In 1960, Science published a paper with a date in its title. 

Friday, November 13, 2026. 

It was supposed to be doomsday. 

Physicist Heinz von Foerster fit a curve to 2,000 years of population data. The curve went vertical. By his math, the human population would reach infinity today. We’d be crushed to death by our own numbers.

“Our great—great—grandchildren will not starve to death,” von Foerster and his co-authors wrote. “They will be squeezed to death.” 

While I’m feeling pretty squeezed to death today, it’s not because of overpopulation. (I’m holding. The stocks are not.)

Turns out, von Foerster’s doomsday curve bent. Population growth peaked in the 1960s and has slowed ever since. 

So it goes… 

Every curve looks like doomsday from the steep part. 

More on the vertical curve—and what it means for the AI trade—in a moment. First, let’s talk about doomsday. 

Planet of the Apes 

Knock over a model of the Statue of Liberty. Throw some dirt on it. Roll film. 

Action! 

That’s the last shot of Planet of the Apes. 

That’s also when, according to sci-fi author Neal Stephenson, Hollywood got hooked on the apocalypse.

Why? Because ruin is cheap. 

Take something that exists today, bang it up, throw some sand on it, and—ta-da—a dark future awaits. Little imagination required. 

Consider the climate. 

The movie came out in 1968. Paul Ehrlich’s bestseller The Population Bomb hit shelves the same year and tightened von Foerster’s timeline: hundreds of millions would starve in the 1970s.

Ehrlich’s core argument was the same one von Foerster made famous: doubling times keep getting shorter, so growth keeps speeding up. 

Simple. That’s the problem. 

Writing up a scary future is so easy, I can do it in three words: Pow. You’re dead. 

Writing up one that works well takes real engineering of the imagination. You have to think through the second-order effects, new kinds of markets with new kinds of things, the regulators, the criminals, and the kids who use those new things in ways the inventors never see coming.

Stephenson has done that work for 30 years. 

Snow Crash gave us the metaverse in 1992—the idea and the word. The Diamond Age gave us an AI tutor that raises a child. Silicon Valley has spent decades building the things he wrote about first. 

So when he sat for an interview recently, he got the question everyone asks him: Are we headed for utopia or dystopia?

He gave three calls.

One: AI will flatten out.

“I’m more of an S-curve guy,” he said. 

Every technology looks exponential until it hits a natural limit. 

He could name one exception: nuclear weapons. Tabletop lab experiments became hydrogen bombs in about 15 years. 

Even then, physics capped the payoff. Past a certain size, a bigger bomb just punches a hole in the atmosphere and vents into space.

He expects AI to bend like everything else. The labs are betting everything on him being wrong. Either way, it might not matter. Because of number two: 

Two: Nobody can hit the brakes.

In February 1940, IG Farben tried to buy every drop of heavy water from the only plant on Earth that made it. 

Heavy water is water made with a heavier form of hydrogen. Germany’s reactor design needed tons of it to keep a uranium chain reaction going.

Months later, Germany overran Belgium and ordered all the uranium the Belgian Congo could mine.

If you sat in Allied intelligence, those two orders told you one thing: launch the Manhattan Project. Now. 

Germany never came close to a bomb. The fear proved unnecessary. Nobody could know that in 1940.

That’s where America sits with China. Slow down, and you bet the country on information you don’t have.

Meaning? The spending continues because no government can afford to stop it. Watch the purchase orders. Whoever is cornering the inputs—chips, power, uranium—is telling you what’s what.

Three: Even the prophet missed it.

Stephenson admits AI “blindsided” him. 

For decades, “AI” meant whatever software didn’t quite work yet. He waved it off. Then it started working, and he came to it late.

If the man who imagined the metaverse didn’t see this coming, maybe nobody did. 

His point? Anyone selling you 100% certainty about where AI goes next is guessing. Some guesses are better than others. But they’re still guesses. 

The ending he won’t write

So which is it? Utopia or dystopia? 

Stephenson refused to pick. 

But if we put his calls together, his future looks like this: 

The curve bends. The spending keeps climbing anyway. Washington won’t allow a slowdown, and the CEOs can’t afford one. Every extra inch of progress costs more compute than the last. 

One thing I agree with him on, though: 

Dystopia is the cheap bet. It’s also boring. 

In 1960, von Foerster had the math. In 1968, Ehrlich had the bestseller. Hollywood had the shot of the Statue of Liberty in the sand. Three different crowds, one easy ending.

The population curve bent. The famine never came. The world kept standing.

It still is. The hard work is figuring out what gets built next to it.

Fortunately, our team here at Paradigm has had—and continues to have—some pretty good guesses.