Subvolts take Robots learn like LLMs: diverse data, not perfect machines, drives general purpose AI. Apr 12, 2026 - Developers
- Learn how diverse data training enables adaptable robots, informing your approach to embodied AI development.
- Business
- Understand the shift towards general-purpose robots and how foundation models could unlock new applications and markets.
This interview provides a technical look at how foundation models could unlock general-purpose robots, a significant area of AI research.
#331 Sergey Levine: The Robot Revolution Nobody Is Talking About
Audio streams from Craig S. Smith. Your place is remembered on this device.
From Subvolts
Summary and analysis by Subvolts editorial, AI-assisted.
Craig Smith interviews Sergey Levine on "Eye on AI" about building foundation models for robots. Levine explains that training AI on diverse data from many robots and humans, rather than perfecting a single machine, enables robots to learn and adapt to the real world. The discussion also covers the debate around humanoid robots and the impact of foundation models on their design.
Imagine teaching a robot to do anything, not just one task. This episode explains how feeding AI lots of examples from many different robots and even people helps it learn to handle the messy real world. This approach could lead to more useful robots everywhere.
This episode offers a technical perspective on the future of robotics, moving beyond the hype. Levine, a leading researcher, explains how foundation models, similar to those powering large language models, can unlock general-purpose robots. This is relevant to frontier labs like Google Gemini and OpenAI as they explore embodied AI and how models can interact with the physical world. It's for anyone interested in the practical challenges and potential of AI in robotics.
- HostCraig S. Smith
- GuestSergey LevineProfessor, UC Berkeley
- Episode Title: #331 Sergey Levine: The Robot Revolution Nobody Is Talking About
- Show: Eye on AI
- Guest: Sergey Levine
- Guest Affiliation: Co-founder of Physical Intelligence and professor at UC Berkeley
- Key Concept: Training AI on broad, diverse data enables robots to learn and adapt.
This episode is sponsored by Modulate. Most voice AI focuses on transcription. Velma takes it further by actually understanding conversations, analyzing tone, timing, stress, and intent using its Ensemble Listening Model architecture. Explore the live preview: https://preview.modulate.ai/ What does it actually mean to build a foundation model for robots? In this episode of Eye on AI, Craig Smith sits down with Sergey Levine, co-founder of Physical Intelligence and professor at UC Berkeley, to explore a fundamentally different approach to building robots, one inspired not by programming a single perfect machine, but by training AI on the broadest and most diverse data possible so robots can learn, adapt, and operate in the unpredictable real world.
Sergey explains why the secret to general-purpose robots isn't perfecting one single machine, but training on massive, diverse data from all kinds of robots and even humans. The more variety the model sees, the better it gets. Just like ChatGPT learned from all the text on the internet, robotic foundation models learn from every robot that has ever moved, grabbed, or interacted with the real world.
We also get into the big humanoid robot debate. Are they the future, or is it mostly hype? Sergey gives an honest and technical take on why the form factor conversation is changing now that foundation models exist, and why that actually opens the door for more creativity, not less.
Finally, Sergey shares what he's most excited about next, building a true data flywheel where robots get smarter the more they are deployed, creating a continuous learning cycle that could change everything.
Subscribe for more conversations with the people building the future of AI and emerging technology.
Stay Updated:
Craig Smith on X: https://x.com/craigss
Eye on A.I. on X: https://x.com/EyeOn_AI
(00:00) Introduction: What Are Foundation Models for Robots?
(01:44) Meet Sergey Levine: Physical Intelligence and UC Berkeley
(02:51) Breaking Down Foundation Models for Non-Technical People
(06:46) Why Real World Data Beats Simulation
(15:00) Building a Broad Robotics Foundation From Scratch
(24:00) The Open World Problem in Robotics
(40:00) Generalist vs Specialist Robots: Which Wins?
(47:00) Humanoid Robots: Real Innovation or Just Hype?
(55:10) The Future: Continuous Learning and the Data Flywheel
(56:23) Guilty Pleasure: Sci Fi and Thinking Beyond the Limits
- foundation model
- Sergey Levine
- A large AI model trained on vast amounts of data that can be adapted for many different tasks.
55 min42 min51 min34 minUkraine's Secret Weapon: The Points System Winning the Drone War | Andrii Hrytseniuk
WithAndrii Hrytseniuk
2 h 12Sam Altman’s Attack, Amazon vs. Starlink, and What Opus 4.7 Actually Means | #248
WithSalim Ismail, Dave Blundin, Dr. Alexander Wissner-Gross
58 min
44 minBuilding AI for Life Sciences - Episode 16
WithJoy Jiao, Yunyun Wang
1 h 02Questions people ask
- What is the new approach to building robots?
- Instead of programming one perfect machine, AI is trained on the broadest and most diverse data possible from many robots and humans to learn and adapt.
- Are humanoid robots the future?
- Foundation models are changing the conversation around humanoid robots, potentially opening doors for more creativity in their design and application.
Page generated Oct 19, 2026. Summaries are Subvolts' own; the audio, artwork and notes belong to Craig S. Smith.