I Hope AI Takes Every Job

I want AI to take your job.

I want it to take mine too.

Imagine I hand you $100 million tomorrow morning. No strings, no catch, no conditions.

Do you show up to work on Monday?

I know my answer. I might keep some of the work. There are problems I love solving and things I still want to build. But I would stop organizing my life around what someone else is willing to pay me to do.

For a lot of us, work isn't the thing we want. It's the toll we pay to get the things we want.

So what happens when the toll booth disappears?

That question gets lost in the conversation about AI and jobs. We talk about protecting employment as though employment and a worthwhile life were the same thing. Sometimes they overlap beautifully. Sometimes a job is how someone gets to do the thing they care about most.

Sometimes it's how they pay rent.

I don't want anyone to lose their livelihood. I want us to reach a point where having a livelihood no longer requires selling most of our time.

The question is whether we can get there without leaving people behind along the way.

What money is actually for

Most conversations about AI stop at jobs. AI will take jobs. Robots will replace workers. In some industries, it's already happening.

But there's a deeper question underneath. What if machines eventually do all jobs?

Not the routine ones. Not the ones we condescendingly call low skill. All of them. Construction workers, doctors, engineers, CEOs, and the people building the machines themselves.

At first, that sounds dystopian. If nobody has a job, nobody has income. If nobody has income, nobody has money.

That's exactly the point.

What is money actually for?

Zoom out far enough and the economy is goods and services. To create them, we combine land, materials, machines, and the one ingredient only people could supply. Specialized labor.

For most of human history, people did nearly everything themselves. Then specialization took over. One person raised chickens, another made blankets, another built tools, and specialization needed a way to trade across time and strangers.

That's the job money does.

Money is stored labor. Stored time. Stored human effort. You do something useful for other people, and society hands you tokens representing that contribution. You carry them into the future and exchange them for someone else's effort. Every dollar in your account is a claim on another human being's time.

Ray Dalio puts the loop simply. One person's spending is another person's income. I manage construction projects. I get paid. I spend it at the grocery store, the gas station, a week in the mountains. My spending becomes their income, and they spend it somewhere else. It looks like an economy, but underneath it's people passing each other's time around in circles.

The part everyone skips

So run the experiment properly.

What happens if AI and robotics can do every job? Not just physical labor. Research, diagnosis, design, management, and the machines that build the machines.

Humans can still do all kinds of things. The system just no longer needs our labor to keep producing.

If machines do all the labor, why would humans need money?

The obvious questions have obvious answers.

Who builds the machines? The machines.

Who maintains the infrastructure? The machines.

Who grows the food? The machines.

I'm talking about a fully automated system, A to Z. If you've watched the last three years, that's no longer a wild premise. It's a timeline question.

The real problem is this.

The people who own the machines will keep everything.

That's not paranoia. Left alone, it's the default outcome. A world capable of feeding everyone could still leave people hungry.

Let's start with the part that goes wrong first.

The economy breaks before it gets better

The most dangerous part of this transition is the period when human labor has become less valuable but everything people need still has a price.

It doesn't begin with abundance. It begins with deflation, and our financial system is deeply fragile under deflation.

Modern economies don't just tolerate rising prices. They require them. Central banks target inflation deliberately. Pensions, mortgages, government budgets, and corporate balance sheets are all built on the assumption that next year's numbers are bigger than this year's.

Technology does the opposite. It makes things cheaper, relentlessly. And it isn't only AI. Robotics, cheap compute, biotech, and energy are arriving together and compounding on each other, the largest deflationary force in human history. Every one of them drives the cost of producing things toward zero.

Which means the thing we're most afraid of and the thing we most want are the same phenomenon. Falling prices are what abundance looks like from inside a monetary system.

The problem is that debt doesn't fall with them.

The savings also arrive unevenly. A company can cut its costs without cutting its prices. Cheap manufactured goods do little for someone whose biggest expense is rent on scarce land. If the price of groceries falls but my income falls faster, the mortgage payment stays the mortgage payment.

Irving Fisher described what happens next in 1933. Prices fall. Debts stay exactly where they are, so the real burden of every debt quietly rises. Borrowers cut spending to keep up. That spending was someone else's income, so incomes fall, prices fall further, and the debt gets heavier again.

Now layer on automation. It doesn't just make goods cheaper. It removes the incomes on the other side of the ledger. When the first person's income disappears, the second person's income goes with it. The debt stays, untouched, indifferent. The Institute of International Finance put global debt at a record of roughly $353 trillion in early 2026, sitting on a system that only works if nominal values keep expanding.

The early tremors are already here, and the full break is closer than most people are ready for.

There are two kinds of deflation. The first comes from getting better at making things. Across much of the late 1800s in the United States, prices fell for decades while real wages rose. People got richer. That's the good kind. The second is debt collapse deflation. The 1930s. Japan's lost decades. Demand craters, debt calcifies, the economy seizes.

Technological deflation is the good kind. But arriving alongside collapsing wages and record debt, it can set off the bad one. Central banks, public spending, and debt restructuring can soften the blow. None of them happen automatically.

It won't arrive evenly either. Software can rewrite an office workflow years before a robot can reliably work a construction site. Customer support, translation, and entry level coding are already feeling it. So one neighbor's profession vanishes while the neighbor across the street hasn't noticed anything yet. The first one is watching an industry learn to function without him. The second assumes he simply failed to adapt.

That gap is where resentment lives. The scapegoating, the strongmen, the identity crises. Not because abundance failed. Because we arrived at abundance still running an operating system designed for scarcity.

Strauss and Howe's Fourth Turning is useful here as a lens, not a prophecy. They argue institutional orders get torn down and rebuilt on roughly eighty to a hundred year cycles, and the tearing down always feels terminal from inside it. Feudalism felt permanent. The gold standard felt permanent. The British Empire felt permanent. Every system feels eternal from the inside, right up until the day it isn't.

We need a way for people to share in the gains before we ask them to trust the destination.

The forcing function

My optimism here isn't idealism. It comes from incentives.

An economy that pays zero wages loses its customers, unless it finds another way to put money in their hands. The owners of the machines are only wealthy if someone can buy what the machines make. Infinite production with no demand isn't power. It's a warehouse full of widgets.

At civilizational scale, that's the pressure that bends capital toward distribution. Not a scheme to take from the rich and hand it to the poor. Not owners suddenly becoming kind. Business owners will figure it out the way they always do. By watching their revenue disappear.

The incentives point the right direction, but not reliably. Each company still benefits from cutting its own wages while hoping someone else keeps customers spending. And the pressure has an expiration date. It only works while owners still need revenue. Once machines can supply everything they want directly, customers stop being necessary.

The bridge has to be built while the incentive still exists.

Ownership, not allowance

Everyone reaches for Universal Basic Income. I think UBI is the weak answer.

A basic income can make you a dependent. Your survival becomes a line item in someone else's budget, and line items get cut, means tested, or made conditional. Conditions are where the danger lives. A monthly payment everyone depends on is the most powerful behavioral lever ever built. Tie it to compliance, to speech, to a social score, and you have a mechanism for control at a scale no government in history has held.

It's a leash with a stipend attached.

That isn't hypothetical. When Henry Ford introduced the five dollar day in 1914, mostly to stop workers from quitting his assembly lines, the raise came with company investigators checking whether employees were living the way the company approved.

The stronger answer, at least in the near term, is ownership. If machines produce everything, the question was never how we redistribute the output. It's who owns the machines.

Norway is the proof this model works. Instead of spending its oil revenue, it invested it. Its sovereign fund now holds, on average, about one and a half percent of all the world's listed companies, roughly $385,000 per citizen on paper, and it funds more than a quarter of the national budget.

Alaska made it personal. Since 1982, nearly every resident has received an annual dividend from the state's oil wealth, and researchers found no decline in people working. But calling it a dividend didn't make it safe. In 2016, the governor cut the payment roughly in half to cover a budget gap, and the state Supreme Court upheld the cut. A fund is only as strong as the rights around it.

Now apply that shape to automated capital. Not a tax on robots. A sovereign equity stake in them. Citizens become shareholders in the automated economy and receive dividends as owners.

The hard part is how the public gets its stake. One route is taxing the largest companies partly in shares rather than cash. Another is public investment in exchange for equity. Another is ownership stakes negotiated when companies receive major public support. Norway and Alaska started with public claims on resources in the ground. Nobody holds an equivalent claim on privately owned AI labs or robot factories. We would have to create one.

The best name for this direction is one I first came across from Yat Siu. Universal Basic Equity. Not income handed down from above. Ownership built into the structure of the economy itself.

A benefit is something you receive. Ownership is something you hold, and done right, it comes with rights that are hard to take away. An owned fund can still pay a universal income. Alaska's dividend is essentially one. What matters is whether the payment is a right or a favor. Ownership done well creates sovereignty. Ownership done badly is a benefit with better branding.

None of this is politically easy. Every entrenched interest fights it. But the pressure to adopt something like it rises with every quarter that productivity climbs while employment doesn't.

Dividends are the bridge, not the destination. They're money, and they matter enormously, right up until the goods they buy cost almost nothing.

Money doesn't get abolished. It gets outlived.

The real constraint was never human time

There's a physical floor under all of this. A robot doesn't eat a salary, but it still consumes energy, takes up space, wears out, and depends on materials.

All production is energy converted into a different arrangement of matter. Human labor is energy conversion too. Calories in, work out, roughly a hundred watts sustained. Every cathedral, every pyramid, every mile of railroad was built by chaining together enough hundred watt biological machines for long enough.

Machines convert energy without a biological ceiling. Under full automation, the binding constraint on civilization stops being human time. It becomes energy and matter.

So this whole thesis rests on one question. Does energy become abundant?

I think the incentives make it close to inevitable. Once labor stops being an input cost, energy becomes the dominant one left, and whoever controls automated production has overwhelming pressure to drive it toward zero. That's a stronger force than any climate policy or subsidy has ever been.

The trend is already running. IRENA reports that the cost of electricity from new utility-scale solar fell about 90 percent between 2010 and 2024. Storage is following. Elon Musk has argued for years that solar plus batteries is the answer, and that the real limit is how fast we can manufacture and deploy.

It isn't automatic. Mining, grid capacity, permitting, and land all take real time, and intelligence alone solves none of them. If energy doesn't become abundant, none of this works. You get a new aristocracy of whoever controls the joules. Feudalism with better hardware.

Energy is only half the floor. The other half is atoms. Every panel, battery, chip, and robot is built from matter, and some of it is rare. Copper. Lithium. Rare earths. Even compute sits downstream of both. In a world built on functional abundance, the metals that actually do something may end up mattering more than the one we've spent five thousand years admiring.

Ecological limits matter too. A system that produces disposable goods faster than the planet can absorb the waste hasn't solved abundance in any sense worth having. Human wants may never stop growing, but ordinary needs are different. I can only eat so much food. A good home doesn't need to double in size every year. An excellent education doesn't get worse because someone else receives one too.

We are not eliminating scarcity. We're relocating the constraint. Away from human time, which is hard capped, non transferable, and painfully finite, toward energy and matter, which have a ceiling higher than anything our species has ever operated against.

That's not the end of limits. It's a limit far enough away that it stops governing daily life.

What money loses

If machines do everything and energy is abundant, does money disappear?

I think eventually, yes. Not by ban, decree, or revolution. It loses its referent.

Money's defining property was never that it allocates scarce goods. It's that it's storable, transferable, and convertible into control over other people's time. Money is a claim on human labor. When no human labor is required, the claim has nothing left to claim.

Money also buys things no one's labor created. Land. A location. An ownership right. So the end of labor doesn't finish money off by itself. It shrinks its territory. Money can survive the end of wages, astonishingly cheap production, even abundance, as long as someone can charge for access.

It becomes unnecessary when access no longer needs to be bought.

Picture a system comfortably supplying what people actually use. Food, housing, transport, education, manufactured goods. Access is secure, and more demand can usually be met without taking something important from someone else.

Why keep distributing money so people can hand it back?

Today, when I pay for dinner, somebody on the other side of that transaction gave up their evening to make it. My money is a claim on their time. When nobody gave up anything to make the meal, the payment has no one on the other end. Now picture that across nearly everything you buy.

That takes three things. Production that meets demand. Access that can't be arbitrarily withheld. And other ways of allocating what stays scarce. Where those hold, I struggle to see what useful work money is still doing.

History has small versions of this. In 1780s Paris, water carriers hauled buckets from public fountains up apartment stairs, thirty deliveries a day, charging by the floor. Water still has a price. But once pipes arrived, the labor vanished and the cost of a bucket collapsed to almost nothing. It wasn't abolished by ideology. It was made obsolete by technology. The need for carriers disappeared, and most of the price went with it.

Food, housing, transport, medicine, energy, and education could follow the same path, one at a time. The price of almost anything is mostly the human time and energy it took to make it, raw materials included. Take human time out, drive energy toward zero, and what's left is the rarity of the matter itself.

Nobody has to announce the last transaction. The territory where payment is useful shrinks until there's nothing important left for it to do.

Who gets the house on Miami Beach

Some things stay scarce no matter how much intelligence you throw at them. Beachfront land. Original art. A specific person's attention. No amount of AI manufactures a second Malibu.

So without money, who gets the beach house?

The obvious answer is some kind of mechanism. A lottery. A rotation. A waitlist. I've never found any of those satisfying, and I don't think the mechanism is the interesting part.

The question contains a hidden assumption.

Right now, money is a universal converter. One score, every game. Win at money and you don't just win at housing. You win at healthcare, legal outcomes, your kids' education, political access, how long you live.

That's what makes scarcity feel unjust. Not that beachfront is limited. That whoever gets the beachfront also gets the better surgeon.

Now take the survival goods out of the market. The beach house stays scarce, but the scarcity stops leaking. Whoever gets it gets a house near the water. Their kid doesn't get a better school because of it. They can't easily turn it into political influence, and they can't use it to command anyone's time.

Think about front row concert seats. Someone else gets them and you don't, and nobody calls that a civilizational injustice, because sitting front row doesn't also buy a better cardiologist.

That's the shift. Scarcity goes from general to local.

The philosopher Michael Walzer argued a version of this decades ago, that an advantage in one sphere of life shouldn't let anyone dominate the rest. What abundance adds is a mechanism. The fewer essential goods we have to ration, the fewer places one advantage can spread.

Once scarcity is local, the question stops being a moral crisis and becomes a question of rules. The rules still matter, but allocating without money isn't a fantasy. Matching systems already assign students to public schools in New York and Boston, and kidney exchanges match donors with patients in a world where buying organs is illegal.

The real danger is trading financial power for administrative power. Whoever controls the queue becomes the new gatekeeper. Any system worth wanting needs transparent rules, appeals, and protection against favoritism. AI can help coordinate it. It can't decide what fairness means for us.

Inheritance works the same way. You'd still pass your house to your kids. What ends is the inheritance of claims on other people's labor. Inherited wealth today is inherited command over human lives, and that's a very different thing from inheriting a building.

None of this makes effort meaningless. The floor rises for everyone. Nobody goes hungry, nobody goes without a doctor, nobody's child gets a worse education because their parents were poor.

Effort doesn't stop mattering. It just stops being the price of staying alive.

Humans will still compete over the things that stay rare, the way we always have, and for most of history that competition wasn't denominated in money anyway. What ends is scarcity in one domain determining your outcomes in all the others.

That's not the elimination of scarcity. It's the separation of scarcity from human subjugation.

The part where this could go badly

The risk isn't that the machines take over. That's the science fiction version.

The real risk is that abundance arrives and access to it gets gated. A world capable of producing everything, with a paywall around it, enforced by whoever owns the machines and no longer needs the rest of us. Not as workers. Not as soldiers. Not even as customers.

Technology determines whether abundance is possible. Politics determines whether it's shared. Only the first one solves itself.

The ownership question is the fork in the road. Everything optimistic in this essay sits downstream of it. I'm not predicting the good version. I'm saying it's available, and that we're sleepwalking past the decision that determines which one we get.

Who do we become

Which leaves the hardest question.

If nobody works, where does meaning come from?

Underneath that question sits an assumption so deep we've stopped seeing it. That human worth comes from economic output. For most of history, to survive was to work. So work became identity, identity became meaning, and meaning got tied to output. We've held them together so long we mistake them for the same thing.

I reject the idea that humans would lose meaning if jobs disappeared. Humans are meaning making machines. We make it from relationships, art, exploration, competition, curiosity, and learning, and most of all from experiences that make us feel something. We were doing it for tens of thousands of years before anyone drew an org chart.

Go back to the $100 million. Most people aren't working primarily for meaning. They're working because survival requires it. Gallup's global surveys find only about one in five employees engaged in their jobs. Meaning gets attached to the structure afterward, because we're built to find meaning in whatever we're already doing.

Some people would keep going. Their work is the meaning. Artists would still make art. Builders would still build. Scientists would still push at the edge of what's known.

Creativity is one of the most human things there is.

But a job gives even the unengaged something real. A reason to get up. People who expect something of you. The experience of being needed. If machines can do almost everything better, one question gets harder to answer.

What do other people need me for?

More leisure doesn't answer that. But an activity doesn't become meaningless because something else can do it better. I don't climb a mountain because the world needs a more efficient way to get me to the summit. The effort is the point. And being economically replaceable doesn't make you interchangeable to the people who love you.

The machines I'm describing don't want anything. We do, and our wants never run out. Hand me five robots that can do everything, and within a week I'd be inventing new things for them to build, some specific want of mine no manufacturer could have predicted.

Abundance doesn't end striving. It ends compulsion.

The difference isn't whether people do things. It's whether they chose them.

Freedom from employment won't automatically produce the habits, communities, and commitments that make freedom worth having. Building those may be the most important work left to us.

The end of one phase

Maybe humanity has spent its entire existence optimizing for survival, and we're approaching the first moment we could optimize for something else. Every civilization we've built, every philosophy we've written, every identity we've inherited was designed by people whose baseline condition was scarcity. We have no idea what we are without it.

The transition will be disruptive, likely more than anything in living memory, and there will be a stretch where the pessimists look right. But a painful transition doesn't mean a bad destination. Most people alive today couldn't survive the world their great grandparents considered normal, and wouldn't want to. But we adapt. It's the single most reliable thing about us.

So maybe the future where AI takes every job isn't the collapse of humanity. Maybe it's the end of one phase of civilization and the beginning of another. A phase where people spend their time creating, learning, building, exploring, connecting, and being present. Not because they're forced to. Because they want to. Because they choose to.

Honestly? That sounds like a world I'd rather live in.

I think it's worth considering that the future we're most afraid of might be the one that frees us.

It's hard to picture a world without money, and that shouldn't be surprising. A fish can't see water. It has spent its whole life inside it, so the water isn't an environment. It's reality itself. Money is our water. None of us has lived a single day outside it. The difficulty of imagining its absence was never evidence the idea is wrong. It's evidence of how deep we're in.

The thesis underneath this whole essay is simple. If machines do all the jobs, humans no longer need money to live. Eventually, I don't think we'll need it at all. The economics is hard. The harder part is what comes after. Deciding how we actually want to spend our time, and building a shared vision of the future worth spending it in.

And maybe the real question was never what humans will do when AI takes the jobs.

Maybe the real question is this.

Who do we become when survival is no longer the primary game?