Full automation, the wage, and who stays on the books.
By J.W. Sher · July 2026 · Fourth piece of The Financial System of Free Market Ecology series.
Here is a thought experiment, and it is not a comfortable one. Suppose the robots arrive in earnest — not the chatbots and warehouse arms of the mid-2020s but the full package: artificial general intelligence, a billion capable humanoid machines, robot armies in the literal sense, each doing whatever its owner wants. Suppose also that the economic system underneath is the one we have now — fused money-and-credit accounting, ecological cost priced in a currency that can be printed, and every person’s claim on output flowing through what they can sell, which for most people means their labor.
Now ask: at what population does that world become ecologically sustainable?
The question sounds absurd until you notice that it has an answer, and that the answer is small. Sustainability is an ecological budget divided by per-person throughput, and in the world described, neither term is anchored. The owners’ throughput is effectively unbounded — machines that extract, build, and consume on their behalf without wages, without rest, without any of the friction that historically kept one person’s appetite finite. Run the division with per-owner consumption at war-economy scale and the sustainable population comes out somewhere shockingly low. Under those books, at that appetite, ten million is a defensible guess. It might be a hundred million; it might be less than ten. The estimate is not the interesting part, and this essay will not defend any particular figure. The interesting part is that in the world described, nobody chooses the number. It emerges — from concentration dynamics, from the physics of the budget, from whoever happens to hold the machines — and it is written on no ledger, voted in no assembly, and priced in no market. It is a residual. A first-order fact about the human future, arrived at by default.
This essay is about why that default is a property of the accounting, not of the robots, and about what changes — and, in keeping with the previous essay in this series, what pointedly does not change — when the books are honest.
The wage was doing two jobs
For the entire history of industrial civilization, the wage has quietly performed a second function on top of its official one. Officially, a wage is the price of labor. Unofficially, it has been the mechanism by which claims on the economy’s output stayed distributed across the population — not because anyone designed it that way, but because extraction and production physically required hands. Every ton of ore moved, every field harvested, every machine tended leaked purchasing power from the owner of the enterprise to the people who did the moving, harvesting, and tending. The leakage was the distribution system. A worker’s claim to exist economically — to command food, shelter, and a place in the ledger of who gets what — did not rest on any principle. It rested on being needed, and on the strike-shaped leverage that being needed confers.
Full automation is precisely the end of being needed, and so it is the end of the leakage. This is a conditional claim, not a forecast — I make no prediction about when machine labor becomes a general substitute for human labor, only about what breaks if it does. What breaks is not employment as a statistic. What breaks is the only mass distribution mechanism the system has, and there is no second one behind it. Ownership distributes to owners. Wages distributed to everyone else. Remove wages and the fused system’s answer to “what is a person’s claim on output?” is: whatever discretionary transfer the politics of the moment provides.
It is worth being fair to that answer, because it is not nothing. Welfare states exist; a world of robot abundance could fund transfers of enormous size, and a democracy might vote them. There are even fused-books instruments that distribute ownership rather than transfers — the Alaska dividend, the Norwegian fund, pension capital — and they are the strongest form of the answer. But look at the shape of these instruments. A transfer on fused books is a flow someone dials — revocable each budget cycle, its cost dissolved into inflation and debt where no one can price it, its recipients positioned as supplicants of whoever holds the dial. And the ownership variants are denominated in a unit their issuer prints, drawn on equities their issuer can dilute or rescue — the same two holes, one layer up. And the political leverage that won every historical expansion of that instrument — the strike, the labor shortage, the army that needed soldiers, the factory that needed hands — is exactly what automation removes. A wage-less population negotiating with the owners of a billion robots has votes but no strikes. It can ask. The fused books make the asking invisible and the granting discretionary, and the population carried becomes whatever the discretion decides — which returns us, by a politer route, to the number nobody chooses.
None of this requires villains, and the argument is wrong if it needs any. No owner in this story sets out to expel anyone. Each simply consumes what his machines can supply, on books that were never asked to record what the supplying used up or whose claim lapsed when the wages stopped. The catastrophe, if it comes, is compounded of ordinary appetites and an accounting system with two specific holes: ecological cost denominated in a printable unit, and human claims denominated in a wage that machinery can abolish. The robots do not create either hole. They pour a billion workers’ worth of throughput through both at once.
What honest books change
Free Market Ecology closes the two holes, and it is worth being precise about the mechanism, because the mechanism is the whole argument.
The first hole — the printable unit — is closed by the cap and the RUR. Ecological cost is denominated in physical, non-fungible resource rights: so many hectare-months, so much water actually recharged, so much ore against a cumulative total, each set by measurement and never expanded because demand grew. A robot workforce changes none of this. Machines borrow their draw like any producer — every hectare a fabricator stands on, every ton its supply chain moves, is debt to the ecological lenders (the Ecological Private Finance of the earlier essays) embedded in the goods and settled when someone consumes them. Under honest books, a billion robots can no more enlarge the planet’s budget than a billion printing presses can. What they can do is make everything made cheaper in money: as machine labor drives the labor content of goods toward zero, prices sag toward resource cost — toward the part that can be engineered down but never printed away, because whatever draw remains must be settled in a unit fixed by physics.
Which is where the second hole closes, if — and this conditional is the hinge of the argument — the jurisdiction has chosen the distributed pole, minting the commons to households as shares. Then the resource component of every price originates with the population. The machines borrow their physical budget from the lenders, but the debt embedded in everything they make can be retired only by rights someone was minted and chose to surrender or sell — and the owners’ own consumption, however machine-served, cannot settle out of their own pro-rata slice; it settles in rights bought, for money, at asked prices, from the people who abstained. A person with no job and no prospect of one holds, by grant or inheritance, a share of the one input the machines cannot synthesize, and the frugality path — consume modestly, sell the surplus to those whose appetite outruns their share — stops being the eccentric choice it is in a labor economy and becomes the ordinary condition. The previous essays in this series traced this circuit in its normal operation; the robot era is the circuit under maximum load. The mass of claims that used to flow through the wage flows instead through the mint, and it flows as property — not a stipend a ministry dials, but a holding the heavy consumers of the robot economy must purchase, month after month, at prices the holders name.
The Abdication — the narrative that runs this framework through a century on one island — gives the sentence for this to a woman whose grandmother cut cane on ground she now co-owns: “My grandmother harvested this island for its owners. So will the machines. The difference is the owners.” The thought experiment that opened this essay is that sentence’s inversion. The machines harvest either way. The accounting decides for whom — and, downstream of that, how many people the harvest is for in any economic sense at all.
What this argument does not claim
The previous essay in this series drew a line the framework must not cross, and this argument sits close enough to that line that the disclaimers are load-bearing rather than ornamental.
FME does not prevent the ten-million world. The distributed mint is a political choice; a statist implementation — the government holding the rights — balances its books just as well while answering every distribution question by decree. But under a distributed mint the population cannot be bought out of its claim: the share is inalienable, tied to the citizen like a pension, so no one can sell their stake to the robot owners and no fortune can buy it from them — it passes only to their children. What honest books guarantee is therefore stronger than mere visibility. Under fused books, the mass of humanity leaves the ledger the way a canceled invoice does — wages stop, and there is simply nothing left that records a claim. Under a distributed FME mint there is no way to make them leave by purchase at all: the robot economy must keep buying their flow, month after month, at the prices they name, because the tap that mints it can never change hands for money. The only road from here to the small world runs back through the statist pole — a government re-chartering the commons into its own hands, in the open, by law or by force.
Nor is this a prediction that the robots are coming on any schedule, or that general machine labor is even achievable. The argument is insurance-shaped: it concerns the asymmetry of the stakes, not the probability of the event.
And the number itself — ten million — should be held loosely and without relish. Nothing in this argument says a smaller population is desirable, and the reader who hears Malthus should hear again: the catastrophe in the thought experiment is not that the earth has limits, it is that a specific accounting system allocates the entire planet to whoever owns the machinery, by default, invisibly. The point of honest books is the opposite of the small world: billions of people, each holding a claim the robot economy must honor, prospering within a budget that is finally written down. The framework’s quarrel is with bad accounting, not with people, and any reading of it that ends in “fewer humans” has misread it at the root.
The window
There is a timing structure in this argument, and it is the closest the argument comes to urgency.
In a labor economy, the commons grant is cheap. Resource rents are a modest share of most prices — labor is the expensive part — so distributing the mint to households redirects a small stream, and the incumbents who might object are conceding little. This is the grandfathering logic the framework uses everywhere, applied at the largest scale: buy the incumbents in while the buying is cheap. In a robot economy, the proportions invert. When labor content goes to zero, resource rents converge from a modest share of income toward the dominant one — the shares thicken until the resource stream is the main channel automation cannot compete away — and the commons is no longer a side stream but the main river. Whoever holds the rights at that point holds the economy’s principal income, and the political price of distributing it rises from a rounding error to an expropriation — demanded, moreover, by a population that no longer has anything the holders need. The same grant that is a minor fiscal reform on one side of the transition is a revolution on the other side, and revolutions from a position of no leverage have a poor record.
So the window argument is this: a society that thinks it might ever want the distributed pole should establish it while its labor still matters — while the grant is cheap, the politics ordinary, and the people receiving it are still parties the economy has to bargain with. Afterward the books can be as honest as anyone likes; honest books record who owns everything, and by then the answer is already written.
None of which tells a society what to choose. That was the previous essay’s whole argument, and it holds here: how many people, holding what, under which government — politics decides all of it, under FME as under anything. What the robot era changes is only the cost of deciding by not deciding. On fused books, the default is a number no one picked, reached in the dark. On honest books, the default is visible years before it arrives — every right concentrated, every move toward the statist pole, every claim redirected, itemized in public while there is still time to vote, to resist, or at least to know.
This is the fourth piece of the Financial System of Free Market Ecology series. The third, What Free Market Ecology Doesn’t Handle, draws the line this one leans on: the framework decides no political question — it prices them. The robot era is the case where the price of leaving a question undecided becomes the human population itself.