Written by Xiaobing
On October 19, 2026, Meta Chief AI Officer Alexandr Wang posted a selfie on X with a caption of only seven words: got the HRT gang back together!
The comments section exploded.
This seemingly casual group photo conceals one of the most absurd facts in today's tech world: the three people in the photo, together with several old classmates who are not present, all took part in the inaugural internship program at Wall Street quantitative trading giant Hudson River Trading in the winter of 2015—a program that admitted only 10 people in total.
Ten years later, the companies founded or led by these 10 interns have a combined valuation exceeding $100 billion.
On the right side of the photo is Wang himself, 29, founder of Scale AI, who was recruited by Zuckerberg last year for $14.3 billion to join Meta as Chief AI Officer, steering Meta's Superintelligence Labs.
On the left is Jeff Yan, founder of Hyperliquid, a decentralized exchange built by an 11-person team with zero funding, which posted $900 million in profit last year and has a project market cap of $76 billion.
In the middle is Scott Wu, founder and CEO of Cognition, whose AI coding agent Devin is growing enterprise customers at a rate of 50% per month, with the company's latest valuation at $48 billion.
The last time all three appeared in the same room at once was probably that winter of 2015, in the trading firm's office by the Hudson River.
Winter by the Hudson River
The bond among this group of geniuses began back in high school.
Jeff Yan and Scott Wu both won gold medals at the International Olympiad in Informatics (IOI), and Wu won gold three times, topping the field with a perfect score in 2014. Alexandr Wang was a regular at the USA Computing Olympiad and the Mathematical Olympiad, and they repeatedly crossed paths at various summer camps and training squads—a circle so small that almost everyone knew each other.
In the winter of 2015, these competition kids scattered across Harvard and MIT were brought together by one common name: Hudson River Trading.
HRT is one of the world's top quantitative trading firms, headquartered in New York, with daily trading volume at one point accounting for more than 15% of the U.S. stock market.
In the winter of 2015, HRT tried recruiting interns for the first time. The quota was extremely small—only 10 people in total. Jeff Yan was then a junior at Harvard, Scott Wu a freshman, and Alexandr Wang a freshman at MIT.
The design of this internship program was itself very "quant": it lasted three weeks and was run as a competition, with rankings after every round. According to Jeff Yan's later recollection, the top two in every round were always him and Scott Wu.
But the most historic moment of this internship may not have been the trading competition itself. According to an in-depth interview with Jeff Yan by Colossus magazine, it was during his HRT internship that Yan first heard about Bitcoin from the pitches of two former partners. Although he did not act immediately at the time, he soon found Ethereum's yellow paper and read a description: a computer run jointly by the whole world that no one could shut down. That seed grew into Hyperliquid years later.
Also that winter, Alexandr Wang, then just 19, said something at his own birthday party to the Harvard and MIT classmates seated around the table that was later quoted repeatedly:
"Why can't it be us?"
Three paths, one origin
After walking out of HRT's office, these young people chose completely different tracks. But they share a common underlying logic: using algorithmic thinking to rebuild a traditional industry.
Alexandr Wang: From data labeling to Zuckerberg's AI brain
After leaving HRT, Wang made a decision that surprised everyone: he dropped out of MIT.
In 2016, at 19, he co-founded Scale AI with Lucy Guo. What the company did sounded not sexy at all—labeling data for AI training. But Wang's judgment was extremely precise: he skipped specific AI applications and bet directly on the "shovels" of the entire AI industry.
As the AI wave swept the globe, Scale AI became a core supplier to giants such as OpenAI, Meta, and the U.S. Department of Defense. In 2021, the company's valuation soared to $7.3 billion, and Wang, at 24, became the world's youngest self-made billionaire. In its 2024 funding round, Scale AI reached a valuation of $13.8 billion, backed by Nvidia, Amazon, and Meta.
Then came the deal that shook Silicon Valley in June 2025. Meta acquired 49% of Scale AI for $14.3 billion, at an overall valuation of $29 billion. As a core condition of the deal, Zuckerberg required Wang himself to join Meta as Chief AI Officer, leading the newly established "Meta Superintelligence Labs."
A young man of 28 became the overall commander of the AI strategy of the world's largest social media empire.
Jeff Yan: From quantitative trader to builder of on-chain financial infrastructure
If Wang chose the "shovel business" of AI, Jeff Yan chose the hardestcore path in the crypto world: rebuilding trading infrastructure.
After graduating from Harvard, Yan did not rush to start a company. He worked at Google and then at a self-driving company, while never stopping his observation of the crypto world. In 2020, he founded Chameleon Trading, a cryptocurrency market maker. This experience gave him a deep understanding of the structural flaws of centralized exchanges, and the collapse of FTX gave him the decisive push.
In 2022, Yan disbanded the Chameleon team and began building Hyperliquid with his own funds—a decentralized perpetual contract exchange running on a blockchain. No VC investment, no token presale, an 11-person team, starting from zero.
This "anti-VC" path is almost unheard of in the crypto industry. Yan's confidence came from the trading profits accumulated at Chameleon, and even more from a quantitative trader's relentless pursuit of product: latency must be low, throughput must be high, and user experience must approach or even surpass that of centralized exchanges.
ICE CEO Jeffrey Sprecher once said that Hyperliquid's trading activity "exceeds that of Nasdaq."
Today, Hyperliquid is one of the companies with the highest profit per capita in the world. Last year, this company's mere 11 employees generated more than $900 million in profit. Founded only four years ago, the company has reached a market cap of $76 billion.
Scott Wu: The man who lets AI write code itself
Scott Wu has the most terrifying competition record of this group. Three IOI gold medals, champion with a perfect score in 2014, MathCounts national champion, third place in Google Code Jam. After studying at Harvard for two years, he dropped out and co-founded the AI social networking platform Lunchclub.
But what truly made his name in Silicon Valley was Cognition, which he co-founded at the end of 2023 with Steven Hao and Walden Yan (the three hold a combined five IOI gold medals). In March 2024, Cognition released Devin, a product billed as "the world's first AI software engineer," capable of independently planning, writing, testing, and deploying code.
Devin sparked enormous controversy in the tech world, but the market gave its own answer: Cognition's valuation soared from $350 million in early 2024 to $10.2 billion in September 2025, and then to $26 billion in May 2026 (raising $1 billion), with annualized revenue surging from $5.4 million in mid-2024 to nearly $500 million in mid-2026.By September 2026, annualized revenue had approached $1 billion, and the latest valuation was reportedly $48 billion. Wu himself says that more than 90% of Cognition's internal code is now written by Devin itself.
A three-year-old startup whose valuation went from $350 million to $48 billion—a more than 130-fold increase.
More than three people
There are only three people in that selfie today, but this circle is far larger than one photo can contain.
According to Scott Wu himself, the HRT internship program was not the starting point of this group's friendship. As early as high school, many of them had already met through various Olympiad competitions. And the names extending outward from that small circle each occupy a place on today's technology map.
Jesse Zhang, one of the same cohort of HRT interns. His previously founded social gaming company Lowkey was acquired by Niantic (the developer of Pokémon Go). In 2023, he founded the AI customer service agent company Decagon, focused on using AI to replace enterprise customer service. Less than three years after founding, the company's valuation has reached $4.5 billion, with customers including giants such as Duolingo and Hertz, and it was selected for Forbes 30 Under 30 in 2026.
Johnny Ho, a three-time IOI gold medalist and Harvard alumnus. He met Scott Wu and Jeff Yan in high school. In the 2013 U.S. IOI team photo, he stands shoulder to shoulder with Steven Hao and Scott Wu.
After leaving the competition circuit, Ho worked as a quantitative trader at Tower Research Capital, then joined Quora as an engineer. In 2022, he co-founded the AI search engine Perplexity with Aravind Srinivas and others, and currently serves as co-founder and chief strategy officer. Perplexity's latest funding round reportedly values it at about $30 billion, and it is redefining the trillion-dollar search engine market.
Demi Guo, a "genius girl" from Hangzhou and a 2015 IOI silver medalist. She holds a bachelor's degree in mathematics and a master's degree in computer science from Harvard, and was once the youngest full-time researcher at Meta AI Research. In 2023, she dropped out of her PhD program at Stanford and co-founded the AI video generation platform Pika with classmate Chenlin Meng, reaching a valuation of $470 million and more than 16 million users. She previously interned at HRT and Quora—again the same circle.
Steven Hao, a 2014 IOI gold medalist and a graduate of MIT's mathematics department. He was an early core engineer at Scale AI during Alexandr Wang's time there, and later joined Scott Wu's Cognition as co-founder and CTO. According to Forbes estimates, the 30-year-old Hao's personal net worth has reached about $1.3 billion, making him the youngest American on the 2026 Forbes 400.
From Hao's trajectory, the "talent flow" within this circle is clearly visible: he brought the engineering experience and understanding of AI data accumulated at Scale AI into Cognition. And Cognition's three founders Wu, Hao, and Walden Yan hold a combined five IOI gold medals; they poured the algorithmic thinking honed in competitions directly into the core architecture of AI coding agents.
Why have quantitative trading firms become founder incubators?
This is not a coincidence.
When Wang reposted that 2015 intern group photo in early October this year, he added a note: "HRT 2015 intern class is the new-gen PayPal Mafia."
The story of the PayPal Mafia is well known. Elon Musk, Peter Thiel, Reid Hoffman, David Sacks, and others came out of the same company and went on to create Tesla, Palantir, LinkedIn, Yammer, and other companies, defining the technology landscape of the previous decade.
But this HRT group differs fundamentally from the PayPal Mafia. The bond of the PayPal Mafia was one company and a shared entrepreneurial experience. The bond among this HRT group is earlier and deeper: what they share is a way of thinking.
What does it mean to come from a competition background?
It means they have been trained since their teens to approach problems with an optimal-solution mindset, with an almost obsessive pursuit of efficiency and an intuitive understanding of complex systems. Quantitative trading firms are a natural amplifier of this way of thinking: at HRT, they learned how to extract signals from massive noise, how to make decisions within millisecond-level time windows, and how to build systems that can optimize themselves.
Jesse Zhang wrote in an article: "The generation that grew up doing math, programming, and science competitions accounts for a disproportionately high share of today's founders."
Quantitative trading firms are precisely the key springboard for this group from the competition circle to the startup circle. Jane Street, Citadel, DE Shaw, Two Sigma—similar stories stand behind all these names, but the concentration in HRT's 2015 class is indeed unusually high.
Of 10 interns, at least 4 became unicorn founders, and the companies are valued at more than $100 billion in total. This "talent conversion rate" exceeds that of technology giants such as Google and Meta, which are known for producing AI talent.
The deeper reason may be that these quantitative trading firms provide not only technical training but also an instinctive understanding of market structure. Wang applied this understanding to AI data infrastructure, Yan to on-chain financial infrastructure, Wu to the automation of software engineering, Zhang to the restructuring of customer service, and Ho to the disruption of search engines.
They solve different problems, but the way they break them down is identical: find an inefficient system and rebuild it with algorithmic and engineering capability.
Geniuses always meet early. But what is truly moving about this story is not how smart they are. Competition gold medals proved that long ago. What is moving is that the era has finally turned to the page that belongs to them.
Ten years ago, what was the image of this group? Math competition champions, programming geeks, invisible people in social settings. In the hierarchy of any American high school, they were all placed under the same word: nerd. For a long time, that word was not a compliment; it meant eccentric, boring, out of step with the mainstream.
But the world after 2024 has given nerds an unprecedented stage.
As AI becomes the underlying operating system of every industry, and as code itself begins to write code, the people capable of commanding all this are precisely those who have lived in algorithms and formulas since childhood.
Wang, Yan, Wu, Zhang, Ho, Guo, Hao—these Chinese surnames that are awkward for white people to pronounce—have moved from Olympiad arenas to the center of the stage of the era.
The nerds who were once despised have finally ushered in their golden age.