
What Two Physical AI Labs Taught Me (For Anyone Building Theirs)
·9 min read
I spent a lot of the last two years working on robot foundation models, first as an early member of the GR00T team at NVIDIA (before N1 shipped) and now at Dream Labs, where I lead Dream API, the
I spent a lot of the last two years working on robot foundation models, first as an early member of the GR00T team at NVIDIA (before N1 shipped) and now at Dream Labs, where I lead Dream API, the first public API for a robot foundation model. A surprising share of the results I had to explain away had nothing to do with the model. This post is about that: the part of robot learning that people tend to underrate. In robot learning, the model is a small part of what determines policy performance.
110
0
17
0
