Yesterday I saw the news that Zhang Xuefeng had passed away at 41 from sudden cardiac death. Three days earlier he had been posting his running logs on WeChat Moments, logging 72 kilometers that month. He felt unwell after an afternoon run, was taken to the hospital, and three hours later he was gone.
He left behind nine companies, an education group valued in the hundreds of millions, and a nine-year-old daughter.
I had just written a piece about the bottleneck in AI startups, mentioning that the body is the hardest infrastructure in this wave of change. Work intensity is increasing, the pace is accelerating, but the body doesn't become more durable just because technology advances. I didn't expect to see news like this the very next day.
This is not an isolated case. Ding Yun of Huawei, 53, died after a 28-kilometer run. Sun Jian, chief scientist at Megvii, 45, suffered a sudden illness in the early hours. Zhang Rui, founder of Chunyu Doctor, 44, died of a heart attack. Every year in China, about 550,000 people die from sudden cardiac death—roughly 1,500 a day. These figures existed before AI, and the work intensity of the AI era will only make them grimmer.
But what stayed with me for a long time wasn't the topic of health preservation. It was that in the face of these events, all the narratives about technological change and productivity explosions suddenly seemed very light.
$100 Million Later, Then What
The experience of Peter Steinberger, founder of OpenClaw, belongs in the same conversation.
He spent 13 years building a PDF SDK used by nearly a billion users, with clients including Apple, Adobe, and Dropbox. In 2021 he sold the company to Insight Partners for roughly $100 million.
Then he crashed.
He described himself as "utterly broken." Every morning he woke up with nothing to look forward to, no real challenge. He tried travel, socializing, therapy—nothing filled the void. He told Lex Fridman: "If you wake up with nothing to look forward to, that boredom hits very fast." By the end of 2024, he couldn't write a single line of code.
Later he rediscovered his passion through AI agents, and in three months built one of the fastest-growing open-source projects on GitHub. But more than the comeback itself, I'm haunted by that blank period. Someone who had already "made it" could still be completely stuck when facing the question of meaning.
When external drive disappears, what keeps a person moving? This isn't just his question.
A ¥120 Billion Void of Meaning
On one side is AI anxiety; on the other, AI mysticism.
During the 2025 Spring Festival, DeepSeek fortune-telling shot to the top of the hot-search lists. You input your birth date and time, and the AI casts your astrological chart, reads your wealth luck, and judges your romantic fate. Zi Wei Dou Shu, Bazi, Tarot, I Ching—it covers them all, with more interactivity than a human fortune-teller.
This is not a niche phenomenon. iiMedia Research predicts that in 2025 China's AI fortune-telling market will exceed ¥120 billion, growing 35% annually. One leading platform has over 80 million registered users and conducts 2 million divinations a day. The user profile is also telling: 73% are aged 20 to 35, and 82% hold at least a bachelor's degree.
It's not seniors passing time. It's highly educated young people seeking a kind of certainty from an algorithm.
I don't see this as a resurgence of superstition. It's more like a symptom. When technological change is so violent that the old scripts—work hard and you'll succeed, own a house and a car and you've made it—begin to loosen, people instinctively reach for something more fundamental to hold onto. Where do I come from? Where am I going? GPT can process all it wants, but it can't answer these. The faster technology runs, the more unavoidable these questions become.
Will AI Replace Human Jobs?
Most people's anxiety follows this chain: AI can work → my job is replaced → I have no value.
But there's a hidden premise in this chain: human value equals labor output.
Hold it up to the light, and that premise looks pretty suspect.
If AI massively increases total social productivity, with digital workers running 24/7 and marginal costs approaching zero—which is exactly what AI as an industrial-revolution-level transformation means—then why must everyone still labor personally to obtain resources? With cheaper productivity that doesn't eat, sleep, or need social security, people should theoretically be set free. To soak up the sun, to do what they're passionate about, to create something for fun or simply to idle. Wouldn't that be a better life?
But no one is thinking about this. Everyone is anxious about having their job stolen. The real thing to worry about is something else: how the fruits of production are distributed.
Productivity is becoming ever more abundant, but the distribution mechanism is still the old one: capital concentrated in the hands of a few, operating through market prices. The result is that a tiny minority decides where society's main resources go. Technology amplifies existing disparities, not equality.
Rather than panic, we'd do better to have a serious conversation: when technological change reaches this level, how do we achieve more reasonable redistribution? Once people's time is freed up, where should their needs be directed?
We're Missing a Positive Vision of the Future
A while back, a friend said something that struck me: what we're missing most right now is an effective imagination of the future.
Think about it, and it's true.
For decades, we more or less agreed on what the future looked like: economic growth, urbanization, globalization, living standards steadily rising. That narrative was strong enough that it didn't need to be discussed. All of society was saturated with an expectation of flourishing.
Now that consensus has shattered. Most of our images of the future come from Hollywood. The Matrix, Blade Runner—their palette is dystopian: wealth polarization, corporate rule, runaway technology. Video games occasionally offer interesting visions of virtual-world socializing, but by and large, exciting positive visions are scarce.
This is dangerous. Pessimistic imagination has its value, but if it's all we have, people are left with nothing but fear and defensiveness. A society driven by fear makes choices that are all about contraction.
Technology Is Neutral; Choices Are Not
Technology itself is neither good nor bad. How it's used, how it's distributed, how society organizes itself to adapt to it—these determine the outcome.
These are not questions engineers can answer. Economists, sociologists, policymakers—these are the people who should step to the forefront precisely when technology is running fastest.
AI is shaking foundations deeper than most people realize. The meaning of labor, the logic of wealth, how time should be spent. These assumptions haven't been seriously questioned in decades, and now they're all loosening.
The barrier to content creation is falling, and the cost of film, television, and cultural products is dropping too. Under these conditions, earnestly imagining and investing in a positive vision of the future is no longer a luxury. The United States is stepping back from leading such imagination; if someone else is to pick up the mantle, now is the time.
Back to the beginning. Zhang Xuefeng is gone at 41. In the face of this, all the tech competition, all the industry anxiety—honestly, none of it is worth that much emotional energy. What deserves serious thought are those old questions: how to make society a little better, how to let people live a little more like human beings. I believe AI doesn't amplify skill; it amplifies passion—finding what you truly care about may matter far more than chasing the technological trend.