Every influential scientist has their own origin story, and the Godmother of AI is no exception.
Feifei Li, a Stanford University professor best known for her work on ImageNet, is now the founder of World Labs, a year-old AI startup that has already been valued at more than $1 billion.
But her beginnings were much more humble.
Lee immigrated to the United States from China at the age of 15 and helped his parents run a dry cleaning business in Parsippany, New Jersey, to earn a living.
“We weren’t exactly well-off financially. My parents worked as cashiers and I worked in a Chinese restaurant,” she said in a Bloomberg Q&A. “My family and I decided to run a small cleaning business to earn money to survive.”
Lee said he likes to joke that he was the “CEO.” She ran the store for seven years, from when she was 18 until halfway through graduate school.
According to his LinkedIn profile, Lee went to Princeton University because it was close to his parents’ store. He then continued his business remotely at Caltech in California while pursuing his Ph.D.
“I was the only one who spoke English, so I took all the customer calls and did all the invoicing, inspections, etc.,” she said.
She said the experience taught her the value of resilience, a principle that continues to guide her career.
“As a scientist, you have to be resilient, because science is a non-linear journey. No one has all the answers. To find the answers, you have to go through these challenges. And as immigrants, you learn resilience,” she said.
At World Labs, Lee has big ambitions. She works on building world models. These are AI models that leverage spatial intelligence, which Lee says is “the ability for AI to understand, recognize, reason, and interact.” [with the world]. It comes from a continuation of visual intelligence. ”
A growing number of AI experts believe that world models are what will propel the AI revolution to the next level. As the name suggests, some believe that large-scale language models that are trained on languages and on which major products are currently based have limitations.
Lee said ImageNet, a comprehensive training dataset of visual information, is the forerunner of the world model.
At the core of Lee’s research is the idea that visual information, a passive way of understanding the world, is an important foundation for real-world behavior. That continues to be one of the end goals of some top AI builders, like Meta’s lead AI scientist Yann LeCun, who recently announced he was leaving to start his own global model startup.
The consistency between Lee’s research and her immigrant narrative is the same.
“I was always a curious kid, and science was the outlet for that curiosity, and that really grounded me,” she told Bloomberg. “I wasn’t interested in nightclubs or anything like that. I loved science.”
