Fei-Fei Li Interview with Stephen Kotkin: Key Points

A summary of Fei-Fei Li's Hoover Institution interview with Stephen Kotkin: coming to the US at 15, physics at Princeton, ImageNet, World Labs and how America should handle AI.

▶ Watch the full interview on YouTubeHoover Institution · 21 October 2026

Fei-Fei Li, the co-founder and CEO of World Labs and a professor of computer science at Stanford University, spoke with the historian Stephen Kotkin for the Hoover Institution's series Only in America. The Hoover Institution published the episode on YouTube on 21 October 2026, and it runs 51 minutes^[1]. It had more than 3.1 million views by 10 November 2026. The conversation was recorded at World Labs' headquarters in San Francisco. Kotkin asked mostly about her path as an immigrant, her view of science and what America should do as AI changes the country. This page summarises what Li said, grouped by topic, with timestamps.

Li is also on our list of the most powerful AI influencers.

Topics and timestamps

The chapter names are shortened from the episode description^[1].

Arriving in New Jersey at 15

Li came to the US with her parents when she was 15 and settled in Parsippany, New Jersey, an hour from Princeton^[1]. She did not know the language and the family did not know anyone. She said her parents, like many immigrants, arrived with $20, so at first there seemed to be no opportunities.

More than 30 years later, she said she is grateful that America is "truly a land of opportunities". Her understanding of opportunity is that nothing is handed to you. To her, America is a place that lets you work hard, dream big or small, and take the opportunities you find.

Physics, Einstein and Princeton

Li said she applied to Princeton because of one person, Albert Einstein. Physics was her first love, and she still loves it although her work is now far from physics. As a teenager she and her parents worked many jobs, and she named cashier, cleaning and restaurant takeout. She also applied to Morris County College and to Rutgers, but she "just had to apply to Princeton" because Einstein had been there.

Asked what pushed her, she said her parents did not push her. They did not check her homework or ask about colleges, partly because they did not know how. She wrote about this time in her book The Worlds I See. In her view her parents gave her unconditional love to follow her dream. They never told her that, because the family was poor, she had to become a doctor or an accountant. Her father had an innate curiosity. The family had no money for paid entertainment, so he took her to free parks and to yard sales, which he treated almost like a scavenger hunt. She said she only understood later that this was a gift.

Kotkin said his own story is partly parallel. His family is a fourth-generation immigrant family, he was the first in it to go to college, and Princeton rejected him twice. Li recalled her first week at Princeton, when a physics professor told the class that Einstein used to sit in those seats. She said she felt like she "went to heaven".

Her audacious question

At Princeton, Li realised that what she loved most about physics was not quantum mechanics or special relativity. It was "the audacity to ask the hardest questions possible". She pointed to physicists who asked questions beyond the atomic world later in their careers. Schrödinger wrote What Is Life? and Roger Penrose wrote The Emperor's New Mind.

When Princeton asked her to choose topics for her junior paper and senior thesis, she decided her own audacious question was intelligence. She wanted to know whether people can build machines that are intelligent. She said she did not know then that this was Alan Turing's question, because she came from physics and had not read Turing yet. She went to Caltech for her PhD and worked on both computational neuroscience and artificial intelligence.

Kotkin joked that at Princeton he did not teach two things: creativity, which cannot be taught, and conformity, which students already knew. He said Li's path showed audacity, which he described as reaching for something that is not the obvious path.

Curiosity, courage and crossing boundaries

Li said her audacity comes from curiosity combined with courage and fearlessness, and she sees it as part of her immigration story. She credited her parents and the millions of immigrants who take a step into the unknown. As an immigrant teenager she learned that there was no path to follow, because she did not even understand the language. That left her with almost nothing to lose.

Kotkin said people often link America's striving culture to the frontier. His parents wanted him to become a doctor or a lawyer, and he joked that he disappointed them by becoming "only a history professor". Li added that freedom also gives people agency. As an immigrant child with total freedom who still had to survive, she felt she had to make it by herself.

She said being an immigrant and being a scientist have a lot in common, because both cross boundaries. An immigrant crosses language and cultural boundaries. A scientist crosses between the known and the unknown, and between fields. Human-centred AI, for example, crosses from technology to the social sciences and humanities.

A scientist's duty is to pursue truth

Asked about the moral and civic responsibilities of a scientist, Li said "the most important moral responsibility as a scientist is to pursue truth". She stressed that doing science does not mean scientists know everything or know the truth. Science pursues truth with rigorous methods and instruments, including experiments, debate and review. She said scientists have to be intellectually honest that their goal is to pursue truth. For her, the line between being a scientist at work and a person in life blurs, because she prefers the freedom to pursue and uphold truth.

Kotkin said people may disagree about what is true but usually agree on how to get there: evidence, logic and verification by others. Li said the pursuit of truth can be messy once those methods meet questions of values. Physics and computer science are mostly clean, deterministic and mathematical, but the human world is "way messier". For that reason she does not think scientists from technology fields hold the moral high ground on truth. She said she has a lot of respect for the humanities and social sciences, because some of what they study is hard to measure.

Kotkin then raised deepfakes. He said technologies that fake people's voices and behaviour put what he called the "truth regime" under siege. Li said humility about what scientists do not know matters here too. The technology her generation created now produces artifacts that challenge that truth regime. She said the same technology that simulates environments for robots and self-driving cars, or the molecular pathways of viral infections, can also be used for deepfakes. She said this is why she co-founded Stanford's Institute for Human-Centered AI. She wanted everyone to study these issues, use the tools in positive ways and look at how to avoid harm.

Why she co-founded World Labs

Kotkin said Li founded World Labs in 2024, and she added that she co-founded it with three other co-founders. She said the team shares an audacious technical goal: spatial intelligence is "the next chapter of AI's frontier". In 2024, about two years after ChatGPT, large language models were taking the world by storm. She said much of human intelligence and work is beyond language. She gave these examples:

• how people interact with the 3D world
• design, from homes and factories to cities and large machinery
• the ability to express and create
• simulation and training, for people from surgeons to athletes

She calls these abilities spatial intelligence. World Labs wants to be a frontier model company that builds the frontier model of spatial intelligence. She said the goal is to help creators, robotics, health care, education and enterprise users.

Li said she formed this thesis over years of AI, computer vision and robotics research at Stanford. She also said she loves building. She joked that her first company was a dry-cleaning shop. She said she built ImageNet, which "contributed to the birth of modern AI", and built Stanford's Human-Centered AI Institute so that faculty, staff and students could work on important problems. At this moment, she wanted to build a startup focused on spatial intelligence.

Kotkin described ImageNet as the step that made earlier neural network research by Geoffrey Hinton, Yann LeCun and others work in practice, because those networks needed large amounts of image data. Li gave the size of the dataset as 15 million images. When Kotkin wondered whether the project began at Caltech, Li said it began at Princeton, and Kotkin said she later brought it to Stanford.

Intelligence is more than language

Kotkin said he found the debate about whether language models are intelligent a little strange, and described 3D as a second leap after images. Li said she "firmly" believes intelligence is multi-dimensional. In evolution, animal intelligence started with sensing and perception, not with language. She said humans are, for most practical purposes, the only animal with language, and language began only a few hundred thousand years ago. Intelligence that started when animals first saw light or felt touch is a journey of half a billion years.

She said this journey began under the ocean, when trilobites first saw light. Seeing light made prey and predators the central force of evolution, and intelligence grew more complex in response to that survival pressure. In her account, intelligence involves seeing, touching, planning, talking and communicating. It also involves reasoning in a complex 3D world and imagining what will happen, both spatially and semantically. She concluded that fully realising the dream of AI needs a multi-dimensional approach. The episode description calls this area "world models"^[1].

Mentorship and civic responsibility

Kotkin asked how Li understands her platform and civic responsibility. She said she does not think of herself as a platform, because that is not her personality. She said one of the biggest privileges of her career is being an educator and a mentor as well as a scientist. Even as a founder and CEO, she feels she has a dual role as business leader and mentor, because everyone at the company is younger than her. She takes that as a responsibility.

She described herself as a citizen of America and of the world. In every role, including her mentoring, her teaching and the products World Labs builds, she wants to have a positive impact on human flourishing. She said this is her responsibility and her privilege at the same time.

America and the AI change

Asked what America could do better, Li said "America to me is change". She said the country was founded on the desire to change, and pointed to the story of the Constitution. She wants the country to recognise and embrace the change now coming from cultural, demographic and technological shifts.

She said "nothing pains me more" than seeing K-12 teachers, or even K-16 teachers, left behind by this change. The conversation about this technological change has not reached the country's teachers and parents. She thinks the change should come from a healthy ecosystem of talent creation, public discourse and incentives. As co-director of Stanford's Institute for Human-Centered AI, she often goes to Washington, D.C. to argue for resources for the public sector and academia. She said that is where public goods are produced, including talent, scientific discoveries such as cancer research, and models of how the economy and jobs will change. She worries that society is not responding to this change in the best way.

Kotkin said the US has about 48 million foreign-born residents, roughly a fifth of the population, and that the country has had to adapt from the start. He asked whether Li meant technological or cultural adaptation. She said she meant adaptation at all levels, because they are connected. The scientific revolution 400 to 500 years ago led to the economic, cultural and political change now called the Enlightenment. She believes the current change is bigger, because this technology reaches the layer of human cognition. It changes work, learning and communication, and cultural, political and economic change follows from that. She said improved productivity "absolutely does not guarantee shared prosperity", and that countries have to recognise this and bring people along.

Near the end, Kotkin said the future is still to be invented. Li agreed that this is what America is about: opportunity, the future and building towards it.

Sources

1. Fei-Fei Li on Curiosity, Audacity, and the Quest for Artificial Intelligence — Hoover Institution, YouTube (21 October 2026; views checked on 10 November 2026)