Why Money Survives Abundance

Scarcity, time and trust in an age of machines. Machines can make everything made of thought cheap, but scarcity moves rather than ends, and money, prices and trust return with whatever stays scarce.

Machines are making thought cheap, and an old forecast has returned with them. Keynes made it in 1930, and some of the people building the machines repeat it today: once robots and software produce more than anyone can consume, scarcity ends, and prices, money and the need to economise end with it. This article argues from first principles that the forecast mistakes a relation for a property. Scarcity moves. It does not end. The article then follows the consequences one step at a time: what stays scarce, who decides how it is used, why that takes money, and why money takes either trust or a way of doing without it.

1. A world without money

Iain M. Banks published the Culture books between 1987 and 2012. The Culture is a civilisation that has left scarcity behind. Its machines make whatever anyone asks for, and nobody has to work. Artificial Minds of almost unlimited ability run its ships and habitats. It has no money. The first novel calls money a clumsy way of rationing, made pointless once production outran every demand.

The books are favourites of the people now building such machines. SpaceX’s rocket-recovery ships carry the names of Culture spacecraft, a meeting room at DeepMind is named after Banks, and he has been called one of the technology industry’s favourite writers (The Economist). What they take from him is a forecast. If machines and software agents can produce more goods and services than anyone can consume, nothing is left to economise. Prices fall toward zero, work becomes a hobby, and money is a relic.

A world in which everything is given is a world in which nothing has to be given up. No good is preferred to another for want of means, so there is no order among desirable things and no reason to do one thing before another.

That is not a description of plenty. It is a description of a being that does not choose. The point is Mises’s, and it opens Human Action (1949): to act is to choose, to prefer one state of affairs and set another aside. Robbins (1932) had already defined economics by the same fact, as the study of behaviour where ends are ranked and the means to reach them are scarce and have alternative uses. Wherever someone still prefers anything to anything else, and cannot have both at once, something is scarce.

So the forecast cannot be true in its strict form. Something always stays scarce, and the first task is to say what.

2. Scarcity is a relation: the land of Cockaigne

To see why, start with what scarcity is. It is a relation between wants and means, not a property of things, so abundance moves it without ending it. Menger (1871) defined an economic good as one whose available quantity falls short of requirements. Machines can push many goods across that line. They cannot abolish the line.

The oldest picture of the other side of the line is the land of Cockaigne. In the medieval poems its rivers run with wine, its houses are roofed with cakes, and roast fowl fly into open mouths. Bruegel painted it in 1567: a clerk, a peasant and a soldier asleep on the ground beneath a laden table. It was a peasant’s dream, and what it abolished was labour.

Mises (1949) used it as a test case and found what the dream leaves standing. Even in a land of Cockaigne, where every material want is met, a person must still decide what to do first. Some satisfactions exclude each other and cannot be enjoyed in the same hour. Bruegel’s sleepers cannot feast while they sleep.

Detail of a painting showing a clerk, a peasant and a soldier asleep on the ground beneath a round table laden with food
Pieter Bruegel the Elder, The Land of Cockaigne (Das Schlaraffenland), 1567. Alte Pinakothek, Munich.

Life is finite and satisfactions are sequenced, so time is economised whatever else is free. The only exit Mises allows is immortality with eternal youth, and even then the ordering of incompatible pleasures remains.

Time preference follows from that. A satisfaction now is ranked above the same satisfaction later, because the later one costs waiting and waiting uses up the one stock that cannot be replenished. Interest is the price at which present goods trade against future ones. It would exist in Cockaigne too, as soon as anyone there promised anything to anyone.

Machine intelligence is the nearest thing to Cockaigne yet built, and only for one class of goods: those made of thought. The question is where scarcity goes when thought is cheap.

3. Where scarcity goes

When one factor of production becomes cheap to reproduce, income moves to the factors that complement it and cannot be reproduced. Three older literatures describe the same movement from different sides.

Hirsch (1976) separated material goods, which technology can multiply, from positional goods, whose value lies in relative standing. Baumol (1967) showed that services which need a person’s hour grow dearer as everything else grows cheaper. George (1879) argued that land rent absorbs the gains of everyone else’s productivity, because land is fixed.

All three describe value retreating into what cannot be reproduced: rank, another person’s hour, a location. Rank is scarce by definition, which is why status competition survives any level of output. That is an ordinal argument. It needs only that people rank outcomes, never that anyone measures them, and that is how the tradition from Menger to Mises has always treated value.

Böhm-Bawerk (1889) supplies the classification. Land and labour are original factors. Capital goods are produced, so their prices tend toward cost. If machines make substitutes for labour reproducible, rents shift to the factors that stay original.

The evidence predates the machines. The rise in the net capital share of advanced economies since 1950 is largely housing (Rognlie, 2015), and the long rise in house prices is largely land (Knoll, Schularick and Steger, 2017). Software became abundant, and the returns went to whoever owned its scarce complement.

Scarce in every scenario Why it cannot be produced
Time and attention Lives are finite and action is sequential
Land and location Fixed by geography and permission
Energy at a given place and hour Bounded by physics and by the time it takes to build
Rank and reputation Relative by definition
Proof of origin A copy can be made; the fact of having been first cannot
Rule-fixed units Scarce by protocol, for as long as the rule holds

When text, images and voices can be synthesised without limit, evidence that a thing existed before a certain moment, or came from a certain person, becomes a good in its own right. The last row is the least familiar. It is a unit whose quantity is fixed by a rule that nobody can change. Like land, more of it cannot be made. Land is limited by nature and the unit by a rule. Whether that kind of scarcity carries value is a separate question, taken up in Section 8.

The table says what stays scarce. It does not say what any of it is worth, and that needs one more step.

4. Why cost does not create value

The abundance forecast is at bottom an argument from cost: what costs nothing to make will be worth nothing. For goods that can be reproduced, that is right. For the scarce things of Section 3 it is not, because their value never came from cost in the first place.

Value runs from ends to means, never the other way. Menger called the goods we consume goods of the first order and the goods that produce them goods of higher order. A higher-order good has value only because the thing it helps to produce is wanted. Wieser later named this the law of imputation.

Costs do not escape the law. The cost of a factor is the value of what it could have produced elsewhere, and that value is imputed too. For reproducible goods the price tends toward cost, as Böhm-Bawerk said, but the cost is itself a reflection of other prices. Energy and cost follow price.

A clean example is any asset whose new units are awarded, on a fixed schedule, to whoever spends the most computation competing for them. Let N be the number of awards per year, R the units per award, P the price of a unit, c the price of electricity per kilowatt-hour and φ the share of power in marginal cost. Entry is free, so competitors arrive until the cost of competing equals the revenue. Annual energy use is then:

E* = φ × N × R × P / c    (kWh per year)

Machine efficiency does not appear. Better machines mean more computations per joule, and the same joules. Energy use is proportional to revenue: halve the price and it halves. Make electricity ten times cheaper and ten times as much of it is burned, for the same money. The schedule of awards does not move in either case.

The energy spent does not give the unit its value. The value of the unit determines how much energy is spent. A gold mine obeys the same logic in a softer form. A higher gold price opens marginal mines; it is never the digging that made the gold dear.

The slogan that such an asset is “stored energy” is therefore a cost theory of value with joules in place of labour hours. Its lineage is Ricardo and Marx, then Henry Ford’s energy dollar of 1921, Soddy’s energy economics and the Technocrats’ energy certificates. It is not Menger’s.

There is a practical objection as well. A kilowatt-hour has no single value. At noon in a solar field in May it is close to worthless; on a windless January evening in a city it is worth a great deal. Electricity is thousands of goods distinguished by time and place, and their relative values are found only by exchange. An energy standard would need a price system to say which energy, which is the problem it set out to solve.

What the energy does buy is real but specific. Section 7 returns to it.

5. Calculation, knowledge and the machine planner

Value, then, comes from what people want and cannot all have. That leaves the question of who decides how the scarce things are used. There are two candidates: a planner, or prices.

The Culture chooses the planner. Its Minds are intelligent enough to allocate everything directly, so prices are not needed and money withers.

Four results stand against it. Hayek (1945) showed that the knowledge an economy runs on is dispersed, local and often tacit. Arrow (1951) showed that individual rankings cannot, under a few mild conditions, be aggregated into a consistent social one without a dictator. Paul (2014) and Callard (2018) showed that people’s values change with the experiences they choose, so there is no fixed target to aim at.

The fourth is Mises’s calculation argument of 1920, and it bears hardest on machines. His point was not that the planner lacks data. It was that without exchange of privately owned capital goods there are no prices for them, and without prices no plan can be compared with another. The problem is the absence of the numbers, not the difficulty of processing them.

A machine planner inherits this intact. More intelligence processes given data faster. It does not create the bids of rival owners risking their own property, and those bids are what a price is. The Minds have no answer to it except the author’s say-so.

The natural reply is that the machine could simulate the market. But preferences are not a dataset waiting to be read. Rothbard (1956) argued that they exist only as demonstrated in choice. A simulated bid risks nothing, and it is the risk of loss that makes a bid informative.

This sorts the proposals for machine intelligence cleanly. Russell’s (2019) design, a machine that stays uncertain about what people want and learns from what they do, is close to the view taken here. A machine given a fixed objective is the planner’s error in silicon.

Nor does a world of software agents escape the argument. It sharpens it. Billions of agents will compete for electricity and computing capacity, both of which have other uses. Deciding which task gets the next kilowatt-hour is a calculation problem, and it is solved the old way, by prices. An agentic economy needs more prices than a human one, not fewer.

Prices are stated in money. If no planner can replace prices, money returns with them, and the next question is what money is.

6. Money and credit: an asset and a claim

Prices have to be stated in something, and that something is money. Money is the most saleable good: the one thing everyone accepts because everyone else does. Menger (1892) explained how it arises without anyone designing it. In barter, a trader who cannot find a direct match accepts a good that is easier to sell on than his own. The more people do this with one good, the easier that good becomes to sell, until it is wanted for that reason alone.

The abundance forecast sees only one of money’s jobs. It assumes money exists to buy produced goods, and that job does shrink as goods get cheap. But money also allocates whatever stays scarce, and it carries decisions through time. Neither job depends on goods being dear.

The second of those jobs is the one Section 2 prepared. Present goods trade against future goods through money, so it is the medium in which time preference is expressed. A person who holds money is keeping a decision open.

Credit comes from the same source. People differ in patience. One wants to consume or build now, another is willing to wait, and they trade a present good for a claim on a future one. A claim needs a debtor who can pay, an enforcer who will make him, and a creditor who believes both. This is the point at which trust enters the argument. A base money such as gold coin is an asset that is nobody’s liability. Credit is always somebody’s promise.

Mises (1912) drew the line. A money substitute is a claim payable on demand that circulates at par because nobody doubts it. Where such claims exceed the money held against them, he called the excess fiduciary media. Holmström (2015) describes the same object a century later as information-insensitive debt. A money substitute is debt that nobody examines.

Every monetary system is therefore a hierarchy, with an asset at the base and layers of promises above it (Mehrling, 2013). The promises are convenient and usually sound. They are also the place where trust enters, and where it can fail. The distinction between the asset and the claim organises everything that follows.

7. How a trustless base is built

Why does the base matter in an article about abundance? Because the asset at the bottom of the hierarchy must itself be scarce, and the machine age is the age in which anything digital can be copied without limit. Either someone is trusted to keep money scarce, or its scarcity is secured in a way that needs nobody’s word. The first is the planner’s problem of Section 5 in monetary form. This section describes the second.

Gold was a base that needed no trust: scarce by nature and nobody’s promise. It was also heavy and hard to assay, so people held claims on it instead, and the claims came to be trusted more than they deserved. A digital base faces the opposite problem. It moves at the speed of light, but its scarcity has to be made by rule. The question is who keeps the rule.

If an issuer keeps it, the unit is a claim on the issuer. Building a base with no issuer took four steps, each the answer to a precise problem.

  1. Order in time. Spending a digital unit twice is a question of which transfer came first. Haber and Stornetta (1991) showed how to timestamp data without trusting the notary. Each timestamp contains the fingerprint of the one before, so altering an old entry changes every later one. From the mid-1990s their company’s weekly fingerprint was printed in the classified pages of the Sunday New York Times. The newspaper was the witness: nobody can recall every copy.
  2. A witness with no publisher. Dwork and Naor (1992) and Back’s hashcash (1997) supplied proof of work: a computation that is costly to produce and trivial to check. Nakamoto (2008) put the two together. Anyone may add the next block of entries by paying that cost, and everyone accepts the valid chain with the most work in it. Rewriting history means redoing the work faster than the rest of the world combined.
  3. Issuance by rule. New units are awarded with each block on a schedule fixed in advance, and their total is capped at twenty-one million. The difficulty of the work adjusts so that blocks arrive at a steady pace however much energy competes. Supply is a function of elapsed time and of nothing else.
  4. Verification by anyone. The full history can be checked on modest hardware. Ownership is knowledge of a key. A holder with his own keys has no counterparty, and settlement becomes more final with every block he waits, because an attacker’s chance falls geometrically.

This is what the energy of Section 4 buys. It sets the cost of rewriting the ledger. It makes issuance costly to obtain, so nobody enjoys free seigniorage. And it is a signal that real resources were committed to this history and no other, which Szabo called unforgeable costliness.

The third step is what separates the design from Ford’s energy dollar. His unit was a claim on energy, so abundant energy would have meant abundant money. Here cheaper energy raises the work per block and leaves the schedule alone. A solar glut or fusion would make the ledger costlier to attack in physical terms and add not one unit to its supply.

“Trustless” needs a careful reading. The design does not remove assumptions. It moves trust from persons and institutions to three other things: that certain mathematical problems are hard, that no single party controls most of the work, and that users keep running the rule.

Hayek’s distinction explains why the third is not planning. A general rule announced in advance is not a command. But rules age. Signature schemes weaken, and security must still be paid for as issuance declines. Those are collective decisions, and no code can make them.

8. Scarcity is necessary, not sufficient

Building a scarce base does not make it valuable. Section 4 showed that value comes from wants and not from cost, and the same holds for scarcity made by rule. Scarcity is necessary for value and not sufficient. Menger required four things of a good: a human need, a causal link to satisfying it, knowledge of that link, and command over the thing. A capped ledger meets the last. The first three depend on other people wanting to hold it.

A familiar objection says that an asset with no cash flows can have no value at all. Hutt (1956) answered it for money in general. Money held yields a service, which is availability under uncertainty, and nobody discounts the cash flows of a banknote. That removes the objection without supplying a price.

Pricing a fixed stock

Rothbard’s stock analysis states the pricing rule exactly. When supply is fixed, the price is set where the existing stock equals the total demand to hold it. The supply curve of a rule-fixed unit is vertical by construction, so its price is reservation demand and nothing else.

This is where the popular stock-to-flow ratio belongs. The ratio of the existing stock to a year’s new production measures how little a rise in price can dilute holders. It explains why gold and not copper became money: a high copper price calls forth copper. It is a useful measure of how hard a supply is. It says nothing about demand, so it cannot be a model of price. Stock-to-flow describes the supply side, which was never in doubt.

The regression theorem as an equilibrium selector

Demand for money looks circular. People hold it because it has purchasing power, and it has purchasing power because people hold it. Mises saw this in 1912 and broke the circle with history. Today’s demand for money rests on yesterday’s purchasing power, and the regress ends at a commodity once valued for use.

The modern literature faces the same circularity and concludes that there are many equilibria, including one at zero (Garratt and Wallace, 2018; Biais et al., 2023; Pagnotta, 2022). The two are closer than they look. The regression theorem is an equilibrium-selection device: history picks the price path, and expectations anchor on the last observed price. Neither school finds an intrinsic value.

The first rule-fixed ledger tested the strict version of the theorem, because its unit had no prior use. Economists in Mises’s tradition split. Davidson and Block (2015) argue that the commodity requirement binds only when money emerges from barter, and this unit was born into existing money prices. Luther (2018) argues that the case shows the theorem is wrong or narrow.

One historical detail favours a middle reading. The first published exchange rate, in October 2009, was computed from the cost of electricity. Cost supplied a focal point for the first trade; demand took over afterwards. The requirement of a commodity origin did not hold, but the role of history did.

The practical consequence is symmetric. Theory provides no floor under the price and no ceiling over it.

Hayek’s test

Hayek predicted in 1976 that if people could choose their money, they would choose stable purchasing power. The market has run the experiment. Dollar-pegged tokens stood at roughly $285–290 billion in mid-November 2026 (DefiLlama). They exist to be spent, not held for gain.

So the demand for a stable private medium was met by a tokenised dollar, as Hayek’s argument implies. The fixed-supply unit kept the savings demand. Its design is Rothbard’s, a fixed stock indifferent to the price level, not Hayek’s managed private money.

The tokens are worth a second look for a reason Hayek did not have in view. Each one is a promise by an issuer to redeem. They are claims, which brings us to the boundary.

9. The trust boundary: the base and the claims on it

Section 6 separated the asset at the base from the promises above it. With a base that needs no trust, that line becomes sharp. Only the base asset held with one’s own keys has no counterparty. Everything built on it is a claim, enforced by law, and the market price of each claim is a running measure of how far it is trusted.

The layers run from most trust required to none.

Layer What the holder owns Enforced by Mises’s category Where trust is priced
Equity Shares in a company that holds the asset Corporate law and management discretion Entrepreneurial claim Premium or discount to the assets held
Credit instruments Bonds and preferred shares issued against it Securities law and courts Credit claim that aspires to be a money substitute Yield; discount to par
Secured credit Loans collateralised by the asset Contract law and margin calls Commodity credit, if funded from savings Interest rate and haircut
Titles Fund shares; units at a custodian Custody and securities law Fully backed title, like a money certificate Discount to net asset value
Unbacked balances Deposits at a lender that re-lends them Bankruptcy court, after the fact Fiduciary media Withdrawal freeze
Base The asset, held with one’s own keys Code and energy The asset itself, not a claim Blocks waited; fees paid

There are two prices of trust in this table. Finality on the base is bought with time. Finality of a claim is bought with yield. One layer prices trust in minutes, the others in basis points.

The boundary also explains why law came back. Dai (1998) hoped that contracts among pseudonyms could be enforced by posted bonds and reputation. Greif’s history of medieval traders explains why that does not scale. Reputation works inside small closed groups. Impersonal exchange needs a third party who can compel. Code settles; it does not adjudicate. The base never left code, and claims were never outside law.

10. What people build on a scarce base

Give a society of status-seeking, credit-using people a base that cannot be inflated, and they do three things with it.

They lend against it. The demand for credit does not come from the monetary base. It comes from time preference, which differs between people, and from the wish to hold wealth and spend it too. A technology that permits full reserves does not make anyone choose them. Within fifteen years of the first block, a lending market had formed on the new base, and it had its first crisis.

The crisis is instructive. In 2022 several large lenders that had mixed custody with lending failed. There was no lender of last resort. Losses fell on creditors through bankruptcy, and the supply schedule did not move by one unit. There was liquidation, not socialisation, and no transfer through the base from savers to the rescued.

So the hope that hard money ends the credit cycle is not supported. Minsky’s point stands: the cycle lives in the behaviour of borrowers, not only in the elasticity of the base. What a hard base changes is who bears the loss. That is the result the free-banking school predicted (White, 1984; Selgin, 1988): credit forms, reserves are high because nobody will rescue, and failures stay contained.

They keep their accounts in something else. Credit on the new base is secured by it and denominated in state money. Few borrowers will promise to repay an asset they expect to appreciate. People quote prices in the unit in which their income and debts are already stated (Doepke and Schneider, 2017). And where exchange rates float, they spend the money they expect to lose value and keep the one they expect to hold it (Bernholz, 2003).

The base therefore works as savings and as collateral inside the existing price system. It is not a unit of account. A market in debt denominated in the base would be the sign that people had begun to calculate in it. Until then the base protects savings from whoever manages the money. It does not protect the price system from him.

They compete with it. Hirsch’s argument returns here. As material goods get cheaper, a larger share of income is spent bidding for what cannot be multiplied: the flat near the good school, the seat, the view. A scarce asset is the natural collateral for that contest, because it lets the owner bid without selling.

Cantillon (1755) explains who gains. New credit reaches some hands first, and in a system of collateralised lending those hands belong to the owners of collateral. Borrowing against a trustless base does not abolish that channel. It adds the holders of the base to it. This describes how the present order distributes. It is not an alternative to it.

11. Machines, original factors and who owns the capital

The remaining question is who ends up owning the scarce things. If machines make labour reproducible, income shifts to whoever owns what machines cannot make. Capital theory predicts that, and the table in Section 3 is the list of those things. The theory is less comforting about the transition than its popular versions suggest.

What the theory says

In a competitive market wages tend toward the discounted marginal value product of labour. For a task a machine can do, competition pushes the price toward the cost of the electricity and the hardware time it uses (Hanson, 2016; Korinek and Suh, 2024). That cost becomes a ceiling on what a person can charge for the same task.

Mises generalised Ricardo’s law of association to show why this is not the end of work. Even if machines are better at every task, people keep a comparative advantage while machine capacity is scarce and has better uses. Capacity is scarce while computing power and energy are. So the survival of labour income turns on the price of energy and computation.

Baumol names the tasks that stay outside the equation: those where the buyer is paying for a person. Care, teaching, craft, performance and accountability belong here. A signature that someone will answer for is worth more than an analysis nobody will.

What the theory does not promise

Nothing in it keeps any particular wage above any particular level, or makes the adjustment quick. Allen (2009) documented Engels’ pause: British output per worker rose for about forty years from 1800 while real wages stayed nearly flat. Mises was right that technological unemployment is not permanent. Forty years is a working life.

The capital side needs a correction too. Piketty’s rule that the return on capital exceeds growth treats capital as a fund that grows by itself. Lachmann (1956) stressed that it is a structure of specific, unlike goods that entrepreneurs must keep regrouping. Returns are earned by judgment and lost by error. Owning machines is not the same as owning a rent.

Who owns the robots

The abundance forecast skips this question, and it is the old question of who owns the means of production. Money stops mattering to consumers only if the output reaches them. There are four answers.

The politics of the coming decades will be fought between these four. Two things can be said in advance. Land rent is the one tax base that cannot move, which is why George’s proposal keeps returning. And a bearer asset outside every issuer’s control is the hedge for those who doubt whichever planner wins.

12. Where this leaves us

This week supplied a test case. On 20 November 2026 Levent Alpöge, a mathematician at Anthropic, released a proof of the Riemann hypothesis that he had found working with the AI model Claude. The question had stood open since 1859. A day later the Lean and Mathlib communities certified that both the statement and the proof are correct (joint statement).

A short piece of speculative fiction that circulated online in July had imagined this very week. In the story the hypothesis falls on 21 November and the proof fits in a single post. Within three months machines dispose of P versus NP, the halting problem and Gödel’s theorems. Its first line came true almost to the day. The differences between the story and the event are where the economics lies.

The proof does not fit in a post

It is a library of about 17 million lines of formal code. Kevin Buzzard of Imperial College estimates that this may correspond to tens of thousands of pages of mathematics, and no person has yet read all of it. In his words, “we don’t even know why the theorem is true” (Al Jazeera).

The proof was abundant on arrival. Understanding it is the scarce good, and it is rationed by the oldest constraint in this article, which is human hours. A world with every theorem on the shelf is Cockaigne for mathematicians. They must still decide which one to read first.

It was believed without being trusted

Nobody had to take the author’s word, or the machine’s. The proof was costly to produce and cheap to check. Dozens of people rebuilt it independently, and several separately written checking programs accepted it.

That is the architecture of Section 7, applied to mathematics. A formal proof and a proof of work share one property: the burden lies on the producer, and the check is open to anyone. As machines produce more of everything that is made of thought, cheap verification becomes the scarce complement. Timestamps, signatures and a ledger anyone can audit are goods of that kind.

The boundary sits in the same place

Verification establishes that the proof is valid. It does not say what the result means, which questions were worth asking, or what should be done next. Those are judgments, and a judgment offered to someone else is a claim.

Trust can be removed exactly where checking is cheap. Where it is not, in a borrower’s promise, a planner’s values or a machine’s advice, trust returns and has to be priced: in yield, in law or in reputation. The trust boundary of Section 9 is a verification boundary.

Some limits are theorems

The rest of the story is a different kind of event. The Riemann hypothesis was an open question. The undecidability of the halting problem (Turing, 1936) and incompleteness (Gödel, 1931) are proved results about what any formal procedure can do. More intelligence settles what is open. It does not overturn what is proved. To expect otherwise is the Cockaigne error once more: mistaking a line for a quantity.

P versus NP is the item that matters for money, and it is open. Everything in Section 7 rests on an asymmetry: some problems are hard to solve and easy to check. Signatures and proof of work assume it. Formally they assume even more than that P differs from NP, namely that one-way functions exist. Neither has been proved, and this week’s result does not bear on either directly.

So a trustless base does not eliminate trust. It moves trust from persons to mathematics. That was a comfortable place to put it while mathematics moved at human speed.

Where we are headed

The breakthroughs mean that thought has become a produced factor, even at the frontier. Böhm-Bawerk’s rule then applies. The price of a produced thing tends toward its cost, here electricity and hardware time, and the rents move to whatever stays original.

Keynes’s own forecast is the caution here. He was roughly right about productive capacity in 2030 and wrong about the fifteen-hour week. Capacity is easier to foresee than what people will want once they have it.

With that caution, the direction is clear enough. Anything made of thought that can be sent down a wire will become very cheap. Time, land, energy at a given place and hour, rank, a person who will answer for a decision, and proof of what is real will not.

On this argument money persists because choice does. It does less work buying produced goods and more allocating what cannot be produced. Prices matter more, because machines must be told which scarce kilowatt-hour to use. Credit persists because people differ in patience and compete for rank, and law persists because claims need a judge. Who owns the scarce complements is the political question of the period.

A base that is nobody’s promise has a definite place in that world. It sits at the bottom of a stack of claims and is valued for needing no one’s word. It is valued only as far as others wish to hold it, and theory gives it neither floor nor ceiling.

The open question

Elon Musk, who often cites the Culture novels, said in 2026 that money will not matter by 2036 (Fortune). His reasoning is that people want money for goods and services, and that robots will supply more of both than anyone can use. He also grants that electricity and mass will stay constrained (Business Today).

Whether he is right depends on questions this article can frame but not settle. Will anything remain that people rank and cannot all have? If so, who decides how it is used: a planner whom everyone must trust, or prices? If prices, stated in what money, and resting on whose promise? And who will own what stays scarce?

If every answer goes the Culture’s way, money will not matter. If one of them does not, it will. Either way the day will still have twenty-four hours, and each of us will still be deciding what to do first. That is all scarcity ever meant.

References

The two linked sources in Section 12 and the stablecoin figure in Section 7 were opened on 21 November 2026. The works below are cited from the standard literature; editions and page numbers should be checked before formal citation.