How Every Technology Revolution Gets Captured: The Pattern

To understand how technology gets captured, you only need to hold one question steady while everything around it changes: when a new tool arrives and makes something genuinely easier, who ends up with the benefit? Not who invents it, not who first cheers for it — who owns it once the rest of us can no longer live without it. I have spent years chasing that question backwards through history, from the artificial intelligence being built right now all the way to the first mud-brick granary, and the answer keeps rhyming. The gains are real. The capture is real too. It is the same move, a new machine, every time. This essay is the map behind everything else I write — the pattern the other pieces fill in.

I want to be careful here, because this is easy to get wrong in two opposite directions. One is the reflex that every new technology is a trick, that progress is a lie, that we were happier in the fields. That is nonsense, and I don’t believe it. The other is the reflex that every new technology is a gift, that the market sorts the winners fairly, that anyone left behind simply failed to adapt. That is also nonsense. The truth sits in the uncomfortable middle: the tools really do help, and the help really does get taken. Both things are true at once, and the pattern is how they fit together.

The five steps, in order

When I line up ten thousand years of tools — writing, ploughs, printing presses, spinning frames, steam engines, assembly lines, search engines, recommendation feeds, large language models — the successful captures nearly always move through the same five steps, in the same order:

1. Something genuinely helps. A tool arrives that lets people do more with less. This is not a con; it is a real gain, and people adopt it because their lives get easier.
2. People come to depend on it. The tool stops being optional. It becomes the ground you stand on — the way things are simply done now, so that opting out means opting out of ordinary life.
3. The dependency is enclosed. Someone draws a fence around the thing everyone now needs and puts a gate in the fence. Access moves from open to owned.
4. The terms are quietly rewritten to favour whoever owns the machine — the split of who gets what shifts, a little at a time, in the owner’s direction, rarely by a single dramatic decree.
5. The costs land on everyone else. The people who depended on the tool absorb the downside — lost livelihoods, lost autonomy, lost privacy, lost bargaining power — while the gains concentrate at the gate.

Read those five as a machine, not a morality tale. There is no cartoon villain required. Each step can look reasonable from the inside; that is exactly why the pattern survives. Let me walk it through history, quickly, because the sweep is the argument.

The granary and the ledger

Start at the beginning — the first cities of the ancient Near East, the first surplus grain. Storing grain was a genuine help: it meant you could survive a bad year, feed a town, plan ahead. People came to depend on the store. Then someone had to keep track of what went in and what came out, and so the earliest writing we have is not poetry — it is accounting. Ledgers. Marks pressed into clay to record who owed how much grain to the temple or the palace.

Look at what that quiet little tool did. Writing enclosed the surplus by making it legible to whoever held the tablet. The person who controlled the ledger controlled the count, and the count decided who ate and who owed. The terms got written — literally — by the class that could write, and the costs, the debts and the tribute, landed on everyone who could not. The granary fed people. It also invented the record-keeping that let a few people own the fact of the food. Same five steps, at the dawn of the whole project.

The earliest writing we have is not poetry. It is accounting — a tool for recording who owes grain to whom, which is to say a tool for owning the count.

The commons, fenced

Jump forward to England, and to fields held in common — pasture, woodland, and meadow that ordinary villagers had the customary right to use to graze an animal, gather fuel, and scrape by. Over centuries, and with a hard acceleration into the eighteenth and nineteenth, that common land was fenced off and turned into private property. This is the enclosure of the commons, and it is the cleanest example of step three you will ever find. The land didn’t change. What changed was the fence and the gate.

The defenders of enclosure had a real argument, and I want to grant it its due: fenced, privately farmed land was often more productive, and higher yields fed a growing nation. That is the “something genuinely helps” step, and it was not imaginary. But notice what the productivity argument does — it answers a question nobody was really asking. The question was never only how much grain. It was who gets it. Enclosure raised output and, at the same time, stripped the customary rights of the people who had depended on the commons, pushing them off the land and toward the towns. More was produced. Fewer people owned the means of producing it. The costs landed on the dispossessed.

The factory and the Luddites

Those dispossessed people walked into the next chapter. The power loom and the spinning frame were genuinely productive tools; nobody sensible denied that cloth could be made faster and cheaper by machine. But the skilled textile workers of early industrial England were not, as the cartoon has it, superstitious fools who hated technology and smashed it out of ignorance. That is a slur written by the winners. When you actually look at the Luddites, you find skilled workers making a precise economic objection: the new machines were being introduced in a fashion — by whom, on what terms, and for whose benefit — that broke their trade and handed the entire gain to the factory owner. They did not smash every machine. They smashed the ones owned by employers who were cutting wages and quality. Their fight was about the terms, which is step four exactly.

They lost, and they were hanged and transported for losing, and the framing of them as anti-progress buffoons has done a century and a half of work since — because if the people who object to the terms of a capture can be dismissed as enemies of the future, then no capture ever has to answer for itself. Keep that trick in mind. We will see it again, dressed in newer clothes.

The feed and the attention economy

Now come almost all the way forward. The internet was a genuine gift; I am not going to pretend otherwise, and I use it every hour. Search that answers any question, a map of the whole world’s knowledge, a way to reach anyone — real help, adopted for good reasons. And then we came to depend on it utterly, until being unfindable online is close to not existing. That is steps one and two, and they happened fast.

Then came the enclosure, and it was subtle because nothing was fenced that you could see. What got enclosed was attention — yours and mine, the finite hours of human notice. The business model that won was not to sell us a product but to sell us, or rather our attention, to advertisers, and to optimise every feed and notification to harvest as much of it as possible. This is the attention economy, and it is capture without a fence: the terms were rewritten so thoroughly that the users became the raw material. You think the feed is a window. It is a till.

Follow that logic to where it now points and you get a shape worth naming. When the essential infrastructure of daily life — the platforms you shop, work, socialise, and think through — is privately owned, and you must pay rent (in money, data, or attention) simply to participate, you are no longer quite a customer in a market. You are closer to a tenant on someone’s land. That is what people mean by technofeudalism: not a metaphor for the internet being annoying, but a serious claim that ownership of digital ground has started to look more like a landlord’s power than a merchant’s. Whether or not you accept the full argument, the direction is the pattern’s direction — dependency, enclosed, then rented back.

And now, AI

Which brings us to the machine of this exact moment, and to why I find the pattern so hard to unsee. Artificial intelligence is, once again, a genuine help — I would not insult your intelligence by denying what these systems can do. And we are already at step two, coming to depend on them, wiring them into how we write, code, diagnose, and decide.

So run the checklist honestly. Step three, enclosure: the models that matter are trained on data-centre scale compute that only a handful of firms on earth can afford, which is the whole story of who owns AI — the capability is being fenced by the sheer cost of the fence. Step four, the terms: those systems were trained on the collective output of humanity — our writing, our art, our code, scraped largely without asking — and the value is being routed to the owners of the model, not the makers of the training data. Step five, the costs: they land, as usual, on labour, which is why the honest version of will AI take my job is not a yes or a no but a question about terms — automation doesn’t simply delete work, it shifts who captures the value the work used to carry. Same move. New machine. It could not be more legible if it tried.

The same move, a new machine, every time: something helps, we depend on it, the dependency is fenced, the terms are quietly rewritten, and the bill arrives for everyone outside the gate.

Why “progress” is the wrong word for the argument

Here is where I have to take on the two phrases that get used to end this conversation before it starts, because they are the pattern’s best defenders.

The first is “progress.” The word smuggles in a conclusion. It takes a question with two parts — are we making more? and who is getting it? — and answers only the first, then acts as if it has answered both. Enclosure was progress by output. The power loom was progress by output. Both raised the total and redistributed it upward at the same time. So when someone says a new technology is “progress” and expects the argument to be over, notice that they have quietly deleted the distribution question. More for whom is not a footnote to progress. It is the whole of what progress means to the person living it.

The second is “it creates new jobs eventually.” This one is true, and that is what makes it slippery. Yes: the weavers’ grandchildren found other work; new industries did eventually arise. But “eventually” is doing enormous, cruel labour in that sentence. Eventually can be two generations. The man whose trade is destroyed at forty does not get to live in the equilibrium the economists are describing; he gets to live in the transition, and the transition is where the costs are paid. Telling him that aggregate employment recovers in the long run is telling him that his particular life is an acceptable rounding error. The “new jobs” argument is a statement about the total and a silence about the distribution — the same trick as “progress,” wearing a kinder face. Neither phrase is a lie. Both are ways of not answering the only question that matters: who gets the benefit, and who pays the bill, and when.

The fifth step is the one we can rewrite

If I stopped here you would be right to find this bleak, so let me tell you what a decade of looking at this has actually convinced me of — and it is the opposite of despair. The five steps are not a law of physics. Four of them are close to inevitable: useful tools spread, people depend on useful things, valuable dependencies attract owners, and owners press for better terms. Fighting the first four is mostly fighting gravity. But the fifth step — who ends up carrying the cost — is not fixed at all. It is a political outcome, decided by how much the people outside the gate are organised, and it has been rewritten, hard, many times.

The clearest proof I know is recent and specific. When the Indian government passed farm laws that would have restructured agricultural markets in ways millions of farmers read, correctly, as an enclosure of their bargaining power, they did not accept the terms. They organised — a vast, disciplined, months-long protest that camped on the capital’s doorstep through a winter and a pandemic — and the laws were repealed. That is the story of how India’s farmers won, and I keep it close because it is a live, twenty-first-century demonstration that step five is negotiable. Capture is not a one-way ratchet. It can be reversed by refusal.

It has been reversed before, too, by building rather than blocking. The cooperative movement is capture run in reverse: when working people jointly own the store, the dairy, the bank, or the utility they all depend on, the gate and the people it serves are the same people, and the terms cannot be quietly rewritten against them. Land reform, where it has genuinely happened, is the enclosure of the commons played backwards — taking ground that had been fenced into a few hands and returning use of it to the many. The commons can be re-opened. The point is not that any of this is easy. The point is that it is possible, which is a different and more useful thing than hope.

And it is worth saying plainly: technology itself is often on the side of the many. Writing, once it escaped the temple scribes, armed everyone who could learn letters. The printing press broke a monopoly on knowledge. The same tools that enable a capture frequently supply the means to resist the next one. The machine is not the enemy. The terms are the fight.

How to use this lens

So here is what I actually want you to walk away with — not a mood, but an instrument. The next time a genuinely useful new technology arrives, and it will, and it will genuinely help, run it through the pattern before you decide how to feel. A handful of questions do most of the work:

• What is this quietly making me depend on? The tool you can no longer imagine living without is the one to watch, because dependency is what gets enclosed.
• Who owns the gate? When this becomes essential, who sits at the chokepoint — and how many of them are there? A capability only a handful of firms can afford is a fence, whatever the marketing says.
• Who made the value, and who captures it? Trace the gain back to its source. If the people whose work or data or attention created the value are not the people collecting it, you are looking at step four in progress.
• Where do the costs land, and how soon? Not in aggregate, not eventually — for actual people, in this transition. “It nets out fine in the long run” is an answer about the total and a silence about the distribution.
• What would it take to rewrite the terms? Because step five is the one that answers to organisation, ownership, and refusal — and that step was never the machine’s to decide. It is ours.

None of this tells you to fear the future or to smash the loom. It tells you to keep your eye on the one question the cheerleading and the doom-mongering are both designed to make you drop: not is this new? and not is this powerful? but who gets the benefit? Ask it about the granary, the fenced field, the factory, the feed, and the model, and you get the same answer often enough to call it a pattern. Ask it early enough about the next machine, and you might change the answer. That is the whole reason I keep telling this history — not to mourn the captures we slept through, but to be awake for the one arriving now.