Every Technological Revolution Promised Freedom. Read the Small Print.

Every technological revolution has arrived wearing the same costume. It shows up at the door promising to set us free — free from drudgery, free from ignorance, free from the tyranny of distance and time — and we open the door gladly, because who would turn away liberation? The plough promised freedom from hunger. The printing press promised freedom from the gatekeepers of knowledge. The steam engine promised freedom from the limits of muscle. The internet promised freedom from every intermediary at once. And in each case, something like freedom did eventually arrive. But it rarely arrived for everyone, and it almost never arrived first. The first thing a new tool does is not to distribute power. It is to concentrate it.

This is the small print on the contract nobody reads. The headline says liberation. The clause underneath says for whoever controls the tool. Understanding the gap between those two lines is the whole story of how technology and freedom actually relate — and it is the story I want to walk through here, era by era, because the pattern is too consistent to be a coincidence.

The plough that fed the granary

Start at the beginning, with agriculture, the original productivity revolution. Farming is usually told as the moment humanity freed itself from the anxious, hand-to-mouth existence of the forager. Settle down, plant grain, and you no longer have to chase the herd. Surplus becomes possible. From surplus comes everything we call civilisation: cities, writing, art, leisure.

All true — and all beside the point if you were an ordinary person living through it. A grain surplus is not just food; it is storable, countable, seizable wealth. Someone has to guard the store, and whoever guards it decides who eats. The archaeology of early settlements is not a picture of newly liberated equals. It is the first appearance of granaries, and beside them, the first appearance of durable hierarchy — the story I trace in the first granary and inequality. The tool that could have fed everyone instead created the first class of people who owned the food and the first class who worked for it. Freedom from hunger, yes, but delivered as dependence on whoever held the key to the store.

The foragers, by some estimates, had worked fewer hours and eaten a more varied diet than the early farmers who replaced them. The revolution made the group richer and most of its members less free. That is the template. Keep it in mind, because we are going to see it again and again in fancier clothes.

The press that Rome tried to close

No technology has a better liberation résumé than the printing press. Before it, copying a book meant a scribe and a year; a library was a form of aristocracy. After it, a single workshop could turn out in a week what a monastery produced in a lifetime. The obvious result — knowledge for the many, the end of the gatekeepers — is the version we all learned.

The real early history is messier, and I have written about it at length in the printing press. The first people to grasp what movable type could do were not peasants suddenly reading for themselves. They were the powers already positioned to exploit a machine that manufactured identical copies at scale. The Church printed indulgences by the tens of thousands — reportedly among the earliest and most profitable products to come off the new presses. Monarchs printed proclamations. The most efficient copying machine ever built was, in its first decades, a superb instrument for whoever already owned a message and wanted more of it in more hands.

The headline says liberation. The clause underneath says: for whoever controls the tool.

The broadening came later, and it did not come from the machine. It came from what people did with it and around it — the licensing regimes that were fought and eventually broken, the vernacular Bibles that could not be un-printed once they existed, the pamphlet wars, the slow construction of a reading public that authorities could no longer fully police. Freedom of the press is a legal and political achievement layered on top of a technical one. The press made mass copying possible. Law, conflict, and organising decided whose copies reached whom. Leave the second part out and you have mistaken the tool for the outcome.

The engine and the people who owned it

The steam engine multiplied human strength beyond anything muscle or water or wind had offered. It should read as pure emancipation: less back-breaking labour, more goods, cheaper everything. In the very long run, industrial economies did lift living standards for ordinary people to heights no previous era imagined. That part is real and worth saying plainly.

But look at the decades the people who actually lived through it endured, not the century-scale average that consoles us afterward. The first-order effect of mechanised production was to concentrate it. Factories demanded capital, and capital belonged to a few. The independent weaver working at home, setting his own hours and owning his own output, was pulled into a mill where the machine set the pace and someone else owned the machine. Wages by some accounts stagnated for a generation or more even as output soared. Children went into the factories. The gains were real; they simply flowed, at first, almost entirely to the owners of the new tools.

This is the context in which the machine-breakers acted, and they were not the mindless reactionaries of the insult their name became — I make that case in the Luddites. They did not misunderstand the technology. They understood it precisely: that a specific machine, deployed by a specific owner under specific terms, was going to transfer their livelihood upward, and that the promised general prosperity was, for them and their children, a promise about a future they would not live to see. They were arguing about distribution, not progress. And distribution is exactly the question the word freedom is so good at making us forget.

The network that was supposed to belong to everyone

Then the internet, whose liberation pitch was the most sweeping of all. This time it was not one gatekeeper being toppled but all of them at once. Anyone could publish. Anyone could sell. Anyone could reach anyone, with no editor, no distributor, no boss standing in the middle taking a cut. For a brief window in the early days it even felt true. The web was a scatter of odd personal sites, and no one owned the middle because there was barely a middle to own.

We know how that resolved. The open protocols are still there, technically, but the traffic runs through a handful of platforms that sit exactly where the old gatekeepers used to sit — except larger, because their reach is planetary and their marginal cost of adding one more user is almost nothing. The tool that abolished the intermediary produced the most powerful intermediaries in human history. The freedom to publish turned out to be real; the freedom to be found was quietly rented back to us by whoever owned the feed. That is not a bug in the internet. It is the plough and the press and the engine, running once more: a tool that lowers the cost of doing something valuable will, left to itself, hand the surplus to whoever controls the chokepoint through which the value flows.

This is the mechanism I keep returning to, and it is worth naming directly, because it is the argument underneath everything here: new technology does not distribute power on its own. It concentrates power at the point of control, and it stays concentrated until law, organising, and sometimes plain conflict pry it open — the process I lay out in how technology gets captured. The liberation we remember is the second act. The first act is capture, and the first act is the one that happens automatically.

The pitch that skips the question

Notice what the word freedom does in all of these stories. It answers a question about capability — what can now be done? — and it uses that answer to skip an entirely different question: who gets it, on what terms, and who decides? The plough could feed everyone. The question was who held the granary key. The press could inform everyone. The question was who held the licence. The engine could enrich everyone. The question was who owned the mill. Every time, the sales pitch is framed as a story about what the tool makes possible, precisely so that we never get around to the story about distribution — which is the only story that determines whether the freedom reaches you or stops one rung above you.

The liberation we remember is the second act. The first act is capture, and it is the one that happens on its own.

I do not think this is usually a conspiracy. It is simpler than that, and more durable. Whoever is positioned to control a new tool has every incentive to sell it in the language of universal freedom, because that language disarms exactly the people who would otherwise ask about terms. And the rest of us are willing customers for the pitch, because the alternative — believing that a wondrous new thing will mostly benefit those who already have the most — feels like cynicism, and we would rather feel hope. So the small print goes unread, every single time, for the same human reasons.

Our own moment, and the small print in it

Which brings us to now, and to the technology whose liberation pitch is currently the loudest in the room. Artificial intelligence arrives promising to democratise expertise itself: everyone gets a tutor, a lawyer, a doctor, a research assistant, a creative collaborator, for the price of a subscription or for free. Read that sentence with the whole history above sitting behind it and the shape of the pitch is unmistakable. It is the plough's promise and the press's promise and the internet's promise, delivered in a new register.

And it may well come true — partly, eventually, in the second act. But ask the distribution questions now, while they still have force. The largest models cost sums to train that only a few organisations on earth can afford. They run on hardware controlled by a handful of firms. The data they learned from was, in large part, produced by the very people whose work they may displace. If the pattern holds — and I have given you four revolutions' worth of reasons to think it will — the first-order effect of AI will not be to hand expertise to the many. It will be to concentrate an unprecedented amount of capability in whoever owns the models, the compute, and the chokepoints, and to rent that capability back to everyone else on terms those owners set.

That is not an argument against the technology, any more than the Luddites' real grievance was an argument against looms. It is an argument for reading the contract. The freedom AI promises is genuinely possible. Whether it arrives for you, or stops one rung above you, will be decided the way it has always been decided: not by the tool, but by the fight over who controls it and on what terms — the boring, essential work of law and organising and public pressure that turns a concentrated capability into a widely shared one.

So when the next revolution knocks — and there is always a next one — open the door, by all means. The tools are marvellous and I would not give a single one of them back. But when it hands you the contract with freedom printed in enormous letters across the top, do the one thing the history above should have taught us to do without fail. Turn the page over. Read the small print. The whole of your actual freedom is written down there, in the clause about who gets to control the thing that was supposed to set us free.