Long read: The Chinese whiz kids of Silicon Valley

The central point of Ms Zhou’s superbly reported story is that AI – whether in America or China or in the fluid boundaries between the 2 nations’ tech communities – is now largely a disruption powered by formidable Chinese intellect. There are 3 distinct drivers of Chinese success in this arena.

The first is China’s formidable educational system which is structured very differently compared to India’s increasingly dysfunctional education system. Ms Zhou explains:

“Growing up, I personally witnessed how a generation of whiz kids were targeted for success by the Chinese education system. Coming of age in the 2000s, my classmates and I were children of the one-child policy. Our families invested everything into our education — especially mathematics. In primary school, I was placed in math-olympiad classes, and assigned geometry and algebra puzzles designed to identify whether I was truly gifted…

Chinese society has long promoted scientific and technical proficiency, an effort that traces back to the fall of the Qing dynasty in the early 20th century. Back then, public intellectuals argued that China’s lack of scientific knowledge left it vulnerable to Western invasion. Excelling in the sciences became a matter of national defense and pride.

A century later, the educational system still heavily emphasizes science and technology. When I attended a top-ranked public high school in Hangzhou around 2010, there were 11 science-focused classes, and only two humanities classes. I enrolled in a science class….While many American high schoolers dreamed of becoming athletes and musicians, in China excelling in math and physics was considered the ultimate proof of their achievement.

Eventually,  I moved to Hong Kong to study journalism — a bizarre choice in the eyes of my friends. An education influencer would later famously tell Chinese parents they should knock their children unconscious if they want to study journalism, because there would be no money to be made.”

The second is the meritocratic but Darwinian ethos of both China and Silicon Valley. Ms Zhou writes: “The AI boom has spurred a new class of Chinese founders and engineers into the spotlight. These are people like Carina Hong, a 24-year-old who grew up as a math lover in Guangzhou. As a teenager, she studied calculus and number theory and competed in math olympiads. She went to MIT, then Stanford, before dropping out to found Axiom, an AI mathematics startup that develops systems to verify AI reasoning. The 1-year-old company, based next door to an old Facebook office in Palo Alto, is valued at $1.6 billion. Math olympiad training in China made her resilient to pain and gave her a “Scout” mentality, while her education in America gave her cultural and language fluency, Hong told me over Zoom. She sums up her life philosophy as “scientific ambition.”

Tim Shi graduated from the elite Yao Class at Tsinghua University, a program known for having trained some of China’s brightest students into tech founders and computer science professors. He also came to Stanford and dropped out of the Ph.D. program. He joined OpenAI and left a year later to co-found an AI company, Cresta. The company automates customer service, and was valued at $1.6 billion in 2022. He has since become an investor in other AI firms, including Anthropic.

In Silicon Valley, Chinese researchers have founded AI startups that make everything from video generation models to agentic assistants to smart notetakers. Chinese scientists are among the co-founders of Elon Musk’s xAI, Mira Murati’s Thinking Machines Lab, and Yann LeCun’s AMI Labs. When Mark Zuckerberg unveiled Meta’s Superintelligence Labs, seven of the first 11 hires, including the chief scientist, were from China…

Almost everyone described the Chinese AI world as juan, or “involution,” a buzzword that refers to incessant, grueling competition. Depending on whether someone has connections or specializes in a particular line of research, one engineer can earn $500,000 a year, while another may earn $50 million. A Google scientist making more than $600,000 said he was anxious about not being able to afford an apartment located in a good school district….“Others are making millions, and I want that, too,” he told me.”

The third driver of the Chinese domination of AI is luck i.e. Chinese techies were in the right place at the right time when AI took off five years ago:

“When ChatGPT launched in 2022 and investors poured money into any and every company associated with AI, Chinese talent was in the right place at the right time. The immigrant Ph.D.s, now working from Silicon Valley’s AI labs, put in 16-hour days designing model architectures, preparing training data, and racing to beat benchmarks, with the same persistence and competitiveness they had honed in Chinese schools. “Chinese people move about like an army,” one Nvidia researcher told me.

During my stay at the Facebook House, I reconnected with the kind of math kids who had intimidated me so much in my school days. They spoke on condition of anonymity because they were not authorized by their employers to talk to a reporter. In a Fremont restaurant, I had lamb skewers with two former classmates. One was now a Nvidia scientist; the other was a machine-learning engineer at a midsize software company….

The Nvidia engineer had studied civil engineering in China, but switched to computer science when he came to the U.S. That was the default path for Chinese students in the 2010s, he told me. He recalled developing a spirit of persistence and focus as a student, and applying it to his Ph.D. studies and to building AI systems. When Chinese scientists shared their latest research, they sometimes added the idiom ri gong yi zu — one pawn forward every day…..The Nvidia researcher’s wife was taking a career break to manage the couple’s stock portfolio. Some friends of his had started companies, and he was considering doing that, too.

In Mountain View, I shared blood curd stew mao xue wang with a Google DeepMind scientist and an ex-Meta researcher. Both had graduated from China’s top-ranked Tsinghua University — the Google guy wore his Tsinghua shirt to dinner. “I’ve been good at math since kindergarten,” the ex-Meta researcher told me. Growing up in a backwater town in central China, he had taught himself olympiad math and physics in high school. “The regular curriculum was too easy, so I read up on more difficult things to feel some happiness of learning,” he said.

…In AI, event invites and job openings circulate on WeChat as much as on X. “Chinese is the language of AI,” Jonathan Li, a Chinese American who works at an AI startup, told me at the conference.”