Building the "App Store" for Robots: Hugging Face's Thomas Wolf on Physical AI

Robotics is currently experiencing a transformative shift analogous to the rise of language models, where software adaptability—not hardware—serves as the primary constraint. Hugging Face’s LeRobot project addresses this by providing an open-source framework that integrates policy models, datasets, and physical hardware to democratize robotics development. By fostering a community-driven ecosystem, this initiative enables developers to move beyond hobbyist setups toward mass-market applications. While data scarcity remains a significant hurdle, advancements in world models and simulation offer promising pathways for generating diverse training data. Rather than prioritizing expensive humanoid form factors, the focus lies on creating a diverse, accessible "galaxy" of robots that can be customized and deployed locally. This approach emphasizes safety, data privacy, and the potential for a new era of physical AI where individuals can actively build and innovate rather than merely consume technology.