Julian Simon was right. That is exactly the problem now.
By J.W. Sher · July 2026 · A companion to The Number Nobody Chooses.
In 1980 the economist Julian Simon issued an open challenge, and the biologist Paul Ehrlich took it. Ehrlich, who had written The Population Bomb, was certain that a growing population pressing on finite resources would drive prices up; he chose five metals and bet they would cost more in ten years. Simon had argued the opposite — that human ingenuity is the ultimate resource, that scarcity is a price signal, and a price signal is an invitation, and the invitation is always answered; he bet the metals would cost less. They settled on copper, chromium, nickel, tin, and tungsten. Ten years later every one of them was cheaper. Simon won, and he was right to win.
He was right about more than the metals. The canonical case — the one every market-minded person reaches for the moment an environmentalist starts talking about running out — is whale oil. In the 1850s much of the industrial world lit its homes with oil rendered from sperm whales. The whales grew scarcer; the price climbed; and the climbing price did exactly what Simon said it would. It made kerosene worth refining. By the 1860s petroleum distillate was cheaper and better, the whale-oil lighting market collapsed, and no committee, no treaty, and no ration book was required. The price system found the substitute and routed demand to it. This is not a story the right tells itself. It happened, and it is the correct account of how it happened.
I want to concede all of this at the outset, because I am about to argue that the mechanism has a failure mode, and the argument is worthless if it does not first grant how well the mechanism works. Simon was not lucky. He understood something Ehrlich did not: that a resource is not a fixed pile but a function of what we know how to do, and that rising prices fund the expansion of what we know. For two centuries that has been the more accurate picture of the world, and the people who bet against it have lost, deservedly, nearly every time.
Now watch what actually happened to the whales.
Kerosene replaced whale oil for lighting in the 1860s. The whales were not saved. Whaling continued — for lubricants, for margarine, for the baleen in corsets — and in the twentieth century it became something the nineteenth could not have imagined. Steam and then diesel factory ships. The explosive harpoon. Spotter aircraft, and later sonar. Across the twentieth century industrial whaling killed on the order of three million whales, and it drove the blue whale, the fin, the right, and the humpback to the edge of extinction — not instead of the substitution but alongside it, decades after the price signal had supposedly done its work. The blue whale was reduced by something close to ninety-nine percent. What finally stopped it was not a price. It was the International Whaling Commission’s moratorium — agreed in 1982, in force from the 1985–86 season: a hard cap, imposed politically, on how many whales could be killed at all.
So the honest version of the story is stranger than either side usually tells it. The price system did find the substitute — and the whales were very nearly exterminated anyway, because the substitution only ever governed one use, and every other use met a technology that made killing whales cheaper faster than scarcity made it dear. Simon’s mechanism worked. It simply was not the thing standing between the whale and the harpoon. The thing standing there was cost — the cost and difficulty and time of the killing itself — and when the technology of killing improved, that cost fell away, and the mechanism had nothing left to hold.
This is the variable the cornucopian account leaves out, and it has been safe to leave out for the whole history of industrial capitalism, which is why leaving it out has worked. Simon’s mechanism has a hidden dependency: it needs the marginal unit of the vanishing resource to stay expensive long enough for the substitute to catch it on price. And for two hundred years the thing that kept the marginal unit expensive was labor. Extraction was slow because people did it, or because people had to build and run and feed the machines that did. The last barrels, the last seams, the last herds cost more to take because taking them was hard — and that hardness bought time. Time for the kerosene to get cheap, the shale to get crackable, the panel to get efficient. That friction was the system’s margin of safety — the reason the price could rise gently enough to redirect us before the resource was gone.
Remove the labor and you remove the friction.
That is what the transition to artificial intelligence and robotics is, seen from this one angle: the near-total removal of the human labor cost of doing physical things. When a fleet of autonomous machines can find, take, and process a resource at a cost that falls toward the cost of the energy alone, the marginal unit stops being expensive. The last of something no longer costs meaningfully more to take than the first. And the moment that is true, Simon’s mechanism loses the one input it silently depended on. The price of the vanishing thing still rises — the signal fires exactly as the theory says it will — but the physics now finishes before the economics does. The substitute is still on its way. The resource is already gone. We take the last whale while inventing the thing that would have made whaling pointless, because with tireless machines the last whale costs almost nothing to take.
I do not think this refutes Hayek. I think it follows straight from him. Hayek’s insight was never that markets are magic; it was that prices are information — that they carry knowledge no planner could assemble, and that the system works because the information is true. His entire case against central planning was an argument about the integrity of a signal. What the robot transition does is corrupt that signal at one specific joint: it lets the price of a physically finite thing be driven by a cost that is racing toward zero, so the number stops tracking the scarcity it is supposed to encode. The gauge still moves. It has merely come loose from the tank. And a Hayekian, of all people, should be the first to mind, because the objection here is not that the outcome is ugly. The objection is that the price has begun to lie — and a lying price is the thing Hayek spent his life warning us about.
Here I part company with my own side, and with the other one too, and it is worth being exact about both.
I am not making the limits-to-growth argument. The Club of Rome was wrong in 1972 for the same reason Ehrlich was wrong in 1980: they mistook a few hard physical bottlenecks for a general ceiling on prosperity, and the general ceiling never came, because almost nothing that matters to human flourishing is physically fixed. The robot transition makes that more true, not less. It is about to make manufactured goods, energy conversion, services, and information radically more abundant — an explosion of precisely the growth the cornucopians promised. Degrowth is the wrong lesson, and it is wrong precisely because it commits Ehrlich’s error in a hair shirt. The accurate picture is that growth is about to run nearly unlimited on everything except the small set of inputs that are actually, physically finite — the last of the easy hydrocarbons, an aquifer’s recharge, the stock of a fishery, a hectare of ground that exists only once — and that the whole task is to let the abundance run at full speed while telling the truth about that small hard set, so the growth routes around the bottlenecks instead of through the last whale.
That is not a job for a planner. A planner cannot do it; that is Hayek’s point, and it still holds. It is a job for a price — but a price denominated in a unit the abundance cannot corrupt. This is the whole of what Free Market Ecology proposes, and to this audience it goes in a paragraph. Give each physically finite resource a unit of its own — an oil unit, a water unit, a fishery unit, none convertible into another, none that can be printed, discounted, or bought down by cheap labor. Make the unit a debt rather than a permit: a producer borrows the rights his extraction will use and works the debt off down the supply chain, where a tradeable permit would be an asset to buy and hoard. That difference is the whole of what separates this from cap-and-trade, and it is why the scheme conserves instead of financializes. Then leave the allocation alone — the market decides who gets what exactly as it does now; only the financing of extraction changes. The quantity of the rights is set to what is physically there, the number machine efficiency cannot forge: no fleet, however cheap, makes another hectare of ground or another barrel of the easy oil that is running out. So the price of the right rises without bound as the stock falls, and extraction stops when the borrowed rights run out, however cheap the killing has become. The substitute still gets invented. Capital still chases the markup. The frugal still sell their unused share to the extravagant. Nothing about the price system is repealed. A handful of prices — one per finite dimension, each in its own unit — are made honest, at the exact moment the machinery would otherwise make them lie.
There is a version of environmentalism that wants to slow the machines down. This is not that. Slowing the machines down is both impossible and undesirable; the abundance is real and worth having, and no one who has watched the last two centuries should care to bet against it. The claim here is narrower, and I think harder to wave off: the same efficiency that delivers the abundance also, for the first time, strips out the friction that used to keep our prices honest about the few things that are genuinely finite — so the abundance and the honesty now have to be engineered together, or the abundance arrives having quietly stepped over the last of something on its way in.
Simon won his bet, and were he alive I suspect he would take this one too, and win it again, because the metals will keep getting cheaper and the doomsayers will keep mistaking a bottleneck for a ceiling. But there is a bet he would lose, and it is the one in front of us. Bet him that a fleet of tireless machines, set to take a physically finite thing and given no signal that cannot be gamed by their own efficiency, will leave the last of it standing. That is the bet the whales already lost, in the century after the substitute arrived, when the thing that saved them in the end was a hard cap the price system never produced on its own. The wall is there for the oil and the water and the soil as it was there for the whales; the stock is finite whatever anyone believes about it. What is not fixed is how we meet it. We can meet it with a price that tells the truth about the scarcity, or with a planner who pretends to know what the price would have said, or the way the whales nearly met it — with a fleet that takes the last of them because nothing in the ledger says stop. Only the first keeps the resource and the market both.
This is the Austrian-facing companion to The Number Nobody Chooses, which runs the same regime change from the other side — what the end of wage labor does to the distribution of claims. Together they are one argument: that the machinery of abundance changes what our prices must be honest about, on the resource side and the human side at once.