This AI Wave Favors People Exactly Like Him, but He Studied Business Instead
Provincial first prizes in both the math and informatics olympiads, Tsinghua, then a postdoc at Harvard Business School. In late 2025 a friend asked him to start an AI company together; this wave favors exactly his kind of résumé. Too bad he doesn’t do large models. What he studies every day is the stock market.
Suzhou used to hold a live essay contest. He won second prize every time; first prize always went to his childhood friend. More than a decade later, he is still writing. She stopped long ago.
Talent decides how high you can go. Love decides how long you can keep going.
Suzhou used to hold a live, on-site essay competition: students from across the city were brought into one classroom, given a topic, and had two hours to finish on the spot.
Richard came home with a prize every time. Second prize.
But Richard never thought he had a gift for writing.
Because the first prize, every time, went to his childhood friend.
That friend was genuinely gifted. Winning first prize in essays was effortless for her; she could also paint and play music, and later went to a well-known design school to study interior design.
Yet more than a decade later, Richard is still writing. His friend stopped long ago.
It was never about talent. “I just love writing,” Richard says.
He’d had a strong urge to express himself since high school. He loved Lu Xun’s essays, loved watching that man insult people in roundabout ways. Later he got on Zhihu, China’s Quora, and has been writing there for over a decade, becoming a featured contributor, covering economics, current affairs, and historical research, writing to this day. I asked whether he uses AI to help with his writing. “No, all by hand,” he said. “Writing is the one thing I want fully under my own control.”
Richard’s real talent lay elsewhere.
In high school he won provincial first prizes in both the math olympiad and the informatics olympiad, and got into Tsinghua.
The year he filled out his college application, he saw that the School of Economics and Management said it wanted people good at math, computers, and English. That sounded exactly like him, so he put it as his first choice. One step into the business school, he kept going, all the way to a postdoc at Harvard Business School.
Near the end of his postdoc, he held several decent faculty offers back in China. But he found his interest in academia fading. He loved doing research; he just didn’t love writing papers. Even his internships were at applied companies. He ended up at an investment advisory firm, building factor models.
Markets rise and fall every day, and someone has to figure out what is actually driving them. The common factors are broad categories; “consumption,” for example, is one factor. Richard works at a finer grain, splitting a category like consumption further into a tablet factor, a phone factor, more than 200 thematic factors in total. These factors explain part of why the market moves, and quant funds buy them to design their own money-making strategies.
The funny thing is, he studies market factors all day but barely touches stocks himself. He’s risk-averse.
So has AI helped here?
Factor mining, to this day, still relies mainly on statistical methods, not large models, because statistics is more interpretable. AI helps him speed up R&D, run exploratory studies, and turn his experiment code into production code. A person who had spent his whole career in pure research roles slowly picked up an engineering foundation, because of AI.
He told me a small story. At the end of 2024, the company’s entire data pipeline ran on Amazon S3, and he understood none of it; every time he needed data for research, he had to ask the developers. “Next year I have to learn this,” he told himself. AI did get him up to speed, fast.
At the end of 2025, a friend asked Richard to start a company together.
The people this AI wave favors most are exactly the elite-school, math-olympiad types. On paper, that’s him.
Too bad he doesn’t do large models.
The opportunity brushed past him.
Since then, he has been taking online courses, cramming knowledge about large models.
AI has genuinely helped with Richard’s personal project.
There is a project he has been running since high school, for over a decade now: translating the Project Euler website into Chinese.
Project Euler is a collection of hard problems that fuse math and programming. Math alone won’t crack them; most require writing code. Back in high school, Richard built a small website himself and translated the problems one by one, by hand.
Before AI, this was bitter work.
Before 2015, the site was hosted on the server of someone he knew online. Then that person deleted the server, and all the data vanished overnight. He started from scratch: taught himself CSS, taught himself JavaScript, learned deployment, moved the site to a wiki, then to GitHub Pages.
When ChatGPT came out, he was among the earliest users of the ChatGPT Projects feature, because it let him write a custom system prompt and get the translation running automatically.
Now he has written automation scripts with Claude Code. Whenever a new problem appears on the Project Euler site, the script translates it into Chinese and posts it to the corresponding page.
More than ten years, from hand-typed HTML to Claude Code. The tools have changed, one generation after another. The only thing that hasn’t changed is that he is still maintaining it.
And still writing.