When preparing a TED talk, there is one especially clumsy method called the verbatim-script method: you write out exactly what you plan to say, including the colloquial parts, word for word, then rehearse it and revise it again and again. It works for people giving TED-style talks for the first time. You get half the result for twice the labor.

What follows is one of my own failed examples.

Don’t Let Sheep Queue for You (Transcript)

Hello everyone. I’m T, from a very small business team. Twelve years ago, the first work I helped plan was a smash hit, but only within a small circle. People outside that circle had no idea it existed. Then, a few years later, that book suddenly became inexplicably, overwhelmingly famous. It became so famous that KOLs everywhere began citing it, elevating it into a grand theory and turning it into a classic handbook of business warfare. That book was The Three-Body Problem, winner of the Hugo Award.

If you listen closely, this story probably sounds familiar. We are never short of this kind of story. Take websites: Douban before, ACFun and Bilibili later, and now all kinds of short-video platforms. Originally they were all products with very narrow audiences, but once some new variable suddenly entered the picture, whether by accident or simply through the ferment of time, their audiences changed and broke through their old boundaries.

If that change happens to you, good. If it happens to your competitor, then it is very likely time for a requiem. In today’s language, borrowing again from The Three-Body Problem, it becomes a so-called dimension-reduction strike. Some people, for instance, say that short video dealt a dimension-reduction strike to what we thought was our real battlefield: social products. No one saw it coming.

So have you noticed that in the internet age, the society we inhabit is not much like industrial civilization, which demands massive construction costs? Nor is it like agricultural civilization, because you do not even own your own small plot of land. What is it like instead? A nomadic civilization. The internet shares several features with nomadic life. If every internet team is a shepherd, and every internet product is a sheep being raised by that shepherd, then you discover:

  1. Borders are unstable. No one’s pasture is permanent. Today you win and stand on a broader field; tomorrow you may lose everything and see the pasture ceded away.
  2. Income is unstable. Any pasture may contain countless shepherds all trying to graze there. Once too many people arrive, the field becomes a red ocean. The stronger your sheep, the more fodder it can seize.
  3. The size of the flock determines whether the family survives. If you have only a few sheep, it may be very hard to endure the winter. No matter how sturdy one sheep is, its output is still limited.
  4. Enemies surround you. Shepherds fight over pasture and over sheep. Beyond the pasture roam foreign cavalry and packs of wolves, all staring directly at your life.

Seen that way, we undoubtedly occupy a very fertile piece of pasture. Everyone here, like me, is a shepherd. Then as shepherds, we first have to solve a few problems.

First, we must learn how to herd sheep. Herding sheep is difficult. A flock may be the hardest team in the world to manage. It is loose and disorderly. Left to itself, it blunders blindly in all directions. Yet once one lead sheep starts moving, all the others charge after it without a second thought, never mind that wolves may be ahead or better grass not far away. The image is vivid, isn’t it? Look at our company, or at many products outside it. The group-buying wars in the distance, live streaming and short video nearby: all of them feel like classic “flock effects,” rushing in together and scattering together.

Second, the flock must keep expanding so that it occupies enough safe space. You need constantly to breed lambs of your own, or capture new wild sheep from outside. How to discover those new wild sheep is an enormous challenge.

Third, you need sheep of your own. If you are always herding for someone else, then the day that sheep is slaughtered or requisitioned, your own value disappears and you are expelled from the tribe.

So let us try to look at our work through the logic of raising sheep. I hope this way of thinking can help shepherds avoid rigid obedience, let us test new herding methods more proactively, and improve our chances of survival.

Let us begin with the question of learning to herd. Why sheep in the first place?

Our pasture has to grow; we shepherds have to survive. So your boss gives you a task: go herd sheep. Fine. But what is the boss’s real need? One sheep, or a whole flock? What if he needs to march out to war? Then what he needs is a warhorse. But does he really need a horse? His final goal may simply be to keep the pasture alive. He may be going out to raid, to negotiate a marriage alliance, or to hunt. What matters is achieving those ends. In other words, what I said just now was deliberately misleading. Everyone here is not necessarily a shepherd. You may be a horse breeder. You may be cavalry. Your identity changes with the performance target. It is not fixed forever.

So a clear sense of your current role, and a clear sense of how that role can change, are both crucial.

The second problem is our greatest personal confusion. Once you enter such a gigantic pasture, you discover that it looks nothing like what I described. There is no so-called flock effect, no noisy masses of sheep, no pellet droppings smelling of grass, no shepherds in sheepskin coats. Instead you find a huge, orderly, modern workshop. All the sheep are raised mechanically, lined up in neat rows. They eat when they are meant to eat, they are milked by machines when it is time, and they are shorn when their turn comes. It is like a German documentary on the age of machinery: everyone is simply a worker on an assembly line. All you have to do is squat at your station, and when the sheep passes by, you reach out, pluck a tuft of wool, and clean its tail.

That may be the most frightening fact of all. Sometimes it frightens me in dreams. But it is already happening. I believe everyone here can sense it, at least a little. The engineers who exited the stage in great companies like IBM, Nokia, and Siemens are our fallen predecessors. Leaving aside personal gain or loss, the trend is obvious: the pastoral workers on the assembly line will sooner or later be replaced by AI colleagues like Tingting, YouTu, and Dreamwriter.

But for a giant pasture, that may not even be a good thing. If everyone advances in orderly fashion, the pasture may still be moving forward, yes, but no one will be trying anything new. We already have several sheep so gigantic they may be larger than African elephants. Sheep like that require preplanned routes, material schedules, huge teams to support and assist them, just so they can keep moving.

But where is the next frontier? No one knows. Finding it requires cavalry willing to charge, not merely logistics troops who know only how to transport supplies.

[Section omitted below.]

Now let us turn to the problem of expanding the flock. Sheep look slow. They seem to spend their lives serenely lowering their heads to graze in place. That image is deeply misleading, and it causes many people to slow themselves down to the sheep’s rhythm. But if you slow down, your sheep will slow down too, because sheep are a species shaped by adaptation, not by brute dominance. In fact, sheep can run extremely fast, or they would never survive encirclement by wolves. Their impact is also tremendous. If you have ever seen the folk sport of ram fighting in the villages of northern Shaanxi, you know the appalling crash that erupts when two rams collide headfirst. You do not want to be the one they hit.

If a flock wants to expand, internal breeding and reproduction are not enough. It is even more important to domesticate and bring in new sheep from outside. Internal breeding alone, breeding among close kin, is deadly. Everyone understands that. Ancient royal houses loved inbreeding because they believed their bloodlines were exalted beyond compare. In A Song of Ice and Fire, the Targaryens kept marrying among themselves and eventually produced the Mad King, Aerys the Mad. During the Northern and Southern Dynasties, Liu Song in the south also practiced close-kin marriage, and several emperors in a row turned out mad, ending at last in dynastic collapse.

So if we want our flock to expand, “new business” is inevitable. That means going out into the wilds to domesticate and discover new sheep.

[Section omitted below.]

In 1992 there was a cyberpunk novel called Snow Crash. What was it about? In Snow Crash, a virus spreads quietly through the world, and those infected collapse into a vegetative state. The way it spreads is by tricking you into scanning a QR code it displays. That was in 1992. The people in that novel all live inside an internet-like space called the Metaverse and access it through goggles. Looking back now, it is almost identical to our AR glasses, and one small step farther gets you the world of Ready Player One.

Another example: the Hugo-winning science fiction masterpiece A Fire Upon the Deep was written in 1993. In that book, civilizations throughout the universe communicate through a network called the Net of a Million Lies. How do they communicate? By posting messages, or what we would simply call shitposting. Twenty-five years later, we discovered that all the essential elements of our social tools were already exhausted there. There was no real breakthrough at all. That interstellar net was a Moments feed twenty-five years ahead of schedule.

Why? Why were science-fiction writers back then so formidable? They sound almost like prophets.

One hypothesis goes like this: the explosive technological development of the last twenty years happened because humanity was digesting the fruits of science that had accumulated since the Third Industrial Revolution. The theoretical groundwork was already in place. What followed was harvest season, so naturally there was an explosion. Likewise, the explosive growth of the internet in the last decade came because we were still digesting the previous era’s achievements. As time passes, because there have been no fundamentally new changes in basic science, stagnation becomes unavoidable. We are drawing infinitely close to the technological singularity. After the singularity, human-era intelligence and technology will become incomprehensible to us, just as goldfish cannot understand human civilization. But before the singularity, human creativity will keep declining. We are now on that road.

That seems to explain why innovation is hard, why expanding into “new business” is hard: every idea you have already feels as though someone else has had it first. Four-legged sheep are easy to find. Sheep with three eyes are not.

Is that really true? My answer is no. The singularity hypothesis exerts very little real influence outside basic science. In every other field, what actually matters is what I would call the rearview-mirror theory: every invention, innovation, and strategy depends on mature, already existing technologies. Take Douyin. Once that model caught fire, and even produced major effects on the to-B side of business, did we discover any profound new technical breakthrough or patent behind it? Not really. Every feature, once you saw it, made you say: oh, that’s all? When we thought variety shows had already exhausted every possible trick, Produce 101 appeared, and it was even licensed from Korea. Did it use some radically new camera system or 3D pipeline? No. They were still sheep. Perfectly normal sheep. We, by contrast, had been looking for an entirely new species. The difference is obvious.

This year marks more than six years since we entered the mobile internet era. So what I want to say is this: we need to find sheep, new sheep, not new species. No one is asking us to go out into the wilderness, discover a dragon egg, bring it home, and hatch a black dragon. What we need are endless new bloodlines of sheep. Countless flocks of sheep. That is how we move infinitely close to the singularity.

[Section omitted below.]

Yes, we are the herdsmen of the internet, the light cavalry of Tencent, resisting incursions from alien tribes, guarding the wealth of the herders, summoning formations beneath the stars, and pushing past the singularity’s waterdrop. We have to find our own borders, keep trying, and try again, so that our flocks can keep expanding and produce fighting rams able to survive on their own.

Thank you.