to align human videos and robot data! This gives us a simple way to leverage human videos. Once π0.5 knows how to control robots, it can naturally learn from human video.






















