We tell ourselves a story about technology, and the story is that it sets us free. The printing press took books away from kings and monks. The internet handed a printing press to anyone with a keyboard. We treat it like gravity: a tool can start out expensive and locked in a lab, but it will leak out, get cheaper, and end up in our hands. Wait long enough and the future arrives at your door.
I believed that story for most of my life. I don’t anymore. And I want to tell you exactly where it breaks, because the break is not where you’d think.
Start with a floor.
1986. I’m five years old, sitting cross-legged on the cold tile of a school gymnasium in Paxton, Illinois. Someone has rolled in one of those heavy televisions on a metal cart, the shared screen of an entire generation. We are watching the Space Shuttle Challenger. Seventy-three seconds in, it comes apart in the sky.
What I remember is not the explosion. It’s the silence after it. These were the experts, the smartest people in the most powerful country on earth, working at the very edge of what humans knew how to do. And the system broke in front of all of us, live, with the sound off in the room because nobody knew what to say. I’m not sure I ever fully got up off that floor.
I’ve spent the decades since watching other supposedly infallible systems break the same way. The hollowing-out of the industrial towns in the nineties. 2008, where the people who built the collapse kept their bonuses and the rest of us got the bill. Every time, the same lesson: the system was never as solid as the people inside it insisted. It was a set of choices, made by particular people, in particular rooms. And choices can go the other way.
That’s the lens I bring to artificial intelligence. Not “is it good or bad.” Not “will the robots take the jobs.” The question I can’t stop asking is the structural one. When this technology produces something valuable, where does the power to make it actually live? With you, or with the people who own the machine?
Here’s the claim this whole series is built on. For two hundred years, every major information technology eventually pushed that power outward, toward ordinary people. The telegraph was a brutal monopoly. Western Union owned the wires, but it never owned the words you sent. AT&T owned the phone network, but it never owned your half of the conversation. Radio and television concentrated hard, and then cameras got smaller and cheaper until the studio ended up in your pocket. The channel got captured over and over. The capability to create kept leaking back to us anyway. That leak is the whole reason the optimistic story feels true.
Artificial intelligence is the first one that runs the other way.
When an AI model writes the code, drafts the memo, makes the image, does the thinking, that act does not happen in your hands. It happens inside a building you will never enter, on hardware you cannot buy, owned by a handful of companies, fed by an amount of money and electricity that rules out almost everyone alive. You get the output. You do not get the means of making it. And unlike every technology before it, you cannot reconstruct the thing for yourself once the door closes. The capability doesn’t leak outward. It pools at the top.
I call that the Inversion, and naming it matters, because a named thing can be argued with. If I’m wrong, you should be able to say precisely how I’m wrong, and by the end of this series I’ll hand you the exact conditions that would prove it. (There are a few. I take them seriously. I’ll show my work.)
But I want to be honest about why this isn’t a tech essay for me. I didn’t grow up with capital. I grew up being sorted, taught early which line I was supposed to stand in and how to keep quiet about it. I spent twenty years inside the customer-service and enterprise-software machine, close enough to the controls to see who they were really built for. So when I look at a technology that decides who gets to think at the frontier and who only gets to rent the result, I don’t see an abstraction. I see the oldest arrangement there is, rebuilt in silicon and called progress.
The good news is the part the Challenger taught me. A system that was built can be built otherwise. But only while it’s still being built. And this one is being poured right now, while we argue about chatbots.
So this is the question I’ll spend the next six posts answering, and it’s the one I’d ask you to sit with before the next one. If the machine that increasingly does our thinking lives somewhere we can’t reach and can’t own, what happens to us the day the people who own it change the terms?
Next: Pass, or Get Sorted. How I learned to read systems, and why you follow the power, not the money.