When the Economy Stops Needing Us: What If We Were Never the Main Story? 9

A look at the constants and variables in the coming shift in work.

Something about our prosperity doesn't feel very prosperous anymore.

It takes two incomes to maintain a lifestyle that one income used to support. The house costs more hours of work than it did for your parents, and so does the degree, and so does the retirement that keeps moving further out. We are richer than any people in history by the official measures, and yet the experience on the ground is one of running faster to stay in place: more credentials required, more debt carried, more of the week spoken for, and a particular dread that arrives on Sunday evenings and has become so common we joke about it.

For the generation just entering working life, the dissonance is sharper still. Reports suggest recent college graduates are unemployed at higher rates than the workforce as a whole—an inversion of the entire premise on which they were sold the degree. More than four in ten of those who do find work are in jobs that didn't require the degree at all. The degree itself arrives with an average of roughly $40,000 in federal student debt, and research on student debt and homeownership finds that every additional $1,000 in student loans measurably lowers the odds of ever owning a home. The sequence that defined middle-class adulthood—degree, job, car, house, family—has stalled for millions of young people at the second step. They did everything the story told them to do, and yet the story is not paying out.

I want to suggest that this dissonance is not in your head, and that it's worth sitting with for a minute before we talk about artificial intelligence—because the AI conversation everyone is having is built on an assumption we need to examine.

How New This Arrangement Is

Here is a fact that surprises most people: the way we live—selling our hours to organizations, organizing our identities around our jobs, structuring life as school-then-career-then-retirement—is about two hundred years old. As a mass arrangement, it barely existed before industrialization. For most of human history, the idea of spending your life working on a stranger's schedule, at a stranger's task, for a stranger's purposes, would have seemed strange at best and degrading at worst. In the 1860s, "wage slavery" was not a radical's phrase; it was ordinary vocabulary, used by mainstream newspapers and politicians to describe an arrangement that many Americans considered a temporary station on the way to independence—a farm, a shop, a trade of one's own.

Within two generations, that view vanished. The temporary station became the destination. The first question we ask a stranger became "What do you do?"—meaning, what is your job?—and we stopped noticing that this is a peculiar way to ask who someone is.

What happened in between was not a debate that wage labor won. What happened was that an industrial system with an enormous appetite for human labor built the institutions that would feed it—most importantly, mass compulsory schooling, which trained children in the punctuality, task-compliance, and tolerance for tedium that factories required. Over time, the system's requirements came to feel like life itself. Work for pay. Give your loyalty to a commercial organization. Live for the weekend. Retire when you're used up. These are not human universals. They are the operating requirements of a particular machine, experienced from the inside as simply the way things are.

I'm not saying the arrangement was a swindle. It paid. That's the part we need to look at squarely.

The Deal Underneath the Arrangement

Every protection and comfort that came to define modern working life—the weekend, the pension, the safety regulations, public education, the vote itself in its expanded form—was obtained the same way: it was purchased with leverage. Factories needed hands. Armies needed bodies. Strikes could actually stop production. The system needed its people, massively and continuously, and that need is what made the people impossible to ignore.

It's tempting to read the last century and a half as a story of moral progress—civilization maturing, rights expanding, dignity winning. The less flattering and more accurate reading is that it was a bull market in human capacity. The rising floor under ordinary life was neither a gift nor an achievement of conscience; it was a price paid for something the system was buying in enormous quantities. Workers were never the point of the machinery. Workers were the fuel, and fuel, while it's needed, gets handled carefully.

This is hard to see precisely because the institutions that managed the arrangement told a different story, in which our work served our flourishing, our careers expressed our identities, and the system existed for us. The gap between the story an institution tells and the function it actually performs is, I've argued elsewhere, the single most useful lens for understanding how institutions work. Apply that lens here and the picture reorganizes: the story was that the economy served human beings; the function was that human beings powered the economy. The stories of school, career, family wage, and retirement were the maintenance schedule for the energy source.

This raises the question that the current moment forces: what happens to the fuel when the engine finds something cheaper to burn?

Why This Matters Now

This is no longer a thought experiment. According to one outplacement firm, artificial intelligence became the leading reason American employers cited for job cuts this year, with more AI-attributed layoffs in the first five months of 2026 than in the previous two years combined. Many of the companies making the largest cuts are reporting record profits and directing savings toward AI investment. Whether AI is the primary driver or a convenient explanation in some cases, the public acceptability of framing layoffs this way is itself telling. A story is changing in real time, and you can watch it change in the earnings calls.

The bottom rung is being sawed off the ladder. Entry-level white-collar work, the traditional intake valve of the whole system, appears especially exposed. Reports indicate employment of young software developers has dropped by roughly a fifth in recent years. The young people locked out of the housing market by debt are now facing new barriers to the income that was supposed to service it.

Artificial intelligence is usually discussed as a story about us: our jobs, our incomes, our futures, what we will do, how we will be retrained, how we will be made whole. Notice the assumption: that the system has some continuing obligation to solve the human side of the equation. That assumption made sense for two hundred years, because for two hundred years, the system needed us. The unsettling possibility is that the obligation was never an obligation at all. It was a purchase agreement. And the buyer may be leaving the market.

I don't claim to know how this plays out, and this piece is not an argument for any particular outcome. What I want to do instead is something I think might be useful at the front edge of a large, inevitable change: lay out the elements at play. What's fixed, what's variable, and what historical cases we can calibrate against. A map, not a verdict.

The Constants

Start with what does not change: human psychology, which was shaped over a very long time and will not be updated on the machinery's schedule. Whatever arrangements emerge on the other side of this transition, they will be evaluated by us and by history against a short list of needs that every durable human culture has had to satisfy: coalitional safety, status and relevance, shared narrative, consequence, and—for any culture that intends to exist in three generations—procreation.

Industrial work, for all its extractions, bundled several of these together. The job was where many people found their coalition, status, narrative, and consequence. That bundling is worth naming, because the loss of employment is never merely a loss of income. It is the withdrawal of an entire delivery system for psychological necessities—and the question of what replaces that delivery system is separate from, and larger than, the question of what replaces the paycheck.

The Master Variable: What AI Turns Out to Be, Economically

Nearly everything downstream depends on a question that sounds technical but isn't: whether revenue from artificial intelligence can be captured.

Oil made certain futures possible because oil is scarce, ownable, and sellable at a margin—it generates rents (income from a resource or production), and rents can fund things, including the pacification of populations the system no longer needs. There is clearly a bet being placed right now, visible in the staggering scale of the American buildout, that AI is the new oil.

But there's another possibility: that AI is the new air. If machine intelligence drives the price of cognition toward zero (and that is its visible trajectory), then it may prove enormously valuable and nearly impossible to charge for, with its margins competed away and its moats breached. The rapid rise of open-source models and the ability to run capable LLMs on personal computers tilts me toward this direction; the fences look increasingly hard to maintain when the technology itself wants to spread. Watch the fights over compute access, licensing, regulation, and proprietary data in the coming years; they might best be understood as attempts to build fences around something before it becomes a commons. Whether those fences hold is perhaps the single most consequential open question, because it determines whether there is a revenue stream large enough to fund whatever comes next. (It may not be binary—hybrid outcomes are most likely—but the direction matters enormously.)

There is also a circularity in the bet that is worth noting plainly. The current buildout is capital selling to capital on the promise of future demand. But if the deployment succeeds in replacing labor income, it erodes the consumers who were ultimately to drive that demand. The system is, in effect, borrowing against a customer it is in the process of firing.

The Actors, by Position Rather Than Identity

It helps to see each group not by its label but by its position in the loop—what it supplies, what leverage it holds, and what claim it has on the rents, if rents materialize.

• White-collar workers are, for the first time in the history of mechanization, exposed first. This is an inversion that scrambles every existing political coalition and every parent's advice. 
• Blue-collar and local trades are insulated by physics and trust, but only against substitution; they remain exposed to the second-order effect of a collapse in local demand. 
• Young men deserve their own recognition here: high coalitional energy, the steepest decline in supplied relevance, and the best-documented track record in history of what happens when both go unanswered. 
• Government workers are insulated by politics rather than productivity, and public employment may quietly grow as a disguised dividend. 
• Incumbent professionals and gatekeepers—the credential guilds, the publishing organizations now fighting for training-data compensation and building systems to track actual human authorship, the schools policing AI use—will fight industry by industry, and the useful diagnostic in each fight is whether what's being defended is a function or a story. 
• The tech elite hold the rent claims; finance is leveraged on those claims paying off; politicians stand between the rents and the legitimacy that the rents will be asked to buy. 
• And off to the side, almost never mentioned in the AI discourse, are the high-fertility insular communities, like the Amish, who never sold their leverage in the first place and are quietly compounding while everyone else debates.

The Distribution Variables

If rents do materialize, the next question is distribution. Oil-funded states like the Gulf monarchies pay their citizens well through stipends, subsidies, and guaranteed positions, because oil revenue flows through the state by default. What citizens of such states do not get is power, because a government that doesn't need its people's labor or taxes doesn't develop accountability to them. This is provision without leverage: comfortable, and politically inert.

AI rents, by contrast, flow to private balance sheets, in political cultures with widely varying appetites for redistribution. So even where the money is enormous, the pipe from rents to dividends has to be built—and in some countries it will be built against organized resistance, fought line by line. The fiscal capacity question sits underneath all of it: sovereign debt loads are already heavy, and the futures that depend on funded dividends depend on treasuries that can fund them.

The Time Variables

The speed and shape of the transition may matter more than its endpoint. A sudden displacement, like mass layoffs concentrated in a year or two, would produce a shared narrative, a common identity among the displaced, and therefore coordinated political leverage. A slow erosion over twenty years produces none of that: each cohort is displaced separately, told individually to adapt and reskill, and the structural story never crystallizes.

There is also institutional lag: narratives decay more slowly than functions. Schooling is an eighteen-year pipeline that will keep solemnly funneling children toward careers whose existence nobody can promise, because institutions cannot update their stories faster than a generation. The children entering kindergarten this fall graduate into the mid-2040s.

And there is the possibility of rupture—perhaps a financial crash, maybe even triggered by the AI bet itself failing to pay. A crash would delete the comfortable futures from the menu by simple arithmetic, but it would also starve the displacement engine of capital, as the 1930s starved mechanization. Crisis doesn't choose between good and bad adaptations.

The Possible Outcomes

Put the constants, the variables, and the laws together, and what emerges is not a single future but a set of potentials into which populations can settle—several of which are already visible somewhere in the world.

• Funded spectatorship: The Saudi model—provision without leverage, comfortable and consequence-free.
• Unfunded spectatorship: The post-Soviet model, also visible in the American Rust Belt, with its signature mortality data—the drinking, the opioids, the morose dependency of people whose usefulness was repealed without replacement.
• Make-work without consequence: The late-Soviet variant, where everyone has a job, and nobody has a reason.
• Coalitional violence: The oldest absorber of surplus young men.
• Patronage: Relevance re-personalized as service to wealthy households and figures.
• The engagement economy: Simulated coalition, simulated status, simulated consequence, delivered cheaply and at infinite scale to people whose evolved psychology cannot fully distinguish the simulation from the real thing—a basin that is comfortable, profitable for its operators, and reproductively sterile.
• Community reconstruction: Arrangements in which a person's usefulness is local, visible, and non-substitutable. The Amish are the standing existence proof—not because of buggies or piety, but because they retained the full stack of production, coupled to the modern economy selectively rather than totally, and never put their necessity up for sale.

What to Watch

I'm not arguing a conclusion, but it's worth looking at the map.

Watch whether the fences around AI hold, whether it becomes oil or air. Watch the ratio between the economy that needs mass consumers and the economy that no longer does. Watch the policy language, for the moment when "what about the displaced workers" quietly becomes "what about social stability." That shift in vocabulary is the operative relation surfacing. Watch each industry's defensive fight, and ask the diagnostic question: defending function, or story? And watch what gets built, locally and on purpose, in the meantime.

We spent two hundred years believing we were the main story. It is possible we were the energy source, and that the most important question of the coming decades is not what the machinery will do for us when it no longer needs us, but what we are prepared to do independently of that.