Fei-Fei Li is a pioneering AI scientist breaking new ground in computer vision, a Stanford professor, and currently leading the innovative start-up World Labs. While her career is deeply rooted in technical expertise, Dr. Li's journey is driven by an insatiable curiosity. In this episode, Brad and Dr. Li reflect on poignant moments from her memoir, "The Worlds I See: Curiosity, Exploration, and Discovery at the Dawn of AI," highlighting the crucial role of keeping humanity at the center of AI development. They also explore how government-funded academic research, driven by curiosity rather than profits, can lead to unexpected and profound discoveries that propel innovation and economic opportunities.
Fei-Fei Li is a pioneering AI scientist breaking new ground in computer vision, a Stanford professor, and currently leading the innovative start-up World Labs. While her career is deeply rooted in technical expertise, Dr. Li's journey is driven by an insatiable curiosity. In this episode, Brad and Dr. Li reflect on poignant moments from her memoir, "The Worlds I See: Curiosity, Exploration, and Discovery at the Dawn of AI," highlighting the crucial role of keeping humanity at the center of AI development. They also explore how government-funded academic research, driven by curiosity rather than profits, can lead to unexpected and profound discoveries that propel innovation and economic opportunities.
Click here for the episode transcript.
Brad Smith: I am Brad Smith, and this is Tools and Weapons. On this podcast, I'm sharing conversations with leaders who are at the intersection of the promise and the peril of the digital age. We'll explore technology's role in the world as we look for new solutions for society's biggest challenges.
Fei-Fei Li: Curiosity is absolutely the through line of my life as a scientist, and I think it's really frankly, one of the gift of life. Every child is born with curiosity. I think it's in our DNA. It's who we are as a species.
Brad Smith: That's Dr. Fei-Fei Li, a pioneering AI scientist, entrepreneur, and a Stanford professor who is leading groundbreaking work on computer vision. On this episode, Dr. Li shares the importance of keeping human values at the center of AI development. We also explore how government-funded academic research driven by curiosity rather than profit leads to unexpected but profound discoveries that have long fueled American innovation. Fei-Fei Li, up next on Tools and Weapons.
Well, let me first say, welcome to what I think is the opportunity to talk with one of the world's greatest computer scientists, AI experts, now an entrepreneur, a person who's connected with technology and its societal impact in the United States around the world, Dr. Fei-Fei from Stanford. Fei-Fei, welcome. Thank you for joining me today.
Fei-Fei Li: Thank you, Brad. I'm so excited to join you and just have a great chat.
Brad Smith: Well, you have such an extraordinary personal story. You've written a terrific book about it, and there's one word that I would like to talk about with you today. The word is curiosity, the role it has played in your life and your success. I think the role it can play for everyone who's listening and the future of technology. You, I think have almost always been a curious person. Where did that interest, that spark of curiosity first enter your life?
Fei-Fei Li: Well, first of all, Brad, I'm so glad you picked that word. Really, the title of my book is The Worlds I See. But the first word in the subtitle is curiosity because I think that is absolutely the through line of my life as a scientist, and I think it's really frankly one of the gift of life. So you asked me where did I come from? I actually want to say that I was lucky curiosity never escaped me. I think every child is born with curiosity. I think it's in our DNA. It's who we are as a species. Frankly, scientists have studied that even primate species and other animal species have curiosity. So this is a huge gift that humanity has for itself.
And what is really a gift was that my early years of development or growing up one way or the other, and a lot of credit to my parents, they protected that. So despite whatever ups and downs of life and the changes of circumstances, I was lucky since I was a little girl, that curiosity was protected. It grew. It became a fire that continued to drive my life. You mentioned that I'm now an entrepreneur. It's a weird choice in this stage of my life at least. A lot of people would ask me, "Why are you doing this?" And the fundamental answer to that question is curiosity. I'm just so excited to be doing what I'm doing.
Brad Smith: I would love for you to share a couple of stories about how this came together for you, because I think they're very relatable and they're very interesting for people, and you tell these stories in your book. But for someone, and we'll get into this, who's been a leader in computer vision and images, pattern matching and the like, when you were growing up, you accompanied your father, I think to quite a number of yard sales. Tell us about that.
Fei-Fei Li: Yeah. So you're referring to our early immigration days. When I was 15, my mom and I joined my dad in New Jersey of all places, per se, pretty in New Jersey. And being an immigrant or immigrant family is really tough. We literally went from a middle-class family in China to really impoverished in New Jersey, and one of the way we survive is just buying items from yard sales and garage sales, that they are cheaper and so on. But I think the thing, and I think I wrote it multiple times in the book, in different light, the things that you're describing is on the weekends I would go with my dad to yard sales, but now that we're talking, it sounds like such a boring or not even great thing to do. But you have to transport yourself to my perspective as this teenager, going to these yard sales with her dad who didn't speak English, but the way he approached this tiny event in life on a weekly basis is so much curiosity.
He would approach every little stand or table of whatever items with that pure child like, "Oh my God, Fei-Fei, come and look at this cup. It has a owl on it, or look at this garden tool I never saw. I've never thought of, you can design it this way so that it's easy to uproot a plant." He would just see everything with that beginner's mind, child like delight. And that was just something that, it didn't occur to me then, but now that I've been a career scientist for decades, I realized that delight in seeing things for the first time and keeping that curiosity is the same as what we do as scientists. The first time I open a paper, look at the results someone gives us or heard about an algorithm, you have to have that delight and curiosity.
Brad Smith: One of the wonderful lessons I think from that experience that I just feel is of universal applicability is that almost every moment in life, even the seemingly most mundane, often creates an opportunity to learn something new. But you have to look, you have to listen, you ask questions, you retain that sense of wonder.
Fei-Fei Li: Yes, the sense of wonder. Yes.
Brad Smith: And it's a wonderful thing, that sense of wonder, and obviously you built on that. You got this scholarship to go to college in New Jersey at Princeton. You were obviously not just curious, you were a hard worker. And one of the stories you also share is how you poured yourself into your classes, your academic work, even your tests. There was the day you took an exam, I think really outside a hospital room where your mother was. Can you tell us a little bit about that, and how that fit into your development as a person?
Fei-Fei Li: Well, thank you, Brad, for reading the book with so much details and also university I went to, you know a thing or two about that.
Brad Smith: True.
Fei-Fei Li: It's our shared alma mater. So first of all, this book is written as a scientist coming of age and also a discipline, a scientific discipline, AI coming of age. It's not meant for a sob story, and that is not how I approach any of these story. The story, the thread that you're referring to right now to me is a love story for immigration, is that millions and millions of American immigrants worked so hard with a sense of hope and dream, and I was just one of them. And it's true that it was difficult years.
My mom didn't have good health and especially because she is a cardio patient, so her heart is very ill. And that thing you talked about was during freshman year in Princeton, I had to take my mom for a quasi-emergency surgery in a hospital, and I had to translate for her because it was a surgery that part of her remains awake and has to respond to the doctor. And I was the only translator and I was also taking exams. So I got Princeton to allow me to do that literally in actually in squats because I have to be in the area where she was and take my exam while being there to support her and support the doctors.
But overall, it was a tale of two towns, the tale of two cities. One is where I study at Princeton and physics, and I loved every minute of it. It was painful, it was hard, but it was fun. But it was also the story of Persephone where my family lived and we had a dry cleaning shop that in Silicon Valley language, I was the CEO of the dry cleaner, and I employed my parents. And it's a weakened business as most dry cleaning business are, and I was working hard there as well.
Brad Smith: Well, what I find interesting is those two towns came together for you. Your journey progressed. It took you from Princeton on the East Coast, Caltech on the West Coast, and I think there was a point where you had a fork in the road. As you were at Caltech, you had the opportunity to decide what to do next. And one path led to McKinsey, obviously great management consulting firm. They offered you a very high salary, and the other was to continue with a quest for science, which certainly at the time would not have appeared to have offered as much money. And you had a conversation with your mother about what you should do. What did your mother tell you?
Fei-Fei Li: Yeah. I think a lot of students would experience that, but my situation probably was exacerbated by the family financial situation. Is you're in the middle of grad school, the second third year of PhD is the hardest. You're in the PhD blue and papers get rejected and you don't know if you're going to get a good job. Especially at that time, AI as a field was still in the AI winter. So it's unlike today's AI generation. If you're an AI student, you'll have so many jobs waiting for you. At that time, that was not true at all. If you AI student, the industry was not quite there yet. And faculty job is so competitive and I was just so worried. And on top of that, my mom's health continue to deteriorate and I cannot. Dry cleaning is not a career aspiration for me.
Brad Smith: We all appreciate people who do devote their lives and careers to dry cleaning, but I can understand what you're saying.
Fei-Fei Li: Yes. So it was very easy career fair, senior graduate student. One thing led to another. I got offers from McKinsey, and that was not a huge surprise, but what was the surprise is the difference between the salary I would've been getting from McKinsey versus what my family has ever experienced in this country. And that was when I was wavering. I was just not sure, but the fact I was wavering also shows I loved my science and my mom was always the first person there. She's just always, she's just like, "Go for what you love. What are you agonizing about?" For her, it wasn't as agonizing as it was for me. And that is a courage. Again, as a kid, you never appreciate your parents till you become too old.
And in hindsight, it's an incredible courage. It's also an unconditional love. It was a gift of love. She was absolutely selflessly thinking about me, not thinking about her. Her lack of medical insurance. Her lack of healthcare and all that. She just thinks, "Go for what you love." And back to curiosity. Curiosity is not just moments of delight or sense of wonder. Curiosity is a state of being, and that state of being makes me happy and being a scientist makes me happy because I'm entitled to be curious all the time. And my mom knew that, and that's what she encouraged me to do.
Brad Smith: And your curiosity and that passion, that quest for science, took you down a path in the early two thousands when you created something that did not exist. I don't think people understood that it would become as important as it has. But basically you started to build a great data set of images. You called it ImageNet. Tell us about ImageNet, what it is, how it started, what gave you the inspiration to put it together.
Fei-Fei Li: Yeah. Thanks, Brad. ImageNet was a project that took us three or four years starting 2006, 2007 to build till 2009, 2010. And at the end, it was the largest data set that AI field has ever seen, consisted of 15 million images, hand curated with labels and cleaned up to across 22,000 visual object categories. And these images are internet images from everywhere in the world. Each one meticulously curated and organized, and they became a essential training data set for all AI algorithms, especially engendered the deep learning revolution because neural network algorithms will become the biggest consumer of ImageNet. And I think the significance of ImageNet is not the number of images per se or the modality that it's a visual data set. It's the concept of big data because now fast-forward to '24, 2025, big data is nobody even thinks about or questions about it.
If you want to do AI, you need three things. You need GPUs, you need neural network algorithms, and you need data. Everything we've seen from AlphaFold to ChatGPT to Gemini, to whatever, the Transformer, the latest and greatest AI algorithms, models, products, is built upon these three ingredients. But before ImageNet was developed, big data was not a concept. People were in fact working with small data. In fact, data was not appreciated at all. Data was just afterthought. After you build some algorithm and then you try to just grab something off somewhere and hope that you can train a model.
We observed before we made ImageNet that that's just a wrong way of thinking. Mathematically, these machine learning models need to be able to learn and generalize. And the way to drive that learning and generalization in addition to the architecture of the algorithm is actually through data and diverse data, a large amount of data. So that concept was initiated by ImageNet and therefore, very luckily for us, ImageNet had an outsized impact to the revolution of deep learning.
Brad Smith: And I think one of the reasons you could pursue ImageNet and certainly something that has been fundamental to your career has been a commitment to be part of academia, to work in a university to pursue basic research around, in this case, images and computer vision. Tell us a little bit about why you're so passionate about academic research for computer science.
Fei-Fei Li: Yeah. So Brad, I think my fundamental passion is about the ecosystem. And as a beneficiary of America's innovation ecosystem, I think academia and public sector plays a critical role. And of course we have seen that, is that when you have a very good academia ecosystem as part of the ecosystem to technical and scientific innovation, you can pursue ideas with curiosity. You can train students who are just with boundless energy and creativity and passion. And then you can work with the rest of the ecosystem to turn that into real technology and products that can be delivered in the hands of people. And that ecosystem is rare.
I mean, America truly has the best in both compared to the rest of the world as well as in the history of human civilization. And personally, curiosity-based research is best done in my opinion in academia by and large. Even today, we can talk about today, there are challenges. I still think academia is a fertile playground where researchers and students are allowed to just imagine the unthinkable or imagine something that people will laugh at. That's what happened to ImageNet. But then you do turn around and change the course of the technology and science.
Brad Smith: And share if you could a little bit more about that phrase, which I think is commonly understood in the world of academia, curiosity-based research, what does that mean and what is curiosity-based research compared to say, other types of research?
Fei-Fei Li: That's a great question, Brad. I think curiosity-based research fundamentally is linked to the word freedom, is that when you are a researcher or professor in a university and academic setting, you don't have a manager or a director who tells you, "Fei-Fei, come here to Princeton or Stanford and these are the three things you should work on." No, you are just given a desk, an office, and an opportunity to apply for funding and then do whatever you want. Of course, it's not as free as you think. Whatever you want is based on your expertise. I was trained as a AI scientist. I'm not going to do chemistry. Even if I try to do that, at least at the beginning without a credential, I would not have street cred to get funding.
But because of that freedom, so specifically like ImageNet, I was a young assistant professor. At that time, I just landed in Princeton. I went back to Princeton as a faculty. I had a lab of two students. Nobody told me what I need to do. I was exploring. I was passionate about this big data idea, and then I start just do it. In the meantime, of course, I was so excited. I chat with my colleagues and some of them will say, "Well, that's a bad idea." Some of them would even go as far as saying, that might be an idea that hurts your career. But even with these pushbacks or debates, nobody said, "You're not allowed to do it." They would just say, "This is my opinion, it's a bad idea, but you are still allowed to do it."
I also got rejected by grants. But again, even with that rejection, it's still freedom. You can do it if you can pull it off, but no one says no. And also on top of that, no one says you should be doing what. And that allows so much curiosity to flow. And of course, I don't want to rosify it. It also requires resilience and perseverance and courage because you do hear comments and opinions and you have to decide for yourself, do you want to change course or stick it through?
Brad Smith: One of the keywords that you mentioned is funding. You're given the freedom to pursue your curiosity, but you do need funding. It may be for lab equipment, it may be for computing time, it may be to employ graduate, help postgraduate fellows and the like. In the United States, a lot of that funding comes from the federal government, the National Science Foundation, the National Institutes of Health. What role did the NSF or other federal funding play in the development of your career, in the pursuit of your research?
Fei-Fei Li: Overall, it's critical, I would say. Of course, you have to take it down and analyze which project got funding. But as you said, for my early years, absolutely, the National Science Foundation is one of the major funding source. Another major funding source was the Navy Research Lab funding, we call Office of Navy Research, ONR. The Office of Navy Research is especially known in the field of computer vision to create this multi-university research initiatives. And they would fund a group of professors from different university to try to motivate multi-university collaboration and multidisciplinary collaboration.
And throughout my career, I have been on so many ONR, we call MURIs, multi-university research initiatives, and that's critical. Also, industry played an increasingly important role. When I started, gosh, 20 years ago as a young faculty, my very first industry grant was from Microsoft. Thank you, Microsoft. So a new faculty fellowship, I remember because I was really new. But gradually, my lab has received funding from Google, from Microsoft, from Amazon, from international partners like Panasonic and Toyota, Nvidia. So that ecosystem of federal funding as well as industry funding back into academia to grease the wheel of innovation is so critical.
Brad Smith: And the other aspect of this ecosystem is for those of us in industry. It's a little bit like standing on the shoulders of giants. There is so much research that comes out of academia that we build on top of. And I think this has been true since really the dawn of serious computer science research. It's been true of many industries in the United States, and the research comes out of the universities in multiple ways. I mean, you all publish, we read your papers, we derive insights and bring those into our products. And another critical aspect of just the economic and innovative dynamism of the United States is that people start in a university, and then they leave and they start a company or they remain in a university these days and they still start a company. You are today, both an academic and an entrepreneur. Tell us a little bit about this entrepreneurial pursuit that you are now focused on.
Fei-Fei Li: Yeah. You're right, Brad. And that's what I love about America's tech innovation ecosystem. It's very porous. Nobody gets to be forced to be put in boxes. And you can even be put into multiple boxes maybe. We're all Schrodinger's cats. And it's because this tech is, especially AI these days, is so fast moving. People from industry and younger generation than academia, there's so much flow of knowledge and capabilities. It's almost impossible to be so boxed in. And another thing about academia is you're allowed for leaves and sabbaticals, and of course you are also allowed to leave academia.
So just a year ago when my new round of sabbatical or leave is coming up, I was feeling pretty restless because I realized my own field of computer vision is going through a new revolution. The technology such as Transformers, diffusion models, neural rendering, and all this is recombining, deep learning, computer graphics and computer vision into a new generation of possibility, which I call spatial intelligence. It's really the fundamental technology to understand and generate 3D worlds and allow the interactivity that we have never imagined before.
And as I was getting so excited by this idea of spatial intelligence, I was recognizing that it was a particular idea that academia doesn't have enough resource for. Because it requires data, it requires computing, and it requires also a larger group of talent focusing on solving that. So together with my former students and just young colleagues in the field who are so at the forefront of the field, we started chatting about this idea. We get more and more excited. We decided, you know what? Let's be entrepreneurial. Let's be curious. Let's be bold. And we want to solve the problem of spatial intelligence. We want to create world models that would allow a new family of products that allow users to experience the ability to create 3D worlds to interact in it, to use it for their productivity and creativity. So that's the path we're on for at World Labs.
Brad Smith: Well, as we come to a close, I think it's an interesting moment in time to look at the issues of our day and reflect on, I think the various lessons that emerge from the conversation we've had. I think 2025 is a year when many people will ask, quite rightly, so, where are our tax dollars going? How is the federal government spending money? What are programs worth funding? And what are programs that are not? When the conversation turns to basic research, the role of universities, the federal support for universities, the ability of universities to stay financially healthy. As someone who came to this country and has pursued this career, what advice would you have for lawmakers in Washington as they think about this aspect of federal policy?
Fei-Fei Li: Yeah. So Brad, for the past five years, I maybe at the beginning reluctantly, but now I'm actually proactively have been advocating for resourcing our public sector research and technology because not only I personally benefited from it, but as we discussed, I see this as a critical component of a unique ecosystem of innovation that no other country or no other timing history has ever had. And I want to contribute to protect and to grow this ecosystem.
So even as an entrepreneur who are on the side of the industry research side, rather than the academic research side, I gain even more understanding of how important it is to resource our public sector and academia, because that's where so many innovative ideas come from. That's where students are trained, and that's where the academia can also play a role to arbitrate and convene challenging debates and just really get people to have rigorous discussions of very critical and challenging topics. So I would say to any lawmaker and policymaker that treat this ecosystem, treat the public sector and academia's role in this ecosystem as a strategic asset to invest. And that asset is both people asset as well as knowledge asset, public good asset. And we need to invest in this to... That's, in my opinion, the only way to keep this ecosystem healthy.
Brad Smith: And as you can probably tell, it's something with which I emphatically agree. It's so interesting because I think as I travel the world and over the years have met with leaders in other countries, I've often encountered government officials who've said, "We have studied the United States," because we wanted to understand what was making the US so successful when it came to economic competitiveness, research, innovation, the real dynamism that I think is a hallmark of the United States. So many times people have said, "As we've studied the United States, the thing that has jumped out is what we've been talking about here, this ecosystem that is in so many ways grounded in this basic research, the federal support, the financial health. It's just impossible to imagine the United States being successful without it."
But I also think sometimes it is easier to see something in someone else than to recognize it in the mirror. Maybe we in the United States don't often understand it well enough ourselves. What I think is fascinating about your personal story, your experience is you bring two critical pieces together. A young girl in New Jersey who could never have afforded tuition if you had to pay for it yourself, but could go because you got a scholarship that was made possible because of the financial health and other scholarship programs that are available. And then once there, could work, study more and then pursue this passion and bring research to life yourself.
As it turned out, you saw something that no one else was seeing at the time. In the middle of this AI winter, you saw a brighter future that would come out of all of this big data, all of these images, and you created something that became a big building block for all of the advances that we're seeing today. So personally, I want to thank you not just for everything you've done, Fei-Fei, but for using your voice to help other people understand its importance. As we look to the rest of this decade and beyond, I think it's a lesson that we can't afford to forget, and you do us all a service by reminding us of it.
Fei-Fei Li: Well, thank you. That's very kind you, Brad. As you said, I am a story of this ecosystem, a story of a successful immigrant, a journey, but really I owe this to every individual who has helped me. In my book, I talked about so many heroes. We just talked about my mom, but really there are so many heroes, from high school math teacher, to graduate school, colleagues, to my students, to mentors, and they're all here because there's an ecosystem that supports all of us. So I want to see this continue.
Brad Smith: We all stand on the shoulders of each other, so thank you. Thank you, Fei-Fei.
Fei-Fei Li: Thank you.
Brad Smith: You've been listening to Tools and Weapons with me, Brad Smith. If you enjoyed today's show, please follow us wherever you like to listen. Our executive producers are Carol Ann Brown and Aaron Thiese. This episode of Tools and Weapons was produced by Corina Hernandez and Jordan Rothlein. This podcast is edited and mixed by Jennie Cataldo, with production support by Sam Kirkpatrick at Run Studios. Original music by Angular Wave Research. Tools and Weapons is the production of Microsoft, made in partnership with Listen.