The assignment sounded simple until one tried to do it. In 2003, Charles Zhang, the founder of Sohu, asked a young computer science graduate student to build a search engine. Wang Xiaochuan was still at Tsinghua University. He had worked part time at ChinaRen, the campus network Sohu had acquired, but building a search engine meant putting a large technical system together from the inside. Eleven months later, Sogou was online.
An engineer might have taken the launch as the end of the story. Wang had built a working machine. The awkward part came next: people needed a reason to use it. He noticed that a simple directory site called hao123 was sending about a third of the traffic to Sohu’s home page. Sogou’s sophisticated search work was fighting for attention while a page of links quietly did the everyday job of getting people where they wanted to go. It was a lesson in the difference between making software and making a habit.
Wang would eventually find his route through the most ordinary of actions: typing. Sogou Pinyin, the Chinese input method released in 2006, would become the first stage of a business strategy that took users from the keyboard to a browser and then to search. The idea sounds tidy in retrospect. Living through it required years of persuasion, a stubborn browser project and an ability to accept that the most consequential product might be the least glamorous one.
The prize before the problem
Wang was born in Chengdu in 1978. As a teenager, he won a gold medal at the International Olympiad in Informatics and entered Tsinghua University in 1996. He studied computer science there as an undergraduate and then as a master’s student. Competition had offered a familiar bargain: a difficult problem, a clear answer, a medal if the answer was right. Business was less obliging. A search engine could be technically sound and still fail to command the home page of a stranger’s day.
While at Tsinghua, Wang joined ChinaRen part time. After Sohu bought it, he stayed connected to the company and moved from campus software into search. At 27, he became a Sohu vice president. The title recognized technical ability, but it did not exempt him from a founder’s more humbling discovery: the market does not grade elegance. Wang later described seeing the traffic flowing from hao123 as a turning point. The site was simple; its route into people’s lives was not.
The pinyin input method supplied a different answer to the traffic problem. Chinese characters do not map one to one onto Latin letters on a keyboard. A pinyin tool interprets what someone types and offers the characters they intend. That turns a small strip of software between hands and screen into a place users visit constantly. Sogou’s version spread rapidly. A Tsinghua alumni account puts its market share at 2 percent in its first year and 40 percent in its second. The figures describe an adoption curve, but the human fact beneath them is simpler: the product saved people effort many times a day.
“If you focus, you’ll think up a hundred ways to solve a problem.”Wang Xiaochuan, in a 2018 interview
A rocket assembled from daily routines
Wang’s next thought was that the input method could lead somewhere. Someone typing a query might open a browser; someone in a browser might use Sogou Search. He called this the “three-stage rocket.” It was less a metaphor for explosive speed than a practical map of user behavior. Search had been the destination all along, but the keyboard gave the company a departure point. The browser was the bridge.
The plan met resistance. Zhang opposed building Sogou’s own browser, according to Wang’s later account. Wang pressed ahead with a small team anyway. This was not the theatrical defiance of an overnight founder myth. He worked inside a large company, where disagreement could stretch across meetings, product cycles and budgets. In a speech to Tsinghua graduates years later, he described replaying his own mistakes and arguments repeatedly. The browser eventually launched, and Wang said its traffic helped Sogou Search climb to roughly 10 percent share over two years.
There is a useful bit of modesty in this arc. Wang did not discover that users secretly wanted a search engine with better engineering. He discovered that they were already doing something else. The input method met that behavior; the browser connected it to search. His work changed from solving an isolated technical problem to designing a chain of ordinary choices. The insight travels well beyond one Chinese internet company: every product must answer the question of how it will enter a routine.
The long route to the opening bell
In 2010, Sogou began operating independently from Sohu, with Wang as CEO. The move gave the team its own business functions and a clearer identity. It also required relationships beyond the engineering room. Alibaba backed the new company; Tencent later became a strategic partner. Wang’s technical project had become a negotiation over capital, traffic and control. The founder who once trusted that a good search engine on a home page would find its audience now had to work through partners who could help provide one.
Tencent’s 2013 partnership gave Sogou a new ally in a crowded search market. It also put Wang in a role far removed from the competitions of his youth. There was no single algorithm that could settle whether Sogou should rely on a rival platform, remain within Sohu, or pursue a separate future. Each choice changed the audience it could reach. The partnerships did not erase Sogou’s engineering story; they made that engineering visible to more people. By then, Wang had spent a decade learning that the route into a user’s day could be as important as the product waiting at the end of it.
Sogou listed on the New York Stock Exchange in November 2017. The interval from the original search assignment was fourteen years. Wang had been there through the early build, the keyboard’s climb, the browser dispute, the spinout and the partnerships. The listing gave the journey a neat date, though it is the middle years that explain his character. In a 2018 interview, he said difficult conditions sharpened his attention. That sounds less like a taste for adversity than the result of having spent a long time solving a problem that refused to stay solved.
His connection to Tsinghua continued during those years. He helped establish support for students and research, and in 2016 represented Sogou in a 180 million yuan gift to the university for an institute devoted to intelligent computing. The gesture joined the place where he had learned to build software to the company that taught him how hard it was to place that software in the world. He also helped start a campus intelligent-agent competition that, by the time of his 2016 alumni profile, had been running for two decades.
The second act speaks the same language
After Tencent acquired Sogou, Wang stepped down as its CEO in 2021. In 2023 he founded Baichuan AI in Beijing. Two decades had changed the technical landscape. Language was no longer merely something a machine helped people type, find or translate; a new generation of models could generate and organize it. Baichuan began by releasing open language models, including two in its first hundred days, according to the company. In 2024, TIME named Wang to its list of influential people in AI.
The continuity is more interesting than the change of company name. At Sogou, Wang repeatedly worked at the seam between human expression and a machine’s response: pinyin prediction, search queries, speech tools. Baichuan extended that interest into language models. The challenge of making a capable system useful has followed him. An impressive model can produce a demonstration; a lasting product must fit into a real task and earn repeated trust. Wang’s first career gave him an unusually literal example of that distinction: a keyboard people used every day.
His public comments also reveal a founder willing to revise course. In a 2026 interview, he discussed the difficulty of narrowing Baichuan’s strategy after an ambitious start. He did not present the company’s path as a straight line. That candor belongs beside the older story of the browser and the stubborn years before Sogou’s listing. Again, he was working through a problem whose first answer did not hold.
The teenager from Chengdu won a competition with a correct answer. The adult founder kept discovering questions whose answers changed once users entered the room. One concerned search. Another concerned how Chinese speakers typed. Today, with Baichuan, he works on a broader version of the same problem: what should software do with language, and how will people actually use it? The opening move remains wonderfully unromantic. Watch what people do. Then begin there.