I recently came across the story of Zhang Xue and his motorcycle brand, and it hit me hard.
Zhang Xue builds motorcycles. Born in 1987, he dropped out of junior high and became a motorcycle repair apprentice at fourteen. At nineteen, he chased a Hunan TV crew over a hundred kilometers of mountain roads in a downpour on a beat-up bike, just to show off his riding skills on camera. In 2013, he arrived in Chongqing with twenty thousand yuan in his pocket, and later co-founded Kove Motorcycles, growing annual sales from 800 units to 30,000. In 2024, at thirty-seven, he gave up his equity in Kove and started Zhang Xue Motorcycles, aiming to build self-developed engines.
Most people only see the end results: in 2025, Zhang Xue Motorcycles posted revenue of 750 million yuan, and this past March, they won back-to-back races in the WorldSSP middleweight class at the WSBK World Superbike Championship Portuguese round, finishing 3.685 seconds ahead of second place. It was the first time a Chinese motorcycle brand had won at this level, breaking decades of dominance by Ducati, Yamaha, and Kawasaki.
But what keeps replaying in my mind isn't any of that. It's the small details he shared in interviews.
I'll Run the 800 Meters
One segment left a deep impression. They were trying to improve the machining precision of a core engine component—from five si down to three si. One si is 0.01 millimeters, so five to three means tightening the tolerance from 0.05 mm to 0.03 mm, a 40% precision improvement. This isn't just tweaking a parameter. It requires upgrading cutting allowances, tool control, material consistency—the entire process chain—while maintaining uniformity at scale. No one in the domestic supply chain had done it before, and suppliers refused to cooperate, deeming the risk too high.
Zhang Xue told them: I'll cover the R&D and trial-and-error costs. You don't need to worry. If it succeeds, we'll share the rewards.
Think about that scene. He wasn't exactly a big player in the industry chain—just a downstream buyer, probably seen as a small client by those suppliers. But what he said followed zero "buyer" logic.
What was it, then? It was that the act of doing this thing held meaning beyond profit, beyond risk. He desperately wanted to see it happen, and he was willing to bet himself on it.
Ten Years Grinding on One Pair of Glasses
Zhang Xue isn't the only one.
Misa Zhu, founder of Rokid, is known to everyone as Misa. In 2012, he was still at Alibaba when he attended the Google Glass launch event. Seeing someone skydive while wearing the glasses shook him to his core. He later said only two thoughts ran through his mind: this thing will change the world, and we can do it better.
He decided to leave Alibaba and start a company. Jack Ma personally talked with him for four hours, asking why he had to go. His answer: "AI and AR will change human lifestyles, and eventually they will become the same thing."
Rokid was founded in 2014. But the AR glasses supply chain was nowhere near mature, so they had to survive by making smart speakers first—the Rokid Alien and Moonstone. Then Baidu drove smart speaker prices down to 89 yuan, while Rokid sold theirs for 1,399. There was no way to survive the subsidy war against BAT. Misa himself said publicly that the AI industry was heading into a twelve-to-eighteen-month winter.
As soon as they caught their breath, he threw himself back into glasses.
By 2025, Rokid had built 49-gram AI+AR glasses—the lightest in the world, supporting real-time translation across more than fifty languages. This February, German Chancellor Merz visited Hangzhou and tried on a pair personally. After experiencing the real-time translation, he said two words: "Accurate. Fast." Seven or eight German business executives in his delegation placed orders on the spot.
The guy who posted on social media ten years ago saying he wanted to make glasses—today, he put them on the German Chancellor.
Misa still posts two or three product-related updates on his social media every day. It's not marketing. He genuinely thinks it's cool and wants people to see it. He once said something that struck me as exactly right: "Entrepreneurship is paying 99% of your time learning things you don't particularly like, for the 1% that is your hobby and dream."
Not Businessmen
These people keep making me ask: what's the real difference between an entrepreneur and a businessman?
China has never lacked businessmen. From ancient times to today, too many people have sharpened their minds trying to figure out how to make money. Businessmen find the optimal solution within existing frameworks—they do whatever makes money, however efficiently possible. There's nothing wrong with that; society needs them.
But that's not what Zhang Xue and Misa are doing.
I later came across the Austrian School economists, and realized they spotted this distinction long ago. Mises called the entrepreneur "the driving force of the whole market system," distinguishing entrepreneurs from managers: managers optimize within given parameters, while entrepreneurs change the parameters themselves. Schumpeter put it even more sharply. He said what entrepreneurs do is "creative destruction"—dismantling old structures from within and creating new ones. He even argued that the primary motive driving innovators isn't profit, but a "will to conquer."
Looking back, Zhang Xue covering his suppliers' trial-and-error costs wasn't because he ran an ROI calculation and found it profitable. He simply had to see three-si precision become reality on a domestically made motorcycle engine. Misa spent ten years grinding on glasses and nearly went under, yet in 2014, right after Google Glass had failed, everyone thought AR was dead.
These people aren't running in a lane—they're paving their own.
Unitree's Wang Xingxing is the same. In 2016, he spent two months at DJI and left to found Unitree. At the time, quadruped robots were almost all hydraulic—expensive and clunky. He bet on electric drive and built core components in-house. In 2021, they launched Go1 for under 75,000. Today, Unitree controls over 60% of the global quadruped robot market. Liang Wenfeng moved from quantitative finance to AGI, hiring for passion and curiosity rather than degrees. He once said: "The follower's phase is over. It's time to lead."
What AI Actually Amplifies
Back to AI.
I used to think AI's value was in amplifying skills. Programmers code ten times faster, designers generate images ten times faster—skills multiplied by a coefficient. That sounds reasonable.
But in the process of pushing AI adoption within my team, I realized that's not it.
The bottleneck isn't whether AI can help you do something. It's whether you know what to use it for. It's the same problem as the million-dollar gap for coding agents—the tools are already there; what's stuck is cognition.
Before, ability was measured by "what can you do well," and finishing the job was the standard. Now the threshold for finishing something has been drastically lowered by AI. So what becomes scarce? The drive to keep doing something well, and doing it differently from everyone else.
The answer to that comes from passion.
I've observed a clear split within my team. Some people are improving rapidly with AI, not because they have strong technical foundations, but because they know what they want. Their direction is clear; they might be orchestrating ten different AI agents on different tasks at once, staring at one problem for the long haul. Others have perfectly decent skills but simply don't know what to do with AI. Their previous workflow was waiting for assigned tasks, completing them, getting feedback. That loop is broken. Now you have to find your own direction, drive yourself, and possibly go through stretches with no one cheering you on. You have to judge for yourself whether it's worth continuing.
Before, that kind of judgment mattered less, because organizational structure was making the decisions for you. Not anymore.
Builders and Promoters
Following this line of thought, I increasingly feel that the company as an organizational form will gradually weaken.
The future may look more like MCNs. Two core roles: the Builder, obsessively focused on building products and turning their vision into reality; and the Promoter, wielding influence to make good things seen.
Many Builders are themselves Promoters. Zhang Xue's interview stories spread organically. Misa posts product updates on social media every day. Wang Xingxing's robots performed at the Spring Festival Gala. These creators are their own distribution nodes.
Between these two roles, the loop might already close. The remaining organizational form may resemble an MCN: providing infrastructure, taking a thin slice of profit, offering some incubation support.
Go Find Your Passion
My mindset is quite different from before. Things I used to value—position in an organization, amount of resources, whether others recognize you—no longer feel like the most critical factors. Instead, I feel a greater sense of freedom.
What AI ultimately offers is the possibility of turning passion into reality. Before, if you had passion but no team and no capital, many things were simply impossible. Now, one person plus AI can accomplish more than a small team used to. This also means the way teams collaborate has to change—passion may have become more important than skills.
So the better question isn't "Will AI replace me?" It's "Where does my passion actually lie?" Do you want to build something, or do you want to help good things be seen by more people?
In moments of anxiety, it's easy to overlook one shift: AI is enabling more Zhang Xues and more Misas to emerge from unexpected corners. They may have no big-company background, no resources, just a direction they feel compelled to pursue, and they're pushing forward with AI.
I don't know exactly where or when, but within six months we should see plenty of them.