Abundance Moves Scarcity

Chasing the light

To live through the night, you needed fire. Step away from the campfire and you were in the dark. Light was bound to fuel and to place. What was scarce was light itself — keeping the wood burning, staying close to the flame.

Then light came loose from fire and could be carried. The tallow candle. Put up with soot and smell, and you could take the dark with you. But fat was dear. Even in 1800, an hour of reading light burned a real fraction of a day's wages. People rationed light. Scarcity had moved from light to fuel, and the time it bought.

Gas lighting arrived; light rode the pipes and became a city. Streets, theaters, factories lit up. For the first time the night became usable time. Factories began to run after dark. Scarcity left fuel behind and moved to the grid — who was connected — and to the night itself. The night for sleeping and the night for working began to fight.

With the incandescent bulb, light cut free of flame entirely. Clean, safe, instant. Homes filled with it, and people stopped thinking about light at all. Scarcity moved to electricity — light was now just one draw on the wires — and to darkness. What became ordinary was light; what stopped being ordinary was the dark.

The fluorescent tube made light cheap per lumen and filled every room with it — the flat, indifferent glow of the office. The closer light came to free, the more precious good light and restful dark became. Then the LED. Lumens went to almost nothing; we lit everything, and paved the night with screens. Light leaked out, and a third of humanity can no longer see the Milky Way. Something, somewhere, is always glowing. By this point the scarce thing is no longer light. It is turning it off — unlit time, a night without notifications.

From fire to LED, what we made cheap was "light." But each time, scarcity walked next door — to fuel → power → the night → attention. Follow the light, and you can watch scarcity move. That is the whole of this essay. Scarcity doesn't vanish. It moves next door.

Scarcity is conserved

We usually tell progress in one direction: "scarcity → abundance." But what we actually observe is "scarcity → (abundance) → new scarcity." Scarcity isn't a property of a thing. It's a property of the system's binding constraint. Make one thing abundant, and the constraint moves to its complement. A complement is the partner good — what runs short in order to use whatever just became plentiful. When light flooded in, what ran short was the attention to spend on it. Anything cheap always has an expensive partner clinging to it, the thing you need to make use of it. That partner is the complement.

In 1865 Jevons noted something strange: when the steam engine learned to burn coal efficiently, coal use didn't fall — it rose. Efficiency unleashed demand, and scarcity moved to coal itself, then on to organization and management. Scarcity doesn't end. It only moves — and once you suspect that, you see the same face everywhere.

It happened with food, too. For most of history the scarce thing was calories themselves. Famine toppled dynasties and decided wars. In the twentieth century the Green Revolution — nitrogen fertilizer and better seed — flipped it over. For the first time in history, in rich countries, calories were in surplus. And scarcity moved from calories to restraint. Not eating, choosing, keeping your metabolism in order. A body optimized for famine is fragile in plenty. What now eats away at affluent nations isn't lack but its opposite. Cheap is sugar and fat; dear is the will to refuse them, the cooked meal, the healthy hour.

And the newest abundance has the same shape. Information (Simon, 1971): the moment it overflowed, "a wealth of information creates a poverty of attention" — scarcity again moved to attention. Sequencing (NHGRI): the cost of reading one genome fell from about $95 million in 2001 to a few hundred dollars by the 2020s — faster than Moore's law. But the cheaper reading got, the more scarcity moved to meaning — what the sequence actually says. Light, food, information: all walking with the same gait. FIG.1 puts the repetition on a single page. The collapsing curves all look alike. That likeness is no accident. It's because scarcity is conserved. Each time a line hits the floor, something in the right-hand column lights up, newly scarce.

Why it moves to the complement

The very word "limiting" comes from agronomy. Liebig's law of the minimum: a field's yield is set by whichever nutrient is scarcest. Add all the nitrogen you like; if what's missing is phosphorus, phosphorus is the limit. A system's output is bound by its rarest input — so that's where the surplus pools. Value flows to the binding constraint. Pile up any non-limiting input and output won't budge. Baumol pointed at the flip side: as manufactured goods get cheaper, the relative price of what resists automation — care, education, live human time — keeps rising.

The root reason comes down to one line. Abundance reveals what couldn't be made abundant. That is what migration really is. When strategy says "commoditize your complement," it's the same force wearing a different face. And the destination has a habit — it goes to what doesn't scale. Attention, capped at 24 hours. Judgment and verification, which matter more the freer generation gets. Trust and provenance, dearer the easier forgery becomes. Restraint, in an age of excess. Coordination, more valuable the more each individual is amplified. And, at the far end of the bits, energy and matter.

When scarcity doesn't move

Here, honestly, let me doubt the law. Does scarcity always move? Not always. Sometimes it vanishes. Smallpox did. Vaccination and eradication didn't relocate "the scarcity of immunity" — they ended it. The disease itself left the earth. Scarcity migrates only when the complement can't scale. If the complement can also be made abundant, scarcity doesn't move — it shrinks.

So the question reduces to one: the complement of whatever is now becoming abundant — does it scale, or not? Malthus said scarcity wins; Boserup said scarcity is precisely what breeds invention. Which is right isn't a fixed fate; it's a moving answer, decided by how far the complement can stretch. Attention looks like the "terminus" exactly because it doesn't scale. A day is 24 hours and can't be enlarged. Where the complement stretches least, scarcity finally settles.

So the law isn't iron. "Scarcity usually moves. Sometimes it vanishes." — and on that boundary lies the question of how far the complement can grow. To guess where scarcity goes next, look at whatever is becoming abundant and ask whether its complement can stretch. And right now, the largest abundance in human history is arriving.

When intelligence becomes free

Everything so far was a warm-up. What's arriving now is neither light nor food — it's the abundance of cognition itself. The power to think at a human level (AGI), made cheap and endlessly copyable. In time, beyond human (ASI). The scarcest of all scarce things in history — thinking — becomes something that comes out of a tap.

AGI and ASI are two stages of the same flood. AGI makes human-level intelligence free — "who does it" changes. ASI takes it past human and begins to set its own direction — "what is possible" changes. Even the first moves scarcity hugely. The second shoves it to the complement by an order no past migration matched. Because intelligence was the most expensive resource of all, the rebound when it goes free is the largest of all.

The public asks, "Will AI take jobs?" But that only asks where the old scarcity — cognitive labor — goes. The law asks the opposite question. When cognition becomes free, what is its complement? Where does the new scarcity appear? Worrying about employment is staring at the collapsing curve. What to watch is what lights up, after the line hits the floor, in the right-hand column.

Every past migration points at the shape of the answer. Scarcity moves to what intelligence can't stand in for. And that splits in two directions. One is down — the physical floor that no amount of cleverness escapes. The other is up — the human ceiling that no flood of answers can fill. We go down first, then climb.

The physical floor (energy & atoms)

Scarcity heads down first, back to the oldest place.

Intelligence isn't free of thermodynamics. Every thought is, in the end, the flipping of a bit, and flipping costs energy and sheds heat (Landauer's principle). One inference is tiny. But run cognition at civilization scale and the constraint becomes joules, cooling, the grid, and land. Even the much-discussed "intelligence explosion" — AI improving AI, faster and faster — can't escape this floor. A design a thousand times smarter overnight still has to be built, in steel and power and fabs and years. Atoms move at the speed of construction, not of inference.

And work that touches the world stays bound to physics. Bits are infinite; atoms are finite. Making things, growing cells, raising bridges — even as intelligence overflows, this stays jammed. It needs hands, machines, materials. The more the mind floods, the more the body of civilization — the power to light it, the water to cool it, the land to put it on, the hands to actually move it — becomes the limit.

History has rehearsed this many times. Steam raised coal, electrification raised generation and the grid, into a master resource. General-purpose technologies always push scarcity into their complement: energy and infrastructure. Intelligence is no different. The master resource of a civilization where thinking is free becomes, once again — and probably more urgently than ever — energy. Who can move electrons cheaply, cleanly, and in vast amounts: that is the oldest limit on the smartest future.

Truth and trust

Scarcity heads up next, toward judgment.

When generation is infinite and free — of text, images, proofs, even personas — producing something plausible costs nothing. So scarcity flips. What's dear is not making but telling which is true. Generation and verification were always asymmetric; checking is harder than writing. In a world of infinite generation, the capacity to verify becomes the constraint. The power to find signal in an ocean of plausible lies, the eye that picks the real — that is the new scarcity.

And in a world where anything can be forged, being real becomes precious. Provenance, attestation, the assurance that "this was truly made by this person, at this time." Trust — between people, in records, in institutions — becomes the currency. In a world where fakes are free, only the "verified real" holds value. Trust was treated like air. Only when the air thins do we notice it was precious.

And one level up from the individual, the same thing happens to society. When anyone can mass-produce plausible reality for free, shared reality itself turns scarce. That everyone can start from the same facts — and the institutions and trusted records that hold that up. The "common ground" we treated as air turns out, once thinned, to have been the foundation of civilization. In a world where truth wobbles cheaply, the dearest infrastructure is that we can all stand on the same ground.

Will and meaning (and ASI)

Now to the deepest place — where superintelligence (ASI) sharpens everything.

AGI gives answers. It does not give the question. The freer and more superhuman optimization becomes, the scarcer the objective — what to aim it at, what to want. ASI maximizes whatever you point it at. So pointing becomes the scarcest, most dangerous, most human act of all. Will. Taste. "What for." The more answers go to infinity, the more the question becomes king.

And setting that aim correctly becomes the most expensive scarcity of all. A superintelligence maximizes the goal you give it with a relentlessness no human imagined. Misstate the goal even slightly, and all that force pours in the wrong direction. The power to decide what to maximize with care — to translate values into words and constraints — becomes a scarcity on which civilization's survival turns. In a world where optimization is free, the dearest thing is a correct objective function.

And who may set that aim becomes its own scarcity — legitimacy. When capability gets cheap, the scarce thing is the right to use it: who may decide, who may act. Power was never just capability. Now that capability is free, power becomes, almost entirely, legitimacy — not what you can do, but what everyone agrees you may do.

And at the level of the group, scarcity moves to agreement. In a world where everyone commands a genius, what's precious is that we can face the same direction — coordination, and the will to choose together. The stronger each individual becomes, the scarcer "we can decide together" gets. More power makes alignment harder.

So in an ASI world, scarcity migrates almost entirely to two poles. The lower pole: energy and matter. The upper pole: judgment, will, and meaning. And the middle — the vast band we once called "knowledge work," even called "intelligence" — collapses toward free. Value, power, and meaning are all drawn to the poles.

Not those who made something abundant, but those who hold a pole, take the top — the law from the opening returns, one last time, at the largest scale. And at the very top of the upper pole sits the oldest human question, made scarce at last by abundance: not "can we," but "should we, and what for."

So what to cultivate

So what, holding this map, should we do? The law itself points the way. Each time abundance makes something free, cultivate its complement — the side that doesn't scale.

The faster things change, the scarcer the time to adapt. Technology turns over overnight, but law and institutions and culture and the human heart, like concrete, take time to set. Cleverness is in surplus; adaptation is short. So the power to take change in, drop what's stale, and recompose — the speed of relearning — grows dearer than ever.

At the level of society, what to cultivate is on the same side. In a world where capability reaches everyone, the precious thing is a legitimate way to decide. Who may move what, by what process. To get things done by agreement and trust rather than force. As the individual cultivates judgment and will, society can only cultivate coordination and legitimacy — infrastructure that doesn't scale, that only humans can build.

And the more machine output overflows, the more the fact that a real human actually cared becomes a premium. Something made by hand. Being there. A real relationship. In a world where AI can write anything, what's precious is the single fact that someone wrote it for you.

So what's asked of us in the coming age is no longer "to produce answers" — machines do that. What's asked is to pose questions, to discern, to choose, to trust, to care. From answerers to askers. From makers to choosers. Wherever value moved is exactly the work left to human beings.

What I drew

In 1930 Keynes wrote that once the economic problem is solved, humanity will face, for the first time, the real problem — how to live. He wrote it half in dread. But line up the record of migration, and it reads less like dread than a schedule.

Each time abundance advances a step, scarcity is promoted a step toward the human core. From quantity to quality. To truth. To will. And to meaning. The flood of intelligence is the final exam of that promotion. The most abundant future makes the most human things scarce. That isn't a defeat. It only means the real frontier has, at last, come into view.

Since the night we first made fire cheap, we have been doing one thing: by making things abundant, we slowly reveal what truly matters. What gets revealed next is — probably — ourselves.