cryptopath on Nostr: Too Cheap to Meter Light went from a rationed luxury to something you never think ...

Too Cheap to Meter

Light went from a rationed luxury to something you never think about — thousands of times cheaper, in hours of your life. Intelligence is about to make the same journey. The only question that matters is whether you'll get to keep the windfall, or whether it'll be printed away before it reaches your hands.

To read after dark in 1800, you worked for it. A tallow candle cost a real slice of a day's wage and gave you one dim, smoky, flickering hour. "Burning the midnight oil" was a line item, not a metaphor. Light was rationed — you spent it the way you'd spend anything scarce, candle by candle.

The economist William Nordhaus measured what happened next, and found that the official price statistics had missed almost all of it, because they tracked the price of candles instead of the price of light. Measured properly — in hours of human labor per unit of actual illumination — light has collapsed in price by a factor of thousands. An hour of your work today buys a quantity of light a Babylonian couldn't have bought in a lifetime. Flip a switch; never think about it. Light stopped being a resource you ration and became an ambient, invisible, functionally unmetered condition of being alive.

That phrase — "too cheap to meter" — is worth pausing on, because it's a warning disguised as a promise. It was coined in 1954 by Lewis Strauss, head of the US Atomic Energy Commission, predicting that nuclear electricity would soon be too cheap to meter. It became famous precisely because it was wrong. Nuclear never got that cheap. So the slogan we reach for to describe glorious abundance is, historically, a monument to overconfidence about abundance. Carry that scar into everything that follows.

Because intelligence is next in line to make light's journey — and whether that turns out to be a gift to you or a gift to someone else comes down to a collision between two economic forces that most people never see fighting.

The two forces

Here are the forces. They push in opposite directions, and almost everything about your economic life is decided by which one wins.

The first is doing more with less. This is the deep fact of economic history productivity, the steady grind of producing more output from fewer hours, less material, less energy. Technology is, at its root, a deflationary force: it lowers the real cost of making things. Left completely alone, progress means prices fall over time. The collapse in the price of light is this fact in its purest visible form — thousands of times more illumination per hour of work.

The second is money expansion. Print more units against the same pile of goods and each unit holds less; purchasing power deteriorates; prices drift up. It's not a stopwatch — where the new money lands matters, timing matters, it shows up unevenly — but over any real stretch of time it's gravity. Expand the money, and you erode what it holds.

Now put them in one room. Productivity pulls prices down. Money expansion pushes them up. And a flat, or gently rising, price level does not mean nothing is happening. It means the printing has been tuned to exactly cancel the deflation you earned.

The flat line is the crime scene

Sit with that, because it's the whole argument.

When a central bank aims for "two percent inflation" in an economy that does more with less every year, what it is actually doing is printing fast enough to erase the price fall that progress would otherwise have handed you. You got more efficient. The dividend of that efficiency — cheaper everything, a little more purchasing power every year for simply holding your money — was owed to everyone. Instead it was printed away and delivered, first and in full, to whoever stands nearest the new money.

The stable price line isn't the absence of a transfer. It's the evidence of one. The gift of progress arrived at the door and was quietly intercepted. These are the cheaper apples you never see — except now you can see why you never see them: the deflation they'd have caused was cancelled on purpose.

Why light was the lucky case

Light mostly escaped this, and it's worth knowing why, because it's exactly what might not repeat.

Cheap light became a near-universal commons. The windfall reached almost everyone, and it didn't arrive as "cheaper candles" — it arrived as a restructuring of life. Cheap light decoupled human activity from the sun: the night shift, the 24-hour city, indoor work in winter, safety after dark, the whole second half of the clock opened up for living. The chandler of 1800 could not have pictured a floodlit stadium. The biggest effects were second-order and unforeseeable, and they landed broadly, because light was competitive, un-ownable, and arrived before the printing press had learned to cancel everything.

That last part is the catch. Intelligence, and who gets to keep it

If artificial intelligence is what its builders claim, it is the largest doing-more-with-less event in history — the biggest deflationary force ever created. Which, under a money that fears deflation above all things, sets up the grimmest default outcome imaginable: the response will be to print against it, hard, to stop prices from falling. And so the entire intelligence windfall gets absorbed into the monetary machine and never reaches you as a cheaper, freer life. The abundance will be completely real. Your share of it will have been confiscated on the way in, by the force that deteriorates price in the other direction.

And intelligence has a second problem light didn't. Light was un-ownable. Intelligence is concentrated — a few labs own the models. So even the part of the windfall that isn't printed away may not fall into your hands as a commons; it may arrive as a rented utility, free at the point of use because you're the product or because the dependence is the business. Cheap light made everyone richer. Cheap intelligence could just as easily make a few owners richer and leave everyone else renting their own minds back by the month.

So the fork is sharp, and it's the only thing worth watching. Does the abundance arrive as falling prices and freed time for everyone — or as a metered utility whose gains accrue to whoever owns the model and whoever owns the printer? Same "doing more with less." Opposite recipient.

What stays scarce

One consolation, and it's a real one. When a thing gets infinitely cheap, the scarcity doesn't vanish — it moves. When light got cheap, light stopped being valuable and what you did in the light became everything; darkness itself turned into a luxury. When intelligence gets cheap, thinking stops being the scarce thing and the complements become precious: judgment about what's worth thinking about, taste, the question rather than the answer, and above all trust — because cheap light was obviously light, but cheap intelligence can be confidently, invisibly wrong, and a wrong answer looks exactly like a right one. Verification becomes the rarest skill in the building. The value runs to the things the machine can't supply: the goal, the responsibility, the meaning, the being-there. Guard those. They're what you trade the freed attention for.

The sound-money fork

The Bitcoin frame falls straight out of the two forces, and it needs no slogan.

A money that cannot be expanded lets the deflation happen. It lets "doing more with less" show up honestly, as rising purchasing power — progress paying its dividend to everyone who simply holds the money, your savings buying more each cycle because the world got more efficient and nobody printed the gain away. That's the shrinking-coin thesis stated as pure economics. A money that can be expanded pays the identical dividend to whoever prints. Same productivity. Same abundance. The only variable is who ends up holding it — and that variable is decided entirely by whether the money can be debased.

So "sound money" and "you get to keep the fruits of progress" turn out to be the same sentence. In a world of honest money, the intelligence windfall reaches you by default. In a world of printed money, reaching you is the exception that has to fight the printing press — and usually loses.

The honest edge

Keep the blade sharp, because the orthodoxy this argues against isn't pure propaganda.

The fear of deflation points at something real — but it conflates two opposite diseases wearing the same symptom. Good deflation is prices falling because we produce more efficiently; the late-19th-century US lived through decades of gently falling prices alongside strong growth, and it was fine. Bad deflation is prices falling because money and credit are collapsing — the 1930s debt spiral, where falling prices and vanishing money feed each other down. Those are not the same thing, and the sleight of hand in "deflation is dangerous" is using the genuine terror of the second to justify always printing — which also happens to erase the benign, earned, first kind. The precise claim is not "all falling prices are good." It's: the deflation that comes from doing more with less is the dividend of progress, and printing to cancel it is theft of that dividend — and it is not the debt spiral they frighten you with, however hard the two are sold as one.

And the rest of the honesty: money-to-prices is a gravity, not a stopwatch, so don't promise anyone a mechanical link. "Too cheap to meter" is a famous broken promise, so hold your optimism about AI abundance loosely. "Unmetered at the point of use" almost always means "metered somewhere else by someone who owns the meter." And the whole dream runs through energy — intelligence gets cheap only if compute scales, which means energy scales, and AI's own firm, inflexible demand may keep energy dear. Cheap intelligence is a bet, not a law, and it's the same bet as cheap energy, which is exactly the contested thing.

The one line to keep

Progress does more with less — that's the fact. Printing deteriorates money — that's the other fact. And a flat price level is simply the second fact cancelling the first, so the dividend you earned gets handed to whoever's nearest the new money instead of to you. AI is about to run that collision at the largest scale in history.

When light got cheap, the gift was never the light. It was the night — everything we chose to do with the hours the sun used to forbid. When intelligence gets cheap, the gift won't be the intelligence either. It'll be whatever we do with the attention it frees — if we keep the judgment to choose well, and if the windfall is allowed to reach us instead of being metered and printed away at the door.

Guard the complements the machine can't give you. And hold the one money that lets you keep the fruits of a world doing more with less — because under everything else, that fruit has your name on it, and under printed money, it always ends up with someone else's.