The last two weeks, the U.S. indices made a remarkable run to all-time highs, but I keep seeing Ray Dalio promoting his new book with some very valid points while thinking back to that Citrini Research graphic — the one that shows no natural brake for the theoretical debt spiral spurred by AI job job losses. Most economists and financial reporters wrote off Citrini’s (very imaginative) report as an out-of-touch work of fiction, but E.C. Harwood and Paul Warburg and Roger Babson were all ignored ahead of 1929.

I guess the experts just assume buying a mattress on a thirty-six-month financing plan is “progress”?

The reality is the entire global economy is a bet that tomorrow will be better than today. Not in a vague, inspirational way. In a literal, mathematical, load-bearing way. The system requires it. The system was designed to require it. And for eighty years, that design has more or less worked, which is exactly the kind of track record that makes people assume it will work forever, which is exactly the kind of assumption that precedes the moment when it doesn’t.

Since the end of World War II, the global economy has run on a Keynesian engine, which is a fancy way of saying that debt is not a bug in the system — debt is the system. The whole apparatus needs two things to keep functioning: low interest rates and a growing population. Low rates keep incomes rising and inflation slow. They also make borrowing irresistible. Companies borrow. Investors borrow. Governments borrow. Everyone borrows, because why wouldn’t you? Money is cheap. The future is going to be bigger than the present, and bigger futures generate bigger returns, and bigger returns pay off today’s debts. It’s elegant, if you don’t think about it too hard. It’s terrifying, if you do.

The thing is, this arrangement works only as long as the future reliably pays for the present. And the future pays for the present through people — specifically, through a growing workforce producing expanding economic activity, generating the cash flow that services all the debt we’ve stacked up like a civilizational game of Jenga. More people means more workers means more tax revenue means more consumption means the debt gets paid and we do it all again next quarter. This is not a conspiracy. It’s just arithmetic.

Still, a lot of respectable minds suggest that artificial intelligence is going to displace a significant chunk of white-collar jobs within one to five years. Not factory jobs. Not the jobs we’ve been collectively hand-wringing about since the 1970s. The other jobs. The ones performed by people who went to college specifically so they wouldn’t have to worry about automation. The accountants. The analysts. The copywriters and the paralegals and the middle managers who synthesize reports that no one reads. Those people.

Now, if a large portion of the workforce is no longer working — if the robots are doing the producing but the humans are no longer earning or borrowing or spending — then who exactly is paying for all the debt? This is not a rhetorical question. It’s a structural one. The system was built on the assumption that the future would always have more participants than the present, and suddenly we’re looking at a future that might have fewer. Or at least fewer who are doing the thing the system needs them to do, which is participate economically. You can’t service a debt-based economy with people who don’t have jobs. You can’t grow your way out of leverage if the growth is happening inside a server farm that doesn’t take out mortgages.

And this is all happening against a backdrop that makes it worse. The post-war financial order — the rules-based system, the one where the United States sits at the center as the issuer of the global reserve currency — that system ran on something even more fragile than low interest rates. It ran on trust. Predictability. The shared understanding among nations that everyone would more or less play by the same rules, and that the rules would more or less stay the same. That predictability is what allowed leverage to build up in the first place. You can borrow against the future only if you believe the future will resemble the present, and you believe the future will resemble the present only if the geopolitical order feels stable.

It does not, at the moment, feel particularly stable.

Trust among nations is weakening. What backs money is becoming an open question again, which is the kind of question that sounds philosophical until it becomes extremely practical.

Sovereigns are answering it, quietly, by selling off U.S. Treasuries. Not in a panic. Not all at once. In the steady, deliberate way that countries do things when they’ve collectively decided something has changed but would prefer not to say so on the record. China has been trimming its holdings for years. Japan, less reliably than it used to. A handful of other reserve managers have been rotating into gold, which is the asset you buy specifically when you have stopped believing in the asset everyone told you for forty years was the only asset you needed to believe in. The Treasury market was, until recently, the closest thing the world had to a sure thing — the risk-free rate, the foundation layer, the asset every other asset got priced against. If the foundation is being unwound by the people who built on top of it, somebody has to step in and absorb what nobody else wants. That somebody, increasingly, is the Federal Reserve, which buys Treasuries with dollars it conjures into existence approximately ten seconds before spending them. We used to call this money printing. Now we call it “quantitative easing,” because money printing sounded too much like what it actually is.

And the Fed is about to be run by Kevin Warsh, who is walking into the chairmanship under the kind of political pressure that central bankers spend their entire careers pretending doesn’t exist. The pressure, specifically, is to not raise rates. Because raising rates would upset President Trump, and upsetting President Trump would, in turn, do unpleasant things to the value of stocks and bonds and real estate and every other asset class the people who care about these things care about. Stocks. Bonds. Real estate. The whole portfolio of paper wealth held by the wealth-holding class. Which means the institution that was explicitly designed to be insulated from political pressure is now functionally taking political pressure as its primary input. Which is, in fairness, more or less what central banks have always quietly done — just less loudly than they’re being asked to do it now.

The mask, as the kids say, is off.

When the relationships between nations become uncertain, the financial relationships built on top of them become uncertain too. And when everything becomes uncertain at once, you get deleveraging — which is the clinical term for the moment when everyone tries to unwind their bets simultaneously, and everything in the markets goes down. Not some things. Everything. Because in a leveraged system, the selling feeds on itself.

So to summarize: We built an economy that requires perpetual growth, funded by perpetual borrowing, sustained by a perpetually expanding workforce, stabilized by perpetual geopolitical trust, and quietly backstopped by a central bank that everyone agreed to pretend was apolitical. And now we’re looking at a world where AI might shrink the workforce, where borrowing costs are no longer trivially low, where foreign sovereigns are quietly heading for the Treasury exits, where the Fed is printing the difference, where the new chair of that Fed is being asked to choose between economic gravity and political consequence, and where the geopolitical consensus that underwrote the whole arrangement is fracturing in real time.

Other than that, things are fine.

The honest answer to what happens next is that nobody knows — which, if you think about it, is the most destabilizing answer possible in a system that was built entirely on the assumption that somebody did.