Smart people keep underestimating technology. Exponentials break intuition, and that same error bends how people read the economy.
The paper fold
Take an ordinary sheet of paper, about a tenth of a millimetre thick. Fold it in half. Fold it again, and keep going. Ask a room of clever people how thick the stack is after 50 folds, and most will guess something you could hold, a few inches, maybe a metre. The stack reaches most of the way to the sun.
The folds in between show why intuition breaks. After 20 folds the stack is about 100 metres, the height of a tall building. After 30 folds it’s past the edge of space. After 42 it’s past the moon. At 50 you’re most of the way there. So at which fold are you halfway? Fold 49. One fold from the end. Each fold doubles everything that came before, so the last few steps do almost all the work. Nothing about the process ever changed. What changed is where you’re standing on it.
So smart people keep underestimating technology even with the information in hand. They misjudge it because intuition measures change in equal steps, and compounding change doesn’t arrive in equal steps. It arrives in doublings. A doubling feels like nothing for a long time, and then it feels like everything at once.
I put technology at around fold 33 or 34. Everything you’ve ever seen computers do, the internet, the smartphone, the AI you’ve used this year, all of it is fold 33. Fold 33 is about 860 kilometres of paper out of a hundred million. And the next fold adds more than all the previous folds combined. That’s what doubling means, and it holds whatever any particular gadget turns out to do. On top of that, the doubling itself is getting faster, because the tools now help build the next tools. Capability jumps that used to take decades now take years, then quarters.
Why being smart doesn’t save you
Four reasons.
First, the failure is perceptual, not intellectual. Even people who know the maths revert to straight-line forecasts the moment they stop consciously correcting. It works like an optical illusion. Knowing it’s an illusion doesn’t make you see it straight.
Second, experience makes it worse, not better. An expert’s instincts were trained on the flat part of the curve. Blockbuster ran the video business better than anyone, and it answered streaming by adding sweet stands to its shops, because everything it had learned said the store network wins. When download speeds crossed the line, that store network flipped from its biggest asset into its biggest cost, almost overnight. BlackBerry knew phones and bet on the keyboard. The more success you’ve had, the better calibrated you are to a world the curve has already left. That’s why beginners often spot the shift before veterans do.
Third, people measure the system from inside the system. The dashboards a smart person checks were all built for the old model. GDP [the total spent on final goods and services], CPI [the official basket used to track consumer prices], wage growth, all of them move when policy moves. If your instruments are wired into the old machine, they’ll keep reporting that the old machine is fine.
Fourth, incentives. A central banker who let prices fall to their natural level would preside over cascading defaults and lose the job. An analyst whose model says the future is the past plus 3% has fewer awkward meetings. Every one of those people can be doing their job well. The structure selects for the linear read.
There’s a second layer underneath all four. Governments and central banks create new money and credit to stop prices from falling, which deletes technology’s fingerprints from the one place people look for them, prices. So the people most fluent in the standard data are often the most confidently wrong about what’s happening underneath.
What the error does to how people read the economy
When a business learns to make the same thing with fewer inputs, competition hands the saving to the customer, because a rival who cuts the price takes the customer from the one who doesn’t. So the natural direction of prices in a technology economy is down. If the technology compounds, the downward pull on prices compounds with it.
But nearly all of our money is created through lending, and debts are fixed in pound amounts. If prices and wages broadly fall, loan repayments don’t fall with them, so the debt takes a bigger bite of every pay packet and every business’s takings until defaults cascade through the banks. A system carrying that much debt can’t allow prices to fall. So governments and central banks create new money and credit to push prices back up.
If the deflationary force doubles, the money creation needed to cancel it must double too. That’s why every rescue is bigger than the last, hundreds of billions in 2008, trillions in 2020. It’s why the world added roughly 185 trillion dollars of new debt in the two decades before the pandemic to buy about 46 trillion of growth, and why the ratio has worsened since. Smart people keep calling that unsustainable and waiting for normalisation. It can’t normalise, because the thing it’s offsetting is still doubling.
From there the misreadings name themselves.
1.Inflation is measured from the wrong baseline
Picture a moving walkway sliding backwards five steps a minute. That backwards pull is technology, taking your cost of living down. To hold prices level, the money side has to walk forward five steps a minute. To push them up, it has to jog. If technology would have taken your cost of living down 3% this year and the index shows plus 2%, roughly 5% of your purchasing power was taken from you. The official record calls it “price stability”.
2.The evidence erases itself
Because new money cancels the fall in sticker prices, a smart person checking price data concludes technology’s effect must be small. The intervention hides the very force that made the intervention necessary. Meanwhile the deflation shows plainly wherever policy has least reach, in the TV, the software, the phone that swallowed your camera, sat nav, torch, and stereo for near nothing.
3.Rising asset prices get read as new wealth
A house that doubles in pounds is mostly the ruler shrinking, not the house improving. You still need somewhere to live, and the next house, the taxes, and the insurance all rise with it.
4.GDP gets read as progress
GDP counts spending. When a £100 chair becomes a £10 chair, or a paid product becomes free, your life got better and measured output went down. The better technology gets at giving you more for less, the worse the economy looks through the official lens.
5.Economists end up teaching that 2% inflation is health
It’s true for the debt structure, which dies without rising prices, and false for the people inside it, who would be better off with falling ones. The system’s survival need gets rebranded as an economic law.
6.The public debate collapses into “inflation or deflation?”
The debate also becomes a hunt for which snowflake will trigger the avalanche. Both miss it. The instability is baked into the design, an exponential force pressing on a system that must never let prices fall.
If the fold is real
Interventions must keep growing, because the gap they’re papering over doubles, and that arithmetic holds whatever the officials running them intend.
The surprise stays structural, so each new capability and each new rescue will keep feeling sudden to linear minds. “Gradually, then suddenly” is what a doubling looks like from inside.
And you can’t fix the reading by squinting harder at the old dashboards. You fix it in two moves. Measure inflation against where prices would have gone, not against zero. And check your measurements in a unit nobody can expand. That’s the work bitcoin does in this thesis. It’s a ruler, not a trade. Priced in a fixed unit, the fall in prices technology has been delivering all along finally shows up. The same linear habit that missed the smartphone is now pricing AI and bitcoin as if the curve were finished.
Reasons to think the fold is wrong
“Exponentials always flatten. Moore’s law is slowing.”
True for any single technology. Each one follows an S-curve [growth that starts slow, accelerates, then levels off as the technology matures]. But new curves stack on the old ones, and aggregate compute, models, and software keep compounding even as the chip-shrinking behind Moore’s law slows. Pointing at one maturing curve and declaring the whole thing finished is exactly how people missed the internet and the smartphone.
“Why fold 33 and not fold 20?”
It’s an estimate. Count doublings in computing capability from the first computer chips in the late 1950s, at an average of about one every two years, and you land in the low thirties today. Put it at 30 or at 36 and the argument above doesn’t change. The next fold still adds more than every fold before it, and the people forecasting it are still drawing straight lines. The number is there to make the scale visible. The argument rests on the doubling.
“Experts understand this and plan for it.”
The record says otherwise. Blockbuster, Kodak, BlackBerry, and the “this time rates normalise” cycles since 2008, in which markets buckle at the hint that borrowing will cost more, and policy reverses. Awareness of the bias doesn’t remove it. Only constant discipline does, and that discipline is rare.
“But the prices I pay keep rising, so the deflation story must be wrong.”
That’s the walkway again. What you feel is the money side winning in the sticker prices while your wages lag. Where the money system’s push is weakest, prices have been collapsing for decades. Two forces, one visible sum.
“If change were really this fast, we’d all see it coming.”
You don’t feel early doublings. Doubling almost nothing still leaves you with almost nothing, so for a long stretch nothing seems to be happening. By the time a doubling is big enough to feel, it has already landed, and the ones after it are larger still. So you notice the change late, and almost all of it is still ahead of you. The shape does that to everyone inside it.