I have spent the last few years building software that replaces people in some way. I’ve watched demos where a person’s decades of experience were compressed into a prompt. It seemed like it was inevitable. Everyone around me thought it was inevitable. We said things like “progress” and “efficiency” and “democratization” and moved on to the next sprint.
Here is what I have come to believe: the only inevitable things are the things we actually work on. We are building the machinery of human irrelevance and calling it progress, and almost nobody is working on what comes after.
The Three Things We Offer
For most of human history, people have contributed to society and to each other in three ways:
Hands. Physical labor. The farmer, the builder, the plumber, the electrician. The body moving through the world, manipulating matter into something useful. This is the oldest form of human contribution and it built civilization.
Minds. Cognitive labor. The analyst, the writer, the programmer, the strategist. Post-industrialization, the developed world migrated here. We stopped making things with our hands and started making things with our thoughts. This was supposed to be the permanent upgrade — the knowledge economy, safe from automation because machines could never think.
Connection. Relational labor. The doctor’s bedside manner, the teacher who sees a struggling child, the therapist who sits with your grief. This is usually bundled with the other two and rarely valued on its own. We pay nurses, caregivers, and counselors poverty wages precisely because their contribution is relational rather than transactional.
Every economic system ever designed — capitalism, socialism, communism — assumes that people perform some combination of these three in exchange for the means to live. You offer your hands, your mind, or your presence, and society offers you a place in it. This is the barter system. It is older than money. It is the substrate beneath every political arrangement humans have attempted.
AI and robotics are breaking it.
The Squeeze
The argument for hands was settled decades ago. Manufacturing, agriculture, logistics — machines do these faster, cheaper, and with fewer errors than human bodies. The world adjusted, painfully. Workers moved from farms to factories to offices. There was somewhere to go.
The argument for minds is being settled now. LLMs write legal briefs, diagnose medical images, generate code, analyze markets, and compose music. Not perfectly. Not yet as well as the best humans in every domain. But at a fraction of the cost and at a speed that makes “good enough” devastating. A 2025 MIT study found that AI can currently replace roughly 12% of the US workforce outright. Over the last few months that number is probably 3x and it’s only moving in one direction.
The whole cage giving off a horrible stench — the stench of terrified, feathered flesh. The roosters in the coop smell the blood from above. They see the organs of their brothers lying around them. They know they’re next. Yet they do not rebel.
— Aravind Adiga, The White Tiger
The optimist’s retreat is always to the third category. “AI can’t replace human connection,” they say. “People will always want a real doctor, a real teacher, a real therapist.” This is true in the same way that people still want handmade furniture: as a luxury, for the few who can afford it and who care. A shrinking market.
The Speed of the Fall
Previous economic transitions gave us time. The shift from agriculture to industry took a century. The shift from industry to knowledge work took fifty years. Human institutions — schools, legal systems, social norms, identity formation — adapted across generations. Slowly and painfully, but they adapted.
AI capability is doubling roughly every six months. Not linearly improving. Doubling.
Generative AI may create a socioeconomic tipping point through labour displacement — not a gradual decline but a sudden phase transition where consumer spending, tax revenue, and social stability all break simultaneously.
— Occhipinti, Hynes et al., Nature Scientific Reports (2025)
This is the detail that makes every historical analogy — “we survived the Industrial Revolution, we’ll survive this” — break down. The Industrial Revolution gave displaced weavers decades to become factory workers, and their children decades to become clerks. AI is not offering decades. Maybe it’s offering years. The models that couldn’t write a coherent paragraph in 2022 are now producing work that passes bar exams, medical licensing tests, and creative writing evaluations. The structural reasons that prevent every job from being replaced tomorrow are not technical — they are regulatory, institutional, and psychological. They are speed bumps, not walls. And the vehicle is accelerating.
Bigger Than Capitalism
Most commentary on this topic frames it as a crisis of capitalism. It is not. Or rather, it is, but that framing is too small.
Capitalism requires consumers with purchasing power. If AI eliminates wages, consumers cannot buy, and the system collapses. Gilles Saint-Paul at the Paris School of Economics modeled this formally in 2025: the middle class shrinks, demand for mass-produced goods collapses, and even capital owners are harmed because there is nobody left to sell to. The oligarchs preside over a shrinking economy. Capital eats itself.
Capitalism’s drive toward efficiency — replacing expensive labor with cheap machines — is also a drive toward destroying its own consumer base. AI is capitalism being too good at being capitalism.
But swap capitalism for any other system. Socialism redistributes the fruits of labor — but what if there is no labor to fruit? Communism holds the means of production in common — but who operates them, and for what purpose, when machines operate themselves? Every political economy humanity has designed is a set of rules for distributing the proceeds of human effort. Remove human effort from the equation and you don’t just break capitalism. You break the barter.
This is not the end of one economic system. It is the end of the assumption that has underlain all of them: that humans produce and humans consume, and the circuit between the two is what makes a society.
The Meaning Problem
This is where most analysis stops and where I believe the real crisis begins.
The economists worry about income. Give people Universal Basic Income, they say, and the consumption circuit is preserved. Perhaps. Let us grant that UBI or some equivalent redistribution mechanism solves the material problem. People have money. They can buy things. The economic identity holds.
What do they do on Monday morning?
A man has to BE something; he has to matter.
— Hunter S. Thompson
The uncomfortable truth that the “age of abundance” crowd does not address is that for most people, meaning and purpose are inseparable from contribution. Not from consumption. Not from leisure. From the feeling that what you do matters to someone, that your hands or your mind or your presence changed something in the world that would not have changed without you.
Strip that away and you do not get paradise. You get a species-wide existential crisis.
We have run this experiment before, in small pockets. Communities where work disappears — post-industrial towns, regions dependent on a single employer that closes — do not become leisure societies. They become sick. Depression, addiction, family breakdown, deaths of despair. Not because people are poor, though they often are. Because they are useless. Because the thing they organized their identity around — “I am a miner,” “I am a steelworker,” “I am the person who does this thing” — evaporated, and nothing filled the void.
Elon Musk says AI will bring an age of abundance for all. He may be right about the abundance. He is catastrophically wrong about the “for all.” Abundance without agency is a zoo. The animals are fed. They do not hunt. They pace.
Bernie Sanders says slow it down, regulate it, protect the jobs. This is the other failure mode — the belief that you can hold back the tide with legislation. You cannot. The economic pressure is too great. A company that replaces its workforce with AI achieves a cost advantage that no regulation-compliant competitor can match. The incentives are locked in. We are past the point where “stop” is a meaningful instruction.
Neither answer addresses the real problem because neither answer acknowledges what the real problem is. The real problem is not economic. It is existential. What is a human being for, in a world that no longer requires human beings to function?
A Banner Without Bearers
I don’t have a complete answer. I want to be honest about that because the essay you are reading was written in part to articulate the question clearly enough that an answer becomes possible.
What I do know is this: no one is working on it. Not seriously. The technologists are building AI as fast as they can because competition demands it and because the engineering problems are genuinely thrilling. I’ve never seen a group of people run as fast as they are to their own destruction. The politicians are either technophobic or technoutopian. The economists are modeling within existing frameworks that assume the very thing being dissolved.
There is a voice missing from this conversation. A practical, urgent, honest voice that says: we are building ourselves into irrelevance at an unprecedented speed, we do not have a plan for what comes after, and the clock is running.
I am picking up that banner. Not because I have the answer but because I helped create the problem. I celebrated those efficiency gains. I told myself it was inevitable, as if inevitability were something that happens to us rather than something we build with our own hands.
The only inevitable things are the things we actually work on.
If we do not work on the problem of human relevance — what people are for, how they contribute, where they find meaning in a world that no longer needs their contribution — then it will not be solved. It will not solve itself. The market will not solve it. Technology will not solve it. It will become the defining crisis of our century, arriving not in some distant future but in the years immediately ahead of us, and we will have built it with our own tools while telling ourselves it was progress.
I don’t know what the solution looks like yet. But I know that finding it requires people who are willing to stare at the problem without flinching, without retreating into utopian fantasy or Luddite denial. People who understand that this is not about jobs. It is not about capitalism. It is about what it means to be human when the things we do can be done without us.
The abyss is there. The question is whether we enter it on our terms or fall in.
One day you finally knew
what you had to do, and began,
though the voices around you
kept shouting
their bad advice—
though the whole house
began to tremble
and you felt the old tug
at your ankles.
“Mend my life!”
each voice cried.
But you didn’t stop.
You knew what you had to do,
though the wind pried
with its stiff fingers
at the very foundations,
though their melancholy
was terrible.
It was already late
enough, and a wild night,
and the road full of fallen
branches and stones.
But little by little,
as you left their voices behind,
the stars began to burn
through the sheets of clouds,
and there was a new voice
which you slowly
recognized as your own,
that kept you company
as you strode deeper and deeper
into the world,
determined to do
the only thing you could do—
determined to save
the only life you could save.
