Illustration: Alberto Miranda

In May, I sat in a living room on the fringes of Hangzhou, one of the most technologically advanced cities in China and home to Alibaba, Unitree, and DeepSeek. In many ways, the party that was about to start resembled one that you’d find in any coastal metropolis: ten professionals in their 20s padding barefoot across the carpet, snacks on the dining table. Only the laptops scattered across the floor and couch suggested this was a different type of social gathering, one that has grown common in the past three years in China’s towns and cities. Our host, KC, has gathered a small circle of friends to learn AI.

KC’s face is still marked with acne; he is tall and thin, barely past 20, and already a vibe coder of some local fame. Today he’ll teach others how to use Codex and Figma to quickly build websites, Electron desktop apps, install packages — but the first step is teaching everyone to “put up the ladder,” the euphemism for deploying VPNs and relays that bypass OpenAI’s geo-block. Like any self-organized class, it doesn’t really begin until nearly two hours in. Only as the sky darkens do people settle down, grow patient, and start the class.

The demographic of the room is young, hungry, yet refusing to join or be excluded from the rat race of big tech — they self-identify as engineers, designers, influencers, or, to use a more fashionable term, one-person companies (OPCs). They are members of a social species typical of a generation facing 18.9 percent unemployment, astronomical real-estate prices in big cities, and a battered post-COVID economy, which a friend describes as petit bourgeois, overeducated, downwardly mobile young people who can just barely make the rent. A few years ago, low-cost towns like Liangzhu (where KC lives), Dali, and Anji started actively courting digital nomads and vibe coders like them. Amid government campaigns and a stream of social-media posts, young people started washing up there, hoping that learning AI while keeping expenses low might change something.

During the self-introduction session, one woman mentions, a little shyly, her old life at Tencent — which she had disliked so much that she quit to explore “the many ways of not going to work.” Another, with long black hair, introduces herself as a novelist serializing a long book on Jinjiang Literature City, China’s biggest women-focused web-fiction platform. A third says she used to work in fashion and now runs an “aesthetics and shopping-companion” business with a friend and does self-media and AI projects on the side. They are all convinced that AI could empower their work.

Since ChatGPT, a new consensus has hardened across China’s tech industry: AI is the next mobile internet. On the surface, the country seems to have arrived at a near-universal embrace of new technology. Xi Jinping has described AI as an essential engine for China’s “new quality productive forces.” From a Marxist-Hegelian perspective, AI is on the right side of history, and history must march forward. As Evan Osnos, a staff writer with The New Yorker, recently wrote, the future could be made in China. But beneath the AI frenzy lies a real dread: people’s fear of being thrown overboard by China’s next Darwinian round of industrial upgrading, with little choice but to go along.

At once an old Chinese reflex returned: When history accelerates and people lose their bearings, find a teacher. In a society where meritocracy and education still carry near-religious authority, the proper response to a disruptive technology is to embrace it and to enroll. Decades ago, it was “English fever,” when English-learning-as-self-transformation swept factory floors and university lecture halls alike, everyone convinced that fluency was the ticket out. Now the conviction has simply changed languages. Resisting this technology is futile, unwise — so better to seize the window before it closes. The state claimed AI will empower thousands of industries, with social-media timelines filled with proof that someone, somewhere, has already changed their lives with it. In response, for many people I met across China since the arrival of ChatGPT, one thing that kept surfacing was their hunger for learning about AI, a compulsion to keep upskilling. Lessons, lessons, lessons — I noticed how often the word surfaced during casual conversations.

In the U.S., backlash against data centers has united people across the political spectrum and generations. This “AI populism,” a term coined by the writer Jasmine Sun, reflects a belief that AI is not a neutral technology but an elite project driven by tech billionaires to concentrate power and displace workers. In China, on the other hand, a meme that’s now circulating goes: Westerners get restless when they can’t protest; Chinese people get restless when they haven’t found a teacher. People like KC self-organize classes; for others, the dread of getting left behind by AI has presented a business opportunity.

Hangzhou’s Future Sci-Tech City (?????) is one of those municipal declarations of intent to materialize the future. It is home to Alibaba’s headquarters and to the towers, campuses, incubators, talent apartments, and financial-service complexes through which a local government announces that it, too, belongs to the next technological epoch. For the sheer density of engineers packed into its office parks, people jokingly call the area “the center of the programmer universe.”

The industry of paid knowledge, and its priesthood of “knowledge influencers,” is nothing new — in the mobile internet age, people built an entire economy around it. Now this industry has a new direction: evangelize AI. The company I’d come to see was the operator of one of the largest “make money with AI” communities, a membership business that monetizes the promise of AI. The company is called Shengcai Youshu (????), roughly “The Art of Getting Rich.” It offers online classes, offline intensive training, and co-working space. Its Hangzhou office, where the founders sit, is one of the larger nodes in a network that spans the country. I was greeted by Mr. Tree, a middle-aged AI influencer on the company’s payroll, whose bio announces him to the world as being on a mission to “look for 100 business answers in the age of AI.” He is habitually nervous, easily and overly excited. His accent is not the polished, elite-educated kind, and he mispronounces his English words.

In one of the short videos of him teaching a class to a few dozen, he stands onstage before an audience old and young, their faces solemn, pens moving. “What is AI?” Mr. Tree asks loudly. “AI is your Silicon-based labor force! Most companies run nine-to-five, eight hours a day. AI runs 7-by-24 — round the clock, every day of the week!” Mr. Tree’s relentless high spirits are contagious, and they remind me of a Mormon elder in Orange County who once tried to bring me into his church when I was a college student freshly arrived from China. He walked me through the office, recording the whole tour on his DJI Osmo Pocket, as though everything he did needed to become material for his vlog.

The office’s theme color is green, the color of money (in fact, the company offers a class called “Making Dollars Together”). Meeting rooms have names like “Apostle of the Method” (???) and “ Evangelist” (???). Along the corridors, the walls are covered, like those of any company proud of its own legend, with founder lore, milestones, and mottos. I read that the founder, Yiren, after leaving Alibaba in 2017, vowed to “help young people who, like him, had once been poor but hardworking, change their fate by making money,” the kind of declaration you might hear from a gilded-age robber baron or a self-made oracle like Benjamin Franklin or Jack Ma — mission fused seamlessly with the libertarian lore of self-salvation. Another line on the wall managed the same mix of sincerity and self-regarding heroism: Whatever saved you, use it to save the world.

That afternoon, when I visited the office, Yiren happened to be in. We shook hands politely. He was thin and shy, soft-spoken, his hair combed with a precision bordering on severity. It was hard to square this quiet figure with his career: a man with a cult following who had recently turned himself into an AI chatbot (on WeChat) and issued a standing invitation: Ask me anything about growth, entrepreneurship, your projects, or making money.

Yiren first found influence online writing paywalled dispatches with titles like “365 Ways to Make Money in a Year.” As paid followers grew, his company, The Art of Getting Rich, funneled that traffic into something far more elaborate. Today, paid members are encouraged to collect something called Dragon Balls — a goofy, gamified name for the points awarded to those who help the community grow: posting useful content, recruiting newcomers, and then climbing the leaderboard. On one wall, a diagram lays out the order of ascent. You begin as an ordinary member by paying the yearly fee, then become a Dragon Ball holder, then a Navigator, then the Deep-Sea Circle — the executive tier — until, at last, you gain access to the co-working floor among the senior members and stand, in effect, at partner level.The higher you climb, the more people and resources become available to you. Am I in a cult? I wondered. The headquarters of some pyramid scheme?

Membership is not cheap. Entry to the online community costs ¥3,365 today, against an average Hangzhou private-sector salary of ¥7,671 a month, roughly half a month’s wages for an ordinary worker. What keeps people coming, I think, is a residual faith in upward mobility — the hope that behind a changed economy, there remains a way in. The company is now nine years old, with more than 80,000 paying members to date and more than 2,000 offline events hosted per year across Chinese cities. The more I read the company website — which promises to help people “break into new circles, trade notes with bosses and operators, find a community that fits, and never walk the road to wealth alone” — the more it resembled a kind of church, a place where career, marriage, and a child’s education might all be worked out.

Mr. Tree took me into a room with 20-odd people in their 20s to 40s. Some were solo entrepreneurs and senior members using the office as a co-working space, heads down at work. Mr. Tree asked everyone to stop so he could introduce me as a visitor from San Francisco, prompting someone to shout, “Take us to Silicon Valley!” The whole office laughed.

Beyond The Art of Getting Rich, which hosts AI courses of every kind, the cheaper, messier imitators — “DeepSeek 101,” “Get Wealthy With AI,” “Learn to Distill,” “Super AI Agent” classes — have now multiplied to fill every corner of the internet. AI is the only path forward. The last chance for ordinary people to defy fate and remake their lot. AI won’t replace people, but people who can’t use AI will be replaced by people who can. Before the pitch, the sellers usually offer a few tidy success stories: a stay-at-home mother, for example, who used DeepSeek to turn out video scripts at 30 yuan a job. Since last year, state-sponsored media has taken to calling this paid-knowledge economy a mess, warning that ordinary people are being “cut like leeks” — fleeced — by a course-selling economy that “takes advantage of people’s ‘skill anxiety’ about new technologies.” Online, I found one member’s verdict on The Art of Getting Rich’s own classes:

“Tried The Art of Getting Rich for a few days. Pricey! Over a thousand yuan. Got inside and found nothing of substance: they have the students themselves post their money-making stories to fill the place up, and if you post enough you can offset the tuition.”

And yet. After the office visit, I read some of the articles shared on the company’s news feed, and to my surprise, they were not bad. One walked through pairing Claude with Obsidian to create a kind of product-copy engine, for the struggling e-commerce seller managing a sprawl of listings that each need its own picture-and-text blurb. This article explained “the attention that AI tends to dilute,” and how to manage “context window,” and the way to prevent models from suffering “pattern collapse” and turning lazy. Most of it didn’t strike me as novice knowledge; Claude and Obsidian are no strangers to Silicon Valley’s hyperproductive tech set. But picture the reader who might actually pay ¥3,365 to learn it further. Maybe he’s a man in his 40s with a middle-school education, somewhere in China’s export-dependent manufacturing belt, running his business on instinct and hard work but not a word of English. He sells water bottles, desk lamps, plastic toys across the border to the richer half of the world, and in a perpetual battle with the blank spot where the Amazon product copy should go. Picture him handed Claude and Obsidian. The distance the article is trying to close is enormous, and it is closing it in earnest.

These are the people Mr. Tree and The Art of Getting Rich mean to reach: the ones from traditional manufacturing, from trades with no exposure to AI at all. In June, he posted, to his own feed, a thank-you note from one of his students, another typical profile:

“I’m working as a security guard right now. I came across The Art of Getting Rich late last year … for more than half a year I’ve been using all kinds of AI without stop … every day at work I’m interacting with Claude Code, on my phone and on my computer at home. My stint as a guard is almost over now — I’m about to start a job at a semiconductor-packaging company! Once the new work settles, I’ll keep building my AI skills.”

Mr. Tree’s caption reads: “Dude says he’s a security guard. Hell of a thing.”

This spring, a Chinese sociologist named Lian Si published A Journey Without a Map, an ethnography of ten groups of young people. Si is no dissident. He is executive vice-president of the China Youth University of Political Studies. Yet even from inside the system, his verdict lands hard. Many young people in their 20s are living in a reality unlike that of previous generations who could take economic miracles for granted. Uncertainty has become an ordinary fact of daily life, with risk seeping into every part of life — school, work, home, and marriage. I thought, Even a Party-approved scholar is allowed to say that now. His counsel to the young: Stop climbing. “Make peace with the ordinary,” he has said. “Shake hands with failure.”

In my search for a story about China’s AI optimism, I instead found stories about choicelessness and coping — about a crippled sense of agency, narrowed down to the few things still within reach, in a world that offers only one correct way to make sense of your own struggle: It would be your own fault if you got left out. Here is a state with vanishingly little upward mobility, promising “common prosperity” anyway by doubling down on high technology. When the Party holds fast to a Hegelian History — History with a capital H — in which technological ascent is not a choice but a necessity, what options remain?

If Mr. Tree’s story is true, then this silver lining is real — the upward mobility AI promises isn’t something people simply conjured out of desperation. Xiaokai, a 25-year-old film director I met near Hangzhou, is an example.

If you didn’t already know that he employed seven people and ran an AI video studio turning over several million yuan a year, his manner and thick Dongbei accent might place him as a street-corner hustler. Every 30 minutes he gets up to smoke. He wears slippers and a pair of beige shorts, his long hair tied back at the nape. Xiaokai’s humor is the Dongbei kind: bleak and wearily cynical, interspersed with random self-deprecating commentary, no tolerance of pretense or bullshit.

Xiaokai left home in Dongbei, China’s rust belt, at 18 to chase a dream of acting. Like many children of the Dongbei, he grew up while his parents were away as migrant workers. His was a lonely childhood, and he was raised by a grandfather and a television playing Jackie Chan and old Shaw Brothers films. Before he discovered Way to AGI (one of the largest AI learning communities online) in 2023 and taught himself AI filmmaking, Xiaokai thought he was a man with no prospects: a third-tier diploma from art school and behind him a string of failed ventures, livestreaming among them. Dongbei people make the best livestreamers, he told me. When he opens his mouth, the street smarts pour out, and you understand at once that this is a man who will never let himself go hungry.

Beside Xiaokai sat Big Wang, 42, Xiaokai’s COO. He’d been up all night working on a video for a district government in Shanghai, a celebratory piece, all gleaming towers and grand accomplishments, in the house style you can call the old cadres love. The peculiar challenge of this government commission is that the AI video needs to convincingly resemble a real production, complete with drone footage and a big-budget feel. Every building, every detail has to be faithful to the real thing — which is why Big Wang had spent the night hunting down each hallucinated tower and street, until nothing in the frame was something the machine had dreamed up. The trick, Xiaokai explained, is to conduct a whole orchestra of AIGC models into setting still images in motion — the right prompts, endless small adjustments — until a three-minute city documentary leaps onto the screen, sweeping and grand, the camera gliding as though it had really been there. Shot conventionally, he said, the thing would cost well over a million yuan. This one went for around a hundred thousand. At lunch, he insisted on paying and ordered far more than we could eat.

Neither of them could have pictured the life they have now — so much work coming in that they are up through the night without rest. The company does so well that they can’t say aloud what a single commission costs, for fear of the envy it would stir among their peers. They showed me a screenshot of one rival’s message, sent in pure spite. “He saw how much we’re pulling in, saw we’d landed a big one,” Xiaokai said, “and he just lost it.” Xiaokai’s social media is a model template of Chinese Gen-Z success, a timeline crowded with highlights: red carpets at the Beijing and Shanghai film festivals; his work aired on CCTV, the state broadcaster; the press anointing him “the first Gen-Z in China to turn AIGC into money”; and, flaunted for all to see, the 2024 Award of Excellence in the AIGC Short-Film section of the Beijing International Film Festival.

With today’s AI capabilities, Xiaokai explained, the best kind of plot is “a story that keeps losing people”: As it advances, it sheds characters, so you have to worry less about consistency. (Too many faces is a real problem in the current state of the models — the longer the clip, the more the features begin to warp and slide.) A perfect project, then, would be an Agatha Christie–type story — characters are trapped in isolated settings, then disappear one by one. Then he showed me a clip: a knife pressing slowly into human skin. The sound raised goosebumps over my whole body. “Look — the blood coming out, how real it looks. Killer.” He pointed it out to me, delighted with himself.

Even as a success story, Xiaokai is emblematic of a market so grim it demands endless grind. (I visited Xiaokai’s studio on a Saturday; three people were still working and all looked sleep deprived. They only take one day off each week.) Next year he plans to put a down payment on an apartment and get married — at a moment when the overwhelming majority of 20-somethings like him are adrift in China’s simmering deflation and cruel job market. He joked that once he had checked off the to-do list of his life, he’d finally make art about his hometown.

Dongbei is in deep demographic decay. Its young people leave for school and rarely return. A genre has emerged under the name “Dongbei Renaissance,” composed of films, fiction, music, all circling the region’s postindustrial ruin and its bleak iconography: abandoned factory compounds, laid-off workers, family shame, murder, violence. The people who make them are the children of the 1990s state-enterprise collapse looking back on their parents’ wrecked world, a point of view that has carried a constellation of writers and filmmakers into the mainstream.

Xiaokai wants to turn his core memory of Dongbei into an AI film. The memory, as he described it: a wide frozen river beside a vast, emptied factory town. Every winter the river froze thick enough to walk on. To keep from slipping, people scattered coal cinders across the ice; the coarse grains bit under a boot or a tire and held.

Xiaokai described every detail as though the film already existed. It’s a story about a father and a son. Across the frozen river runs a thin black path — coal ash the father has laid down, stroke by careful stroke, so that the son, away in some distant southern city, can find his way home safely. But the father dies before the son returns. The son learns of the death, comes home, and finds the ice long since melted.

In Silicon Valley, the running argument between the accelerationists and the doomers is premised on the idea that AI is the only variable that matters. It could go extraordinarily well or extraordinarily badly; utopia or dystopia will arrive in a matter of years, even months. The great AI researchers assured me, all the same, that one day we will stand to superintelligence as apes stand to us. But when I think of the film Xiaokai wants to make, or read the latest gospel Mr. Tree posts, or catch up on KC’s new projects, what stays with me is not Silicon Valley’s arithmetic about AI or Beijing’s grand plan, but people’s stubborn insistence on their own dignity, defiant of a future the big men are summoning.