The Fold

Smart people keep underestimating technology. Anything that keeps doubling is almost impossible to picture. And technology makes things cheaper to make, but the fall in prices that should follow never reaches your overall cost of living, because governments and central banks create new money and credit to stop prices from falling. So the overall figures carry no trace of the doubling, and the people who read those figures most closely end up the most confident and the most wrong.

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 along the way show where the guess goes wrong. After 20 folds the stack is about 100 metres, the height of a tall building. After 30 it’s past the edge of space. After 42 it’s past the moon.¹ Fold 50 gets you most of the way to the sun, and fold 49 gets you half that. Every fold doubles the whole stack, so the last step covers as much ground as all the steps before it put together, and the first 40 barely register. The folding never speeds up. You just reach the part where each step is enormous.

That’s the mistake people make about technology. We judge change by adding. If something moved this far last year, we expect it to move about that far again next year. Doubling doesn’t work like that. For a long stretch there’s almost nothing to see, because doubling a tiny number still leaves a tiny number. Then it stops being tiny, and every step after that is huge.

Count the doublings in computing as a whole, not the transistors on one chip. Computing has doubled about once every two years on average since the first chips in the late 1950s.² That’s 33 or 34 doublings. Call it fold 33. Everything you’ve ever seen a computer do, the internet, the smartphone, the AI you used this year, all of it is fold 33, about 860 kilometres of paper. The next fold on its own adds as much as the whole stack underneath it. Put the count at 30 or at 36 and the argument doesn’t change. And the doubling is speeding up, because the tools we build now help build the next ones. In AI, what once took decades to get twice as good now takes a few years, and lately a few months.³

Any one technology does level off in the end. Each follows an S-curve, slow at first, then fast, then flat. The shrinking of chips behind Moore’s law really is slowing. But a new one often starts before the old one flattens, and what matters is the total. Processing power, AI models, and software together are still doubling.⁴ Pointing at one technology that has matured and declaring the doubling over is judging by adding again. Judging by adding is how people missed the internet, and then the smartphone.

Why being smart doesn’t save you

Four reasons.

First, knowing the maths isn’t enough. You can follow the doubling perfectly well and still find your mind sliding back to straight lines. The arithmetic is right, and it’s still hard to believe that 50 folds gets you most of the way to the sun.

Second, experience makes it worse. An expert’s instincts were trained on the years when change was slow, and those are the years that no longer apply. Take Blockbuster for example. They ran the biggest video rental business in the world. When Netflix began posting discs out, and then streaming them, Blockbuster answered by putting sweet stands in its shops, because everything it had learned said a big network of stores wins. Home internet speed had been doubling every couple of years, far too slow to carry a film for most of that time, but once it was fast enough those shops turned from an asset into a cost almost overnight.⁵ The more success you’ve had, the more your instincts tell you to do what used to work.

Third, you can’t just look it up. The numbers a smart person would check tell you the wrong thing. GDP is how governments judge the economy, CPI tracks what a basket of goods costs, and wage growth tracks pay. All three are built so that a falling number means trouble. Technology makes things cheaper to make, so it pushes all three down.

Fourth, the incentives point the same way. A central banker who stood back the next time a crisis threatened would watch borrowers default in every direction, and would be replaced by someone promising to stop it. An analyst whose model says next year looks like this year plus a few percent has fewer awkward meetings. Every one of those people can be doing their job well. Nobody has to be wrong on purpose for a whole profession to keep reading straight lines.

What the error does to how people read the economy

Now connect the fold to the till.

When technology lets a business make the same thing with less work and less material, and a rival can copy the cheaper way, competition hands the saving to the customer, because the rival who drops the price takes the customer from the one who doesn’t. Where nobody can copy it, the saving stays with the business. So where rivals can copy, technology pushes prices down, and when it keeps doubling, the push keeps doubling too.

But nearly all of our money is created when somebody borrows, and a loan is written for a fixed amount.⁶ A hundred thousand pounds owed stays a hundred thousand pounds owed, whatever happens to prices. So if prices fall broadly and nothing is done to stop it, businesses take in less, and they can afford less in wages, and pay falls with prices. The repayment doesn’t move. It takes a bigger share of every pay packet and every business’s takings. And where the loan was written against a house, the house is worth less too, so the bank stands to lose more on that loan. The system can take a slow fall that comes with better tools. In that fall it’s the cost of making things that drops, and pay drifts down slowly behind it, so the money to meet a fixed payment is still there. What it can’t take is a sudden fall in the prices credit pushed up, houses first, because then the losses spread through the banks. In that fall it’s the lending that has stopped, so incomes and the house fall together and at once, and the fixed payment is hit from both sides in the same year. So governments and central banks create new money and credit to push prices back up.⁷ The same money that holds incomes and house prices up also cancels the overall fall in shop prices.

If the push downward keeps doubling, what it takes to cancel it keeps growing. That’s why every rescue is bigger than the last, hundreds of billions of dollars of government spending and bond-buying in 2008 and trillions in 2020.⁸ It’s why the world added roughly 185 trillion dollars of new debt between 2000 and early 2018 and got about 46 trillion of growth for it, and why that ratio has got worse since.⁹ Smart people keep calling that unsustainable and waiting for it to settle back to normal. It can’t settle back, because the fall the new money exists to cancel is still doubling.

Once you see that, six familiar readings of the economy turn out to be the same mistake.

1.Inflation is measured from the wrong baseline

The honest baseline is whatever prices would do if nobody touched them. While technology keeps improving, that’s downwards. It takes new money just to hold them steady, and more to push them up. So if technology would have made your cost of living 3% cheaper this year, and the official measure shows prices up 2%, then 5% of what your money could have bought never reaches you. The official record calls that “price stability”.¹⁰

2.The evidence erases itself

New money cancels the fall in prices, so someone checking the data decides technology’s effect must be small. The prices that stayed put are the evidence that something pushed them there. Where new money reaches least, the fall is obvious, and it has shown in televisions, in software, and in the phone that swallowed your camera, your sat nav, your torch, and your stereo for almost nothing.¹¹

3.Rising asset prices get read as new wealth

When a house doubles in pounds, the house hasn’t changed. It has the same rooms and the same roof. Land is scarce and we don’t build enough, but a new shortage of land isn’t what starts a rise in house prices. They jump when borrowing gets cheaper and stall when it gets dearer, and where building is tightly limited they run further still.¹² What changed is the money the house is priced in, and cheaper borrowing is how more of it gets made. And you have to live somewhere, so selling yours just buys you another one at the same inflated price. The tax on the next move went up with the price, and so did the cost of keeping the place standing.¹³

4.GDP gets read as progress

GDP is how we measure whether the economy is healthy. But it counts paid transactions. So when a television that used to cost £1,000 now costs £300, GDP goes down. Your life gets better but the figures say the economy is worse. And when something is free, like calls over WhatsApp, there’s no transaction to count at all. The better technology gets at giving people more for less, the worse the economy looks on the official numbers.

5.Economists end up teaching that 2% inflation is health

Two percent isn’t where prices would settle on their own. While technology keeps making things cheaper to make, prices would drift down. Getting them to plus two takes new money that cancels that drift first, and then adds two on top. Central banks publish why they aim there.¹⁴ They want room to cut interest rates in a downturn, they know wages resist falling, and they say the official measure of prices overstates the rise. Those are the reasons they give, and the debts aren’t among them. The debts show up in the rescues, not in the everyday 2%. Whatever the reason, if your pay doesn’t rise as fast as prices, your real wage has fallen even though the number on the payslip hasn’t. And the saver who kept cash lost too. Two percent was chosen, and it ends up taught as a law of economics.

6.The public debate collapses into “inflation or deflation?”

People argue about whether prices will rise or fall next year. It doesn’t matter which. The push from technology keeps doubling, and the system still can’t take a sudden fall in the prices credit pushed up. That stays true whichever way the figures happen to move.

If the fold is real

If that’s right, the money created to stop prices falling has to keep getting bigger, and no change of government or central banker alters that. The bigger the push it has to cancel, the bigger the amount.

And the surprises will keep coming, because the doubling doesn’t stop. Each new thing technology can do arrives sooner than expected. So does each new rescue, the money governments and central banks create whenever a crisis threatens the debt. Both feel sudden even to the people who watched the last one.

The pound can’t show you the size of the fall, because the pound is what’s hiding it. Every price you’ve ever checked was written in a unit somebody could make more of, so you never saw how far prices would have fallen. To see past that you need to measure prices in something nobody can make more of, and until bitcoin that thing didn’t exist. Bitcoin is still being taken up as money, and as more people hold it, a bitcoin buys more. So prices in bitcoin keep falling for that reason alone, whatever happens to the cost of making things. Which leaves the fall hidden for now, and the push behind it keeps growing, because the doubling doesn’t stop.