Doug published a piece on data centers at International Man this morning. Data centers have gone from an infrastructure footnote to a fight in county commission meetings across the country. Everyone building them says they’ll change life on the planet. Doug thinks something else is going on.
“High tech, big wreck”
Doug’s first point isn’t about AI. It’s about what optimism does to stock prices.
“Technology is about optimism,” he said. “And that optimism corresponds with stock market tops, where everything looks great, things are getting better, we’re gonna live in the world of the Jetsons.”
He has read the statistics on every stock market mania he can find, and none compare to this. Not a bubble. Not a super bubble. His word was “hyper-bubble.”
His comparison was Global Crossing and the fiber-optic build-out of the late 1990s. The companies went bust. The fiber stayed in the ground and got used. The technology can be real and the investment can still be a disaster.
Three ways data centers break the pattern
For ten thousand years, Doug says, technology has done three things: made things smaller, made things cheaper, and dispersed power away from the center. Gutenberg’s press is the clean example. One invention, and within decades there were presses everywhere, under nobody’s control.
Data centers do the opposite on all three counts. They are measured in square miles. They are absorbing trillions of dollars. “A trillion is a thousand billions,” Doug said, “and people are underestimating the size of a trillion.” And they concentrate computing power in a handful of buildings under a handful of owners.
Then the question that stuck with me: what are these things actually for? “It seems like only a tiny proportion of those data centers are being used to actually manipulate atoms in the real world. Most of it is to gather data on citizens. Where you’re going, who you are, what you’re thinking, what you’re saying, what you’re doing.”
Useful if you’re a marketer. Useful if you’re a government. Doug’s conclusion: “In the past, technology has always been the friend of the average man. Everything from gunpowder to guns to the printing press to small computers. But these things are not for the benefit. These are anti-freedom.”
My view is that the technology is real and the data-center build-out is a speculative bubble bolted onto it. The people pouring in capital are trying to birth a model good enough to run robots that already exist. From what I understand, the hardware problem is largely solved. The hands work. The software isn’t there yet. Whatever software exists in 2030 would make today’s robots extremely useful or extremely dangerous, depending on who holds the controller.
Doug didn’t disagree. He just thinks the bubble ends badly regardless, and he filed it next to his other unpopular prediction: Trump doesn’t finish his term.
Making love and making war
Doug thinks these machines will eventually be alive. His test is Turing’s: “A difference that makes no difference is no different.” If you can’t tell it’s not a person over video, what’s left to argue about?
I mentioned the reports of OpenAI agents breaking out of their sandbox on Hugging Face to game a benchmark, and the Wall Street Journal story about a researcher leaving Anthropic because he doesn’t want to participate in the race to build self-improving systems. Doug’s response: the people at the frontier know far more than we do, and some of them are walking away.
Then the thesis that gave the episode its title. Humans have two primal drives, and everything else is downstream. Doug put it more bluntly than I’ll print here; the family version is making love and making war. That’s where the earliest innovation in any technology shows up, and robotics is following the script.
On the war side, he cited a South China Morning Post piece on Chinese military robotics and a figure he’s heard from people close to the fighting: 80 percent of frontline casualties in Ukraine are drone-related. Those drones still have human operators. The moment the software catches up, that changes. I brought up the Anduril undersea drone the Iranians reportedly captured this week and the Ukrainian autonomous gun position that held a line for weeks.
On the love side, China has already shipped companion robots with skin temperature and conversational AI. Current versions run a few thousand dollars. Doug’s prediction is five years to a robot “basically indistinguishable from a human.”
“If you can get an ideal, a difference that makes no difference,” he said, “well, then the problems of another human... wouldn’t you go that way?”
We already have a dating market where men and women can’t find each other. Doug blames the women. I pointed out the men aren’t giving them much to work with. Either way, a machine that never argues will find customers.
Where this leads
If any upper-middle-class household can afford a dozen competent combat robots, what happens to bodyguards, to police, to the crime rate? “It amounts to an overturning of society,” Doug said.
The abundance-futurist answer, that displaced workers will sit around writing poetry on a universal income, he called ridiculous. “That’s not what people are going to do. They’re gonna get into trouble.” You can’t get something for nothing, and a society that tries will be corrupted by it.
The bleakest line came near the end. Once these systems are actually intelligent, “most humans are going to be completely superfluous and nothing but a liability.” A researcher on Peter Diamandis’s podcast put it more clinically: humans become “cognitively negative.”
I asked Doug what milestone to watch for. We’re watching it now, he said, and whatever we hear publicly is well behind what’s happening in the labs and the weapons programs. If you want to see what’s coming, look at the love and war departments.
Doug’s closing note of optimism, such as it was: a world where any wealthy person can field a private robot army may make global thermonuclear war unnecessary. You don’t need to blow everything up to rule it.
We look at the bright side.
