Fei-Fei Li: Spatial Intelligence is the Next Frontier in AI
Key ideas
Li traces her career from ImageNet to World Labs, describes choosing near-delusional problems and feeling at home at ground zero, and offers founders her hiring criterion: intellectual fearlessness.
From a machine transcript of the audio. Quotes link to the moment they are said.
- 1
Bet on the ingredient everyone is missing
Li says generalization kept haunting her machine learning work, and that around 2007 she and a student bet that a data-driven paradigm shift was needed. They open-sourced ImageNet and ran an annual challenge so researchers everywhere would work on it. She recalls three years with very little signal before a 2012 entry brought data, GPUs and neural networks together.
- 2
Look to evolution for the next problem
Li says that when she searches for her next North Star problem, she asks what evolution or brain development has done. She contrasts language, which she puts at well under a million years of evolution, with the 540 million years she gives for vision and understanding the 3D world. From this she concludes that spatial intelligence is a fundamental problem of AI, and she founded World Labs to work on it.
- 3
Treat difficulty as the reason to start
Li lists why world models are hard: the world is 3D, sensing collapses it to 2D, models must move between generating and reconstructing, and spatial data is not as accessible as language. Rather than treating this as a deterrent, she says it excites her, because easy problems would already be solved. She declines to give World Labs specifics, saying only that it takes a hybrid approach to data and cares about quality.
- 4
Hire for intellectual fearlessness
Asked what her best-known students had in common, Li says they differ widely but share a courage to embrace something hard and go all in. She says background and problem matter less than that trait. She tells founders it is also her own hiring criterion as CEO of World Labs.
- 5
Make ground zero your comfort zone
Li describes starting a dry cleaning shop at 19 to support her family, then a couple of times joining departments where she was the first computer vision professor, against advice to seek senior mentors. She says she ran Stanford's Human-Centered AI Institute like a startup for five years. She admits she still has days of not knowing what she is doing, and advises simply focusing on doing it.
In Fei-Fei’s words
“frankly it excites me because if it's easy somebody else has solved it and my entire career is going after problems that are just so hard, bordering delusional”
Fei-Fei Li·Fei-Fei Li: Spatial Intelligence is the Next Frontier in AI·Watch at 21:23(opens the original in a new tab)“that courage, that fearlessness of embracing something hard and go about it and be all in and trying to solve that in however way you want is really a core characteristic of people who succeed.”
Fei-Fei Li·Fei-Fei Li: Spatial Intelligence is the Next Frontier in AI·Watch at 31:07(opens the original in a new tab)“I love the feeling of ground zero, like standing on ground zero. Forget about what you have done in the past. Forget about what others think of you. Just hunker down and build.”
Fei-Fei Li·Fei-Fei Li: Spatial Intelligence is the Next Frontier in AI·Watch at 29:08(opens the original in a new tab)About this entry
- Published
- Jul 1, 2025
- Channel
- Y Combinator
- Type
- Video
- Length
- 44:22
- Language
- English
- Added
- Nov 2, 2026
- Source last checked
- Nov 2, 2026
What’s available here
Original source
Linked. Opens on YouTube.
Player on this page
Available. YouTube’s own player, loaded only when you press play.
Full text or transcript
Not reviewed. No license to show a text version here.
Our summary
Draft. Written by us with AI from a machine transcript of the video’s audio (Whisper) and checked against it. Machine transcripts can mishear; each quote links to the moment so you can hear it. Not yet rights-reviewed.
Ask questions with AI
Not reviewed. Off until there is licensed text to work from.