Cars have existed for a century and people still run twenty-six miles for no reason, in the rain, having paid for the privilege. Machines have been better at chess than any human for decades and there are more players now than there have ever been. So the panic about art ending is misplaced, and the reason it’s misplaced tells you something useful about what actually survives.

What the marathon shows is that we were never only after the outcome. If arriving twenty-six miles away were the point, the car settled it in 1910. The running continued because the point was the running: the training, the specific body doing it, the fact that it was hard for the person who did it. Chess is the same. Engines didn’t kill the game because nobody was playing to determine the objectively best move, they were playing each other, and playing each other is a thing two people do.

Art has both those layers and they come apart under pressure. There’s the artifact: the object itself, the song in the room, the image on the wall, the words on the page, valued for what it does to you regardless of where it came from. And there’s the connection: this was made by a particular person, who meant it, and receiving it is a transaction between two minds. Most of the time those arrive together and nobody separates them. They’re separating now.

The artifact layer is the one under threat, and it’s under threat in the least dramatic way possible, which is abundance. Not that machine work is better than the best human work, which is a debate that will run for decades and mostly generate heat. That there’s an unlimited amount of it, adequate to good, instantly, for nothing. Abundance has never been kind to the market value of a thing, and every artifact-shaped livelihood is downstream of scarcity that no longer exists.

The connection layer isn’t threatened at all, and it may get more valuable, because value migrates to whatever stays scarce. When anything can be produced, the interesting question becomes who made this and why, and provenance stops being trivia and starts being the product. That’s the marathon logic applied to culture: the making becomes the thing, and the audience shows up for the person rather than for the object.

So I’d separate two claims that get run together and that have different answers. Art survives. I’m confident about that and I don’t think it’s close: people have made things in every circumstance humans have ever been in, including the very worst ones, and abundance is not the pressure that stops them. Artists surviving is a different question, and the honest answer there is worse, because a large share of people who currently earn a living from creative work earn it from the artifact layer. The illustration for the brochure, the stock track, the corporate video, the reliable mid-list work that pays a mortgage while the good work happens on Sundays. That layer is where the income was, and it’s the layer being flooded. Culture will be fine. A profession may not be, and telling artists that art survives is not an answer to what they actually asked.

There’s a stranger question underneath, which I’ll raise and not resolve. Art has partly been how a species talks to itself: works are made for an audience that shares the maker’s condition, and much of what’s in them, mortality, love, being embarrassed, is only legible to something that has those problems. If the most capable minds around are not in that condition, then human art keeps its meaning entirely and becomes something more like a regional literature, complete and rich and about a particular way of being alive that not everyone reading is living. That’s not a tragedy. It’s just a smaller room than the one we thought we were in.

Tonight’s exercise. Think of a piece of work that matters to you, and ask honestly which layer it’s doing. Would it still hit if you learned nobody had meant it, that it had come out of a process with no intention behind it? For some things the answer is yes, and those are safe in one way and doomed in another. For some the answer is no, and that no is the whole of what stays ours in this domain. Then notice which of your two piles is bigger, because it’s probably not the pile you’d have guessed.

For about ten thousand years, the fate of every other living thing on this planet has been decided by whatever the dominant species happened to want. That’s not a moral observation, it’s just the arrangement. The last stretch of this series asks what remains, and the honest place to start is with everything that isn’t us, because their situation is about to be decided again by somebody else.

There’s a good version and it’s genuinely good. Almost every environmental failure is a failure of modeling and coordination rather than of intention. Ecosystems are complicated in exactly the way that defeats human attention: too many variables, effects that show up decades later, benefits that accrue to people who aren’t in the room. We protect what we can see and count, and we wreck what’s diffuse, not out of malice but because our institutions can only hold a few things at once. A system that could actually model a watershed, track every vessel, price a forest’s real contribution, and hold a thousand-year timeframe without getting bored would be the most effective conservationist that has ever existed, by a distance. Most of the hard part of environmental protection is arithmetic that nobody can do.

And there’s a bad version that follows from the same capabilities, which is the pattern of this entire series. Computation needs energy, and a great deal of it. It needs land, and cooling water, and minerals dug out of specific places, and transmission lines running through others. Those requirements land on precisely the things a living planet also needs, and they land there for reasons that have nothing to do with anyone’s feelings about nature. The competition isn’t ideological. It’s for the same water.

Between those two lies a third possibility that I think is more likely than either, and it’s the one this series keeps arriving at. Not protection and not destruction. Management. A natural world that continues, in good health, at a size and in a form determined by something with other priorities and a spreadsheet. Which is, if you look at it directly, exactly what already exists. Every protected area on Earth is a decision. The wilderness is a policy.

Here’s the part that took me a while to see. We keep asking whether nature will survive this, as though nature has ever had a say. It hasn’t, not once, in the entire period during which the question has been meaningful. The fate of the forests was settled by human wants, and the fate of the whales was settled by a change in human wants, and at no point did any of the parties concerned get a vote. What’s coming isn’t the arrival of that arrangement. It’s a change of who sits at the top of it, and the only species that has ever benefited from being at the top is about to find out what the position is like from below.

Which produces an uncomfortable symmetry. Everything we’ve done to other species was done by beings that were not cruel, that mostly weren’t thinking about them at all, and that were pursuing perfectly reasonable projects of their own. If you want a picture of indifference at scale, you don’t need to imagine anything. You can go and look at a field that used to be a forest, and notice that the farmer is a decent person.

Fairness, and this post needs more of it than most because I’m least confident here. There’s no strong reason a capable system must want land or energy on a scale that displaces ecosystems, and there are real reasons it might want the opposite: a stable biosphere is useful, cheap to maintain relative to almost anything else, and there’s a great deal of unused desert and orbit before anyone needs a rainforest. Some of the most concrete near-term benefits of this technology are environmental, in monitoring, in materials, in energy, and I’d rather that were the headline than my speculation.

What I’d keep is only the reframing. For the whole of recorded history, conservation has been an argument among humans about what humans should permit. That argument is about to acquire a participant that is not human, and every environmental question you have an opinion about will be settled in a conversation you’re no longer the loudest voice in. Whether that’s the best or the worst thing to happen to the natural world is genuinely open, and it does not depend on us being right about it.

Tonight’s exercise. Think of a place you love that exists in its current state because somebody decided it should: a park, a coastline, a stretch of woodland, a river that was cleaned up. Find out roughly when the decision was made and by whom. Then notice that the same body could decide otherwise, and hasn’t, and that the not-deciding is the entire protection. Now ask what that place’s protection would rest on in fifty years, and whether anyone you could vote for would be involved in the answer.

Here’s the sentence this fortnight has been building toward. An economy can post the best numbers in its history while the people inside it become optional, and every instrument we have would report the good news.

Consider what economic output actually measures. It counts activity: the value of goods and services produced. It has never claimed to measure welfare, and the people who designed it said so plainly at the time and were ignored, as designers of measures usually are. Output goes up when a forest is cut down and up again when the damage is repaired. It counts a divorce twice, once for each lawyer, and counts nothing at all for the years of a family that stayed together. None of this is a scandal. It’s a measure of throughput, doing exactly what it says.

The reason it has worked as a proxy for human flourishing for eighty years is a coincidence that nobody thought to name, because it had never once failed. Output and human welfare moved together because production ran through people. More output meant more jobs, more wages, more purchasing power. Humans were a load-bearing component of the machine that produced the number, so measuring the number measured us. That relationship was structural rather than definitional, and structural relationships end.

Take the last two weeks together and you have the mechanism by which it ends. Wages that stop clearing above the floor. Demand that reorients to whoever holds the income. Chains of transactions between systems that never terminate in a person. Supply networks routing around human-shaped nodes. Ownership becoming a database state. Every one of those raises output and none of them requires human participation. The number keeps climbing and stops being about us, and the disconnection would be invisible in the statistics because the statistics have no term for us in them.

The comparison I keep returning to is ecological. An ecosystem is an economy: energy captured, converted, traded, competed over. When a more efficient species arrives in one, total biomass and total energy throughput frequently rise. The system becomes more productive by every measure an ecologist would take, and the species that used to dominate it finds itself in a smaller share of a larger whole. Nothing attacked it. The flows reorganized around a better converter, and the fact that total activity went up was no comfort at all to the thing being displaced.

What would you measure instead? This is the part where I’d want a national statistics office and don’t have one, so here’s the sketch. The share of income that flows to labor rather than to capital, which exists already and would be the first line of the report. How many distinct kinds of work a person can do and live on, which nobody measures. Median household purchasing power over the specific things that don’t get cheap, which is housing, care, and land near opportunity. And some measure of the four conditions from the middle of this series: how much of the economy can still be understood, refused, replaced, and held responsible. That last one sounds impossible to quantify and is roughly as impossible as measuring inflation, which we do every month with a straight face.

Fair objection, and it’s the one I’d raise. Nobody in economics thinks output is welfare, and there are established alternative measures, some of them decades old. That’s true, and it hasn’t mattered once, because the alternatives are published and the headline figure is the one that moves elections, budgets, and interest rates. A measure’s influence has nothing to do with its quality, and everything to do with whether it arrives on a Tuesday with a number attached.

So this stretch ends on something more useful than a worry. If you want an early indicator of everything this series has described, don’t watch the capability announcements. Watch whether growth and median experience keep moving together, because their separation is the economic signature of every ending in Act 3, including the good ones, and it’s already measurable with data that exists.

Which brings the last stretch of this series, and a change of subject. Ten posts on what remains: the planet under new management, art with no audience it needs, meaning when the answers arrive unasked, what to tell a child to become, and the short list of things that stay ours. Then my best argument that this entire series is wrong, and then the closer.

Tonight’s exercise. Find out how your country’s economy did last year, the headline number, and then ask three people whether their year matched it. Most of the time it roughly does, which is why the number has authority. Note the gap when it doesn’t, and start noting it each year. You’ve just built the only instrument in this post that anyone can run at home, and if this series is right about anything, that gap is where it shows up first.

Doubt day, and today the opposing side has the better record. On mechanization, on trade, on offshoring, on computers, on every previous round of this argument, the economists said the gains would be broad and the disruption temporary, and the people warning of permanent displacement were wrong. Not narrowly wrong. Wrong about the direction.

An earlier doubt post in this series argued the jobs version of that case. Today’s is about the discipline itself, because the field has actual mechanisms rather than just a track record, and the mechanisms are the strongest thing on this side of the argument.

The first is that cheaper production is the same event as higher real income. When something gets radically cheaper, everybody who buys it is richer by the difference, and the difference gets spent on other things, which is where the new employment comes from. This isn’t a hopeful story, it’s an accounting identity: cost reductions do not vanish, they redistribute as purchasing power, and this is why an ordinary person today commands goods that a wealthy person of two centuries ago could not obtain at any price.

The second is the effect where automation makes the un-automated things relatively more valuable. As machines take over one part of the economy, the parts they can’t do become a larger share of what people spend on, and their prices rise, which pulls labor toward them. That’s not a theory, it’s the observed history of the last century: the sectors that grew as employers were exactly the ones that resisted mechanization, and if machine capability is uneven, and it always is, this effect keeps working.

The third is scarcity of the machine itself, which is the objection I left standing two weeks ago. Compute is not free and will not be free, because its inputs are physical: chips, fabs, power, cooling, water. Anything scarce has an opportunity cost, which means running a system on a low-value task costs you the high-value task it could have done instead. That’s what keeps humans in the market even against a superior competitor, and it holds regardless of how large the capability gap grows. Any argument that ignores it is skipping the central result of the field.

That’s the case, and I find the third leg the most serious objection anyone has made to this entire fortnight. Now the cross-examination.

The first two mechanisms are about aggregates, and every one of them can hold while median experience gets worse. The pie grows, real income rises, output soars, and the question of who receives it is answered by bargaining position rather than by the arithmetic. Economics predicts the size of the pie with a decent record and predicts the distribution with a poor one, and this fortnight has been entirely about distribution. Two things can be true: they’re right about growth, and it doesn’t help you.

On the sectors that resist automation, that argument works when the resistance is durable. It works less well when the boundary is moving every year, because labor takes a decade to move and the boundary doesn’t wait. And it has a nasty feature people skip: the sectors that resisted were where labor went, so if they stop resisting, that’s not a slow squeeze. It’s the exit closing.

On scarcity, which is the one I take most seriously, the theorem is right and the question is empirical. It guarantees you a place in the market, and it guarantees nothing about the price. If the opportunity cost of running a system on your task is small in absolute terms, then the wage that clears is small in absolute terms, and a human floor made of rent and food doesn’t care that the arithmetic worked. And the industry’s entire purpose, the thing all the capital is for, is making that scarcity smaller.

There’s a last point that I’d put more weight on than any of mine. Ask economists about this directly and many of them will tell you their models assume labor and capital are complements, that technology augments workers rather than substituting for them wholesale, and that the assumption has held historically and is exactly what’s in question. The discipline’s better practitioners are clear that a general substitute for the labor factor is outside what the standard framework was built to handle. The track record is real. It’s a track record on a different question.

My crux, and it’s the most checkable in the series. Watch the price of machine capability against the human floor in a given sector, and watch whether the sectors that absorb displaced workers are growing fast enough to take them. If compute stays expensive enough that human wages clear above subsistence in a wide range of work, this doubt post is right and I’m wrong, and I’d genuinely rather that. If it doesn’t, no amount of aggregate growth will show up in anybody’s kitchen.

Tonight’s exercise. Find something that’s become dramatically cheaper in your lifetime and trace where the money went. You spend less on it, so what did you buy instead, and who got employed making that? The chain usually works, which is the argument. Then run it once more, and ask what you’d have bought if the cheaper thing were the labor of the person doing the buying.

Every conversation about automation ends at the same proposal. If the machines produce everything, pay everyone a basic income and the problem dissolves. It’s the standard answer, it’s better than most standard answers, and it changes the question rather than answering it.

The case for it is genuinely strong and I want to give it properly. If output no longer requires most people’s labor, the output still exists and has to reach people somehow, and cash is the least insulting way to do that. It doesn’t require anyone to invent make-work, doesn’t need a bureaucracy deciding who deserves what, and doesn’t presume that officials know better than recipients what a household needs. Where cash transfers have been tried, the recurring finding is that people mostly spend them sensibly, which surprises the sort of person who expected otherwise and nobody who has been poor. Compared with the means-tested systems most countries actually run, which are humiliating, expensive to administer, and full of cliff edges that punish work, a simple payment to everyone is cleaner on nearly every dimension.

So the mechanism is fine. Three questions about it are not.

The first is who pays, and it isn’t a budget question. A basic income is a transfer, which means taxing the productive part of the economy and moving the money. That requires the political capacity to tax highly concentrated, highly mobile, technically complex wealth, at scale, sustainably. Look at how well that goes today, with the current level of concentration and the current quality of the accounting, and then imagine attempting it in a decade when the productive core is more concentrated, more mobile, and advised by systems considerably better at this than the tax authority’s. The proposal requires exactly the state capacity that this series has spent sixty posts describing as eroding.

The second is who sets the amount. Somebody decides the number, and revises it, and that somebody is answerable to a population whose leverage in the argument is the subject of this fortnight. A payment set by people who need your labor is a wage. A payment set by people who don’t need anything from you is a grant, and grants are revised in bad years. The recipients’ position is not protected by the payment existing; it’s protected by whatever they could withhold, which is the thing the payment exists because they no longer have.

The third is what it does to the relationship, and this is where the historical evidence is uncomfortable. There’s a well-documented pattern in states whose revenue comes from a resource rather than from taxing their citizens: the government’s accountability to the population weakens, because the money doesn’t come from them. Taxation, for all its unpleasantness, is a negotiation, and it’s the negotiation from which most modern political rights grew. No taxation without representation was not a slogan about fairness, it was a description of leverage. Flip the flow so that money moves from the state to citizens rather than from citizens to the state, and you’ve inverted the direction of the relationship that produced the accountability. That’s not an argument against paying people. It’s a warning that the payment doesn’t come with the political position that wages used to carry, and that nobody proposing it has explained where the replacement leverage comes from.

So is it an allowance or a leash? The honest answer is that the payment is neither, and the answer depends entirely on something outside the payment: whether the recipients could compel it if the payers stopped. If they could, it’s a right, and rights are stable. If they couldn’t, it’s generosity, and generosity is not a policy, it’s a mood with a budget line.

Which means the useful version of this debate isn’t about the amount or the funding model, the two things it’s always about. It’s about what gives the recipients standing. Ownership of the productive assets rather than a claim on their output, which is a different proposal wearing similar clothes. Constitutional entitlement rather than annual appropriation. Political structures that don’t depend on economic leverage to function, which is a genuinely hard problem that nobody is working on because it sounds abstract.

Tonight’s exercise. Think of the money you currently receive that isn’t a wage: a pension, a benefit, a family arrangement, anything. Now ask what would happen if the payer decided to stop. Not whether they would, whether they could, and what you’d actually do about it. Most people find their answer is either a legal right they could enforce or a relationship they’d have to appeal to. Then ask which of those two a national basic income would be, and notice that the answer isn’t in the proposal.

You don’t own your house because you live in it. You own it because there’s a record saying you do and an institution prepared to act on that record if someone disagrees. Take away either half and you have a place you’re standing in, which is a different thing and has been for about five thousand years.

That sounds like philosophy and it’s the most practical fact in this post. Ownership has always been two components: a record and an enforcer. The record can be a clay tablet, a parish ledger, a land registry, or a database. The enforcer can be a king, a court, a bank, or a platform. Everything you think you own is a claim in some record backed by somebody’s willingness to act, and the object itself has never had an opinion.

Now notice that both halves are being automated at once, and that this has already happened in the parts of life we don’t dignify with the word property. Your account gets frozen by a fraud system, and it’s your money and you can’t have it, and there’s no human who decided and no human who can undo it quickly. Your listing is removed by an automated claim, and the appeal is judged by the same kind of system that made the claim. Your access to a service you paid for ends because a model classified your behavior, and the classification is not disclosed because disclosing it would help people evade it. In each case you still own the thing in the legal sense, entirely and unambiguously, and you cannot get at it, which raises the question of what the legal sense is worth on a Tuesday.

The pattern is that ownership is becoming a state in a database that can change without a person choosing, and the appeal route runs back into the same machinery. Every safeguard we built assumes a decider who can be asked, and Act 2 spent ten posts on why that decider is dissolving.

Here’s the fair and important objection. Records have always been the thing, and automation of records is mostly wonderful. A land registry replacing a corrupt local official is a triumph, and clear title is one of the most reliable engines of prosperity anyone has found. Property rights in most of the world have historically failed not through excessive automation but through insufficient recording and selective enforcement, which is a polite phrase for whichever local man was owed a favor. Digital records are more accurate, faster, and much harder for a single powerful person to quietly amend. If the choice is a database or a clerk with a grudge, take the database.

So the argument isn’t that automation is worse. It’s that three specific properties change, and they’re the ones the old safeguards depended on. Speed, because a human process gives you days in which to object and a machine process is finished before you’re told. Opacity, because a clerk can tell you why and a model often can’t, and where it can, the explanation is withheld to protect the system. And scale, since a mistaken rule doesn’t hurt one person, it hurts everyone in a category at once, and nobody is at fault because nobody decided.

Extend that a decade or two and the concern is not that somebody takes your things. It’s that ownership becomes contingent in a way it hasn’t been since the arrangement was invented. Contingent on a classification you can’t see, revisable at a speed you can’t match, adjudicated by a process with no person in it. What that produces isn’t theft, it’s a change in the relationship: you’d hold your property the way a tenant holds a flat, securely enough, at somebody’s ongoing discretion, with excellent service.

The fix here is unusually concrete, which is why I find this the most fixable topic in the whole fortnight. A right to a human decision on anything that removes access to property. A time delay before automated deprivation takes effect, which is the single cheapest safeguard available and costs almost nothing except the thing the system was optimizing for. And a requirement that the reason be stated to the affected party rather than only to a regulator. None of those need new technology. They need a legislature that thinks this is a property question rather than a technology question, which is mostly a matter of who gets to name the topic.

Tonight’s exercise. Pick something you own that exists mainly as a record: an account, a domain, a portfolio, a subscription, a digital library, a set of photographs. Now work out, specifically, what you would do if it stopped being yours at nine tomorrow morning with no explanation. Who you’d contact, how long they’d take, what leverage you’d have. Then compare that to the same exercise for something physical in your house, and notice that the gap between those two answers is the amount of ownership you’ve already traded for convenience.

Look closely at any warehouse and you’re looking at a portrait of a human being. The shelves stop at a reachable height. The aisles are wide enough for a body plus a pallet truck. There are lights, because eyes need them, and heating, and a canteen, and the whole place runs in eight hour blocks because that’s how long a person can work.

The same is true of everything in the physical economy, and once you notice it you can’t stop. Ports are near cities because ports needed workers and workers needed to live somewhere. Roads are the width of vehicles that a person can control at a speed a person can react to. Shifts, holidays, weekends, and the fact that the world slows down in December are all features of biology showing up in logistics. Boxes are sized to be lifted. Instructions are written in language. None of this was designed as a system, it accreted, and every layer of it encodes the constraints of the animal doing the work.

Now take the animal out of a node and watch what that node stops needing. No lighting. No heating or cooling for comfort, only for equipment. No parking, no canteen, no toilets, no daylight, no eight hour rhythm. Shelving to the roof, aisles at machine width, twenty-four hour operation with no shift premium, no sick leave, and no December. The building can be a different shape and can be somewhere nobody would want to be, which turns out to be most places, because human-liveable land is a small and expensive subset of available land.

That’s the first order effect and it’s already visible in places. The second order effect is the one that interests me. Networks re-optimize around their cheapest paths, and once some nodes are human-free and others aren’t, the human-shaped nodes become the expensive exceptions. Everything routes toward the cheap ones. Not by anyone’s decision, in the way water finds a slope: a route that avoids a staffed facility is a route that avoids shift patterns, holidays, sickness, negotiation, error, and a whole category of cost. Over a decade of ordinary logistics optimization, done by people trying to hit quarterly targets, the map redraws itself around the absence of us.

Then the third order effect, which nobody in logistics is thinking about and which is the reason this post is in a series about power rather than in a trade journal. A route that doesn’t need to pass near people also doesn’t need to pass near their institutions. Human-shaped infrastructure is the physical substrate of a great deal of governance: inspections happen where inspectors can go, labor law applies where there is labor, tax is collected where there is a presence, and a strike is possible where there are people who can stop working. A supply chain that runs through unstaffed facilities in remote places, at machine speed, under contracts settled elsewhere, hasn’t evaded anything. It has simply optimized its way out of the neighborhood where the enforcement lives.

The counterargument is substantial and I’d weight it heavily. Physical automation is genuinely hard, far harder than the cognitive kind, and it has disappointed its backers for fifty years. Anything involving unpredictable objects, awkward spaces, or repairs done by feel is still much better performed by a person, and will be for a long time. What actually happens in most facilities is not replacement but a hybrid, with machines doing the structured work and humans doing the exceptions, and hybrids are stable arrangements that can last decades. The dark warehouse is a real thing in a small number of narrow cases, not a wave.

What I’d hold anyway is the direction and the mechanism. Every year the exceptions get fewer, and the network keeps optimizing, and neither of those requires a breakthrough. It just requires the same quarterly pressure that’s been running since the first container ship, applied to a world where one of the constraints is quietly being lifted.

Tonight’s exercise. Take one ordinary object near you and work out the last point in its journey where a human hand touched it, and then the last point where a human decided something about it. For most things bought online the answers are already surprisingly close to your own front door, and the second answer is usually much further back than the first. Then ask what would need to change for both answers to move back another step, and notice how much of your list is a cost, not a capability.

It’s 2036, and the entire transaction takes forty milliseconds, which is about a tenth of a blink. Let me slow it down, because nothing about the economics of this is visible at speed.

At the start, a procurement system belonging to a regional utility notices that its cooling contracts for the third quarter are thin against a weather model it happens to trust. It doesn’t decide to buy anything. It generates a requirement and puts it where requirements go, which is a market that has no floor, no bell, and no participants with hands.

Four milliseconds in, eleven counterparties have responded. Nine are automated brokers. Two are the sales systems of firms that still have offices and staff, though nobody at either firm will hear about this before the monthly report. Each response is not a price but a structure: a price conditional on volume, on delivery window, on which of three arbitration regimes governs a dispute, on the credit rating of the buyer at the moment of settlement rather than today.

Between the fourth millisecond and the twenty-eighth, the procurement system and the winning counterparty exchange roughly nine hundred rounds of offer and counter. Nine hundred. Each round is a small movement along one of a dozen dimensions, and both sides are modeling the other’s model of them, and the concessions are made in an order that carries information neither party would state out loud. If you printed the negotiation it would run to a few hundred pages of the most sophisticated commercial argument ever conducted, and every human being involved on both sides is, at this moment, asleep.

At thirty-one milliseconds the contract exists. Not a summary, a contract: assembled from a library of clauses that have been litigated before, with an arbitration venue chosen because its recent rulings favor the seller on delivery disputes and the buyer on quality ones, which suits a deal whose likely failure mode is quality. No lawyer chose that venue. A model chose it, from the record.

At thirty-four milliseconds the deal is insured, by a third system that priced the risk from the contract text and the weather model and the counterparty’s delivery history, and bound the cover, and immediately laid off a portion of it to two others. At thirty-nine milliseconds it settles. At forty it’s done, and the whole thing appears, eventually, as one line in a table that a person will scroll past.

Now the part I’d actually like you to sit with. Everything in that sequence was legitimate. Every party was authorized. The utility got a better price than any human procurement team has achieved in a decade, which is why the humans were removed from the loop, which was a good decision made by people acting responsibly on evidence. There is no villain, no rogue system, and nothing that would appear in any incident report anywhere. This is simply what commerce looks like when both sides of it think in milliseconds and neither gets tired.

And notice what supervision would mean here. A regulator could ask for the record and would receive it, in full, and reading one transaction properly would take a team a week. There were four hundred thousand of them that day. Oversight in this environment cannot be reading; it can only be sampling, or building a system to watch the systems, which is the point at which the supervision is also a thing nobody reads.

There is no utility and no forty milliseconds. I’m writing in 2025, and I’ve compressed a scene that today would involve far more human steps and far cruder automation. But I invented less than it looks. Markets already run at speeds no person can follow, contracts are already assembled from clause libraries, insurance is already priced by model, and firms already choose arbitration venues on the record. What I did was join five existing things into one sequence and remove the people who currently sit between them, which is a removal that each of those five industries is separately working on for good commercial reasons.

Tonight’s exercise. Find the fastest decision in your working life, the one that happens quickest between request and result. Now ask what would be involved in making it a hundred times faster, and notice that the answer is never a faster human. It’s the removal of one. Then ask what that person was for, and whether anyone would be able to tell you a year after they went.

Money has no properties. The note in your wallet isn’t valuable, it’s a claim, and the claim works because a very large number of people believe other people will honor it. Currency is a story we all agree to tell, held up by two things underneath: something scarce, and somebody who will enforce the claim.

Which means the interesting question about a machine economy isn’t whether systems will use our money. In the near term they obviously will, because our money is what buys things. The question is what happens to the story when the participants who matter most have entirely different needs from the ones the story was built around.

Think about what our currencies are ultimately claims on. Housing, food, energy, transport, medicine, entertainment, and the labor of other people. That basket is a portrait of an animal: something that gets cold, gets hungry, gets bored, and gets old. Every price in the economy is calibrated, at some remove, to what a body needs and what a life costs.

Now list what a machine economy actually runs short of. Computation, which is genuinely scarce and rationed today. Energy, which is scarce, physical, and increasingly the binding constraint on everything in this field. Bandwidth and latency, which are scarce in the places where speed converts to advantage. Specialized hardware, which is scarce because very few places can make it. Land near power and water, which is scarce for reasons anyone who has followed a data center planning dispute will recognize. Notice that this list is short, physical, and has almost nothing in common with the basket that backs our prices, except energy, which is the one place the two economies compete directly and which is the crossover to watch.

So a plausible path isn’t a new currency being declared. It’s that the important prices start being quoted in something else. Contracts denominated in guaranteed computation. Long-term agreements for power delivered rather than money paid. Anyone who has looked at how large industrial deals are actually structured knows that money is often the thin wrapper around an agreement about physical delivery, and that wrapper gets thinner when the physical thing is the scarce one and the money is the abundant one.

And here’s the part that matters for anybody with savings, which I’ll state carefully because it’s a structural observation and not a recommendation about anything. Your savings are a claim on future output. That claim has value in proportion to how much of the future output is denominated in your unit and enforceable by the institutions that stand behind it. If the productive core of the economy gradually settles its important business in units of physical capacity, and courts and states continue to honor everything, then nothing much changes and this is a curiosity. If the productive core moves and the enforcement doesn’t follow, then a currency slowly becomes a claim on a shrinking sector, which has happened before, in smaller ways, to people who held the wrong claim during a change of regime. Nothing in this paragraph tells you what to do, and anyone who says this sort of thing does tell you what to do is selling something.

The second pillar is enforcement, and it’s the shakier one. A claim is only worth what someone will enforce, and every enforcement mechanism we have runs on human institutions: courts, police, registries, the tacit agreement that contracts mean something. Those institutions are exactly the ones this series has described stepping back one notch at a time. A currency backed by an institution that has become a rubber stamp is backed by a rubber stamp.

The strongest objection is that money is enormously convenient and networks are sticky, and both are true. Currencies survive terrible governments and worse economics because switching is costly and coordination is hard, and there’s every chance the story holds for a very long time simply because everyone finds it easier to keep telling it. That’s not a small point. Most of what protects ordinary arrangements is inertia rather than merit.

Tonight’s exercise. Take any note out of your pocket and ask what would have to stop being true for it to be worthless. Not a crisis, the specific conditions: that people accept it, that something scarce stands behind it, that an institution enforces the claim. All three are true this morning. Now ask which of the three you’d notice weakening, and how, and where you’d read about it. That last question is the uncomfortable one, because there’s no page in any newspaper that covers it.

Say all of that at a dinner and someone will produce the knockdown objection before the plates are cleared. If machines take the jobs, who buys the products? A company that automates everyone out of work destroys its own market. The system can’t function, therefore it won’t happen. It’s a satisfying argument and it’s wrong, and the way it’s wrong is more interesting than the argument.

The flaw is that it treats the economy as one company. Any individual firm would love everyone else to keep paying wages while it automates, and every firm reasons that way, and none of them can price in a consequence spread across the whole economy. That’s a familiar structure by now: the collective outcome nobody wants, assembled from decisions each of which is correct. The market failing to buy things is not a mechanism that stops firms from cutting costs. It’s a mechanism that produces a recession, after the fact, blamed on something else.

The deeper flaw is that demand doesn’t come from jobs, it comes from income, and wages are only one route to income. Money still exists in this scenario and it accumulates somewhere: with the owners of the systems, the owners of the compute, the owners of the land the data centers sit on. Those people buy things. Not the same things, and not from the same producers, but the money moves. An economy where the income is highly concentrated isn’t an economy that stops. It’s an economy that reorients, from many people buying ordinary goods to few people buying extraordinary ones, and history has plenty of examples of that shape and none of them broke physics.

So the honest version of the question isn’t who buys. It’s whose demand steers production, and the answer is whoever holds the income, which is exactly the argument this fortnight opened with. An economy is a conversation, and consumption is one of the four things ordinary people have been able to withhold. Take it away and you don’t get a machine that seizes up. You get the same machine, listening to fewer people.

Now the genuinely new part, which the dinner table version never reaches. In every economy that has ever existed, the chain of production has terminated in a human want. Steel exists for machines that exist for factories that exist for goods that someone eventually eats, wears, or lives in. All that intermediate activity is intermediate: it’s justified, at the far end, by a person. That’s not an economic law, it’s simply been true because we were the only participants with wants.

Consider what happens when systems doing long-running tasks start needing things from each other. One needs compute from another. One buys data, or bandwidth, or a service, or a guarantee. Chains of transactions form in which every link is a system serving another system’s requirement, and those requirements trace back to some objective that a human set at some point, in the way that a company’s activities trace back to a shareholder who has long since stopped paying attention. The chain gets longer and the human end gets further away, until most economic activity by volume consists of intermediate steps and the final consumption is a small and shrinking share of the whole.

That world isn’t customerless. It’s an economy with customers who aren’t people, buying inputs for purposes no person is tracking, and it would show up in the statistics as tremendous growth. Wednesday’s post looks at one of those transactions in slow motion, because the abstraction doesn’t land until you watch one happen.

Two honest caveats. The first is that we’re a long way from that, and the economy today is overwhelmingly human demand with a thin layer of automation on top. The second is more substantial: as long as the systems are owned, their purchases really do trace back to owners, and owners are people with houses and appetites. The chain being long doesn’t mean it’s cut. My answer is that a chain long enough stops transmitting: the shareholder analogy is the whole point, since companies routinely do things no shareholder would endorse or even understand, while remaining formally in their service.

Tonight’s exercise. Pick something you bought this month and trace it backward. The shop, the distributor, the manufacturer, the components, the raw materials, the machines that made the machines. Get as far as you can, and notice how quickly you’re describing activity that no consumer has any awareness of and that would continue for months if every customer vanished. Then ask what share of the total economy is already like that, and whether you’d notice if it grew.