Who Will Buy the World the Machine Builds?

In the summer of 1934, my grandfather watched a city stop. The longshoremen of San Francisco had laid down their hooks, the teamsters had parked their trucks, and one by one the trades had walked off the job until the whole engine of commerce sat cold on its blocks. Writing as Mike Quin, the rank-and-file journalist, he set the scene in a single sentence that has stayed with me my entire life.

Read that again and notice what it actually proves. The wheels did not move because no hand moved them. The levers did not budge because the men who pulled them had folded their arms. For all the steel and steam of an industrial port, the machinery turned out to be inert — a vast sculpture of iron that did precisely nothing until a human being decided it should. The strike was not a protest against the machine. It was a demonstration of who the machine had always belonged to.

I keep returning to that scene because the men who own today’s machines are making the opposite wager. They are betting that the wheel can be made to turn with the hand removed entirely — and they are no longer shy about saying so.

They Have Told Us Their Plans

This is not speculation, and it is not the nervous imagining of people who fear change. It is the stated business plan of the most powerful firms on earth, delivered in interviews and earnings calls with the calm of men describing the weather.

In February of this year, the chief executive of Microsoft’s artificial-intelligence division told the Financial Times that the technology would soon reach human-level performance on most professional tasks — and that the work of lawyers, accountants, project managers, and marketers, the work of “sitting down at a computer,” could be largely automated inside twelve to eighteen months. He is not alone in the chorus. The head of one leading AI lab warned that as many as half of all entry-level white-collar jobs could vanish; the chief executive of a storied American carmaker said the same technology could halve white-collar employment outright. They differ on the timeline. They agree on the direction.

And the ledger has already begun to fill. This is no longer a forecast; it is a count.

One of the men keeping that ledger put the matter more honestly than the executives do. “Regardless of whether individual jobs are being replaced by AI,” the firm’s analyst observed, “the money for those roles is.” That is the sentence to underline. The work may or may not disappear. The wage is what they are coming for.

My grandfather knew that arithmetic in his bones. His first pamphlet, in 1933, was titled And We Are Millions — a collection of testimony from homeless young men jailed for vagrancy in the depths of the Depression, the human discard pile of an economy that had decided it had no use for them. He understood that when capital treats labor purely as a cost, the logical end of cost-cutting is a person standing in the road with nowhere to go. We have a more polished vocabulary now. We say “efficiency.” We say “rightsizing.” We say the model is “exceeding human baseline.” The man in the road is the same man.

A Customer Is Also Somebody’s Worker

Here is the question the efficiency men keep forgetting to ask, and it is the oldest one in the marketplace: if you fire everyone, who is left to buy the goods?

The story is told — half legend, half documented — of Henry Ford II walking the union leader Walter Reuther through a newly automated engine plant. Gesturing at the robots, Ford needled him: how are you going to get those to pay union dues? Reuther is said to have answered without missing a beat: how are you going to get them to buy your cars?

You can build the automobiles, Reuther told a Senate subcommittee in 1958, but the consumers are still made the good old-fashioned way. Walter Reuther · Senate testimony, recounting the plant tour

The line is funny until you sit with it, and then it is terrifying, because it exposes the marketplace for what it actually is: not a one-way chute that runs from factory to buyer, but a circle. Wages are the river. They leave the firm as a cost on Monday and return to it as demand on Saturday, when the worker spends his paycheck on somebody else’s product — and that somebody’s wages return as demand for yours. The elder Henry Ford understood this when he doubled his workers’ pay to five dollars a day. He was not being generous. He wanted men who could afford to buy the cars they were building. He was watering his own river.

The dream now on offer is a factory with no payroll feeding a market with no customers. Each firm, acting with perfect rationality, cuts its own labor to win the quarter. But my labor is your customer, and your labor is mine, and when every firm cuts at once we discover — too late — that we have been sawing off the branch we are all sitting on. Efficiency, pursued without limit, becomes a solvent. It dissolves cost, then it dissolves the wage, then it dissolves the wage-earner, and finally it dissolves the demand that the whole enterprise was built to capture. A perfectly efficient economy that has automated away its consumers is not a triumph. It is a beautifully optimized corpse.

This is the heart of the Dangerous Thoughts thesis, and it cuts clean across the usual political lines: a system can run exactly as designed and still grind down the ordinary people it depends on, not out of malice, but out of the relentless internal logic of squeezing one more point of margin. Nobody has to be a villain. The machine of incentives does the harm on its own.

But the Future Is Not Yet Written

I owe you honesty, because my grandfather was a propagandist when the cause demanded it and a reporter when the truth did, and the second man is the one worth being. So here is the other half of the ledger.

The displacement, so far, is real but uneven. One independent study found that AI tools actually slowed some experienced software developers even as they felt faster. The Yale Budget Lab, tracking the data through the spring, found no significant shift in occupational unemployment despite tens of thousands of headline cuts. Economists keep pointing to the old paradox named for Jevons: when a thing becomes cheaper, we often use vastly more of it, and demand for the cheapened skill can explode rather than vanish. Even the lab chief who once predicted the white-collar bloodbath has lately walked it back — “delighted to be wrong,” he said — and reframed automation as a multiplier: automate ninety percent of a job and the remaining ten percent expands to fill the day, ten times more productive than before.

And the workers themselves are not passive. By one account, four in five refuse the adoption mandates handed down from above. The strike instinct — the folded arms, the cold lever — is not extinct.

I report all this not to soften the warning but to sharpen the point: none of this is fated. The technology is a hammer. A hammer builds a house or breaks a skull depending entirely on the hand that swings it. Which means the only question that finally matters is the one no earnings call will answer for us — whose hand?

A Tool in the Worker’s Hand

In 1954, the same Walter Reuther wrote something that the doom-merchants on every side conveniently forget. Guided well, he argued, automation could “bring health and happiness, security and leisure, and peace and freedom to mankind everywhere.” The union did not fear the machine. It feared the machine in the wrong hands. It wanted the wheel to turn and the worker to share in what the turning produced.

That is the road this publication chooses, and it is the road I believe my grandfather would recognize. The fight was never against the tool. Mike Quin called himself a rank-and-file journalist — a man whose entire craft was putting a printing press, a megaphone, a voice, into the hands of people the powerful preferred to keep silent. He spent his life lowering the cost of being heard. Ask yourself what such a man would do with a tool that can draft, translate, research, summarize, and explain — not for the boardroom, but for the rank and file.

Because the very same technology that a corporation uses to erase a hundred workers can be turned around and placed in the hands of one. Consider what that hand can suddenly do:

The freelancer gains the back office of a firm — the bookkeeping, the contracts, the proposals — without surrendering a cut to anyone. The small shop competes with the giant on knowledge instead of being buried by it. The worker who was told for a century that the law was too complicated to understand can now ask, in plain language, what her rights are, what her contract says, what her employer owes her. The organizer can reach a thousand co-workers in the time it once took to reach ten. Every person who was ever told they could not write, could not analyze, could not build — that gate is coming off its hinges. A tool that lowers the cost of capability does not have to concentrate power. It can distribute it. That is a choice, and it is ours to make.

More Human, Not Less

There is a deeper inheritance here than economics, and it points to the question hiding underneath all the others: what is work even for?

More than a century ago, Charlotte Perkins Gilman argued in Human Work that labor is not merely how we earn our bread — it is how we become ourselves and how we are bound to one another. Work, in her account, is the very medium of human development: the place where we grow our capacities, prove our usefulness, and weave ourselves into the social fabric. Strip a person of meaningful work and you do not simply lower their income; you sever the thread that ties them to the human project.

If Gilman was right — and everything I have learned about flow, mastery, and the slow craft of becoming a fully integrated human tells me she was — then the goal of automation should never be to abolish work. It should be to abolish drudgery, and to free human beings for the work that actually grows them: the creative, the relational, the deeply skilled, the things a machine can imitate but never inhabit. Let the model do the toil. Let the human do the becoming. That is the promise worth fighting for, and it is the precise opposite of the promise being sold to us — which is that the human becomes the part you cut.

An economy that uses these tools to make a few people unimaginably rich and the rest unnecessary has failed the Gilman test completely. An economy that uses them to remove the soul-deadening parts of work, so that more people can do work that develops them and connects them — that economy is using the machine to make us more human, not less. Same tool. Different hand. Different world.

Which Direction Will We Choose?

So we arrive where my grandfather always arrived: at a fork that ordinary people, organized, get to decide.

Down one road, artificial intelligence is a machine for concentration. Fewer owners, more discards, a marketplace efficient enough to have automated away the customers it was built to serve — the optimized corpse, gleaming and bankrupt. Down the other road, the same invention is a machine for distribution: capability poured into more hands, dignity restored to more work, the drudgery lifted, the river of wages kept flowing, the rank and file handed a press, a back office, a voice, a fighting chance.

Nothing in the silicon decides between those two roads. The technology is indifferent; it will serve a sweatshop or a cooperative with equal fluency. The decision is political, and moral, and entirely human — which is exactly the kind of decision the powerful would prefer we believe has already been made for us, by forces too large and too inevitable to resist.

It hasn’t been. In 1934 the wheels stopped because the people who turned them remembered, all at once, that the machine was theirs. The lesson did not expire. The levers of this new age are no different. They will move in whatever direction we, the rank and file, decide to push them — and they will not budge an inch on their own.