Sergey Levine on Physical Intelligence: scaling robot foundation models from Pi Zero to Pi 0.7
adityaag · x · 2026-10-23
A wide-ranging conversation with Physical Intelligence co-founder Sergey Levine. His founding thesis: 'The universe is really big, and almost none of it is inside computers' — so his team filled a repurposed Google conference room with robots grasping objects all day. It worked until they hit a wall: lots of experience, no prior knowledge — and that wall became Physical Intelligence.
Key segments:
- From video games to robotics, and why the PI founding timing and team mattered
- Pi Zero: building a foundation model for robots
- From Pi 0.5 to Pi 0.7: teaching robots to reason and generalize
- What makes a great robotics researcher
Related event: Sergey Levine on Building Robot Foundation Models at Physical Intelligence(2 posts)→
More from AGI Musings
- gfodor defends dad teaching kid to build with AI: critics have 'Stockholm Syndrome' — gandamu_ml · 2026-10-23
- AI safety researcher: labs aren't building anything capable of killing everyone — nitarshan · 2026-10-23
- "Front-end is sacrificial": the case for headless, agent-built UIs — genmon · 2026-10-23
- Security lead warns AI vulnerability tsunami is months away; basics buy only ~18 months — EthanJPerez · 2026-10-23
- Frontier models are accelerating research, letting every lab ship more models with the same compute — almmaasoglu · 2026-10-23
- Karina Nguyen: Anthropic warned of AI cyberattacks and sycophancy years before they were legible — karinanguyen · 2026-10-23