Twenty Years Is a Long Time to Be Wrong About Everything

Someone sent me the clip of the Musk interview with The Economist on Sunday morning, the way people send you things now, with no context, just a link and a “have you seen this.” I watched it while the coffee was still too hot to drink. Worth doing, if only for the bit at the end where he predicts AI and robots will make money essentially meaningless within twenty years, then in the next breath predicts civil war in Britain within ten. Two forecasts, wildly different timelines, no attempt to reconcile them. When the interviewer points out the tension, he just says the AI will probably stop the war before it gets going. Convenient.

I want to be fair here, because the pile-on happening in the comments under that clip is enormous and not all of it is interesting. A chunk of it is just “he looks unwell,” which, sure, he does, but that’s not really a point about anything. The more useful thread, buried under all that, is about a specific kind of failure mode: being genuinely excellent at one thing and mistaking that for being excellent at everything.

I’ve watched this happen at a much smaller scale in my own industry. Someone becomes a brilliant engineer, gets promoted into management because that’s the only career ladder anyone’s built, and turns out to be catastrophic at managing people. Different domain, same muscle memory, doesn’t transfer. Usually the damage is contained to one team and one budget cycle. Musk’s version of this plays out with satellites, elections, and a car company’s stock price attached to whether he can convincingly LARP as a genius on a livestream. The scale is the whole difference. The underlying mistake is the same one I’ve watched happen in open-plan offices in Docklands.

The twenty-year prediction thing bugs me for a reason that’s not really about him specifically. Somebody in that thread pointed out that in 2006, Bush was still president, MySpace was still the dominant platform, and you couldn’t buy a laptop without a disc drive. That’s the honest answer to any twenty-year forecast: nobody knows, and the confident ones are the least trustworthy, not the most. I say this as someone who works with technology every day and still can’t tell you with any real conviction what my own job looks like in five years, let alone twenty. The pattern-matching part of my brain that gets paid to write DevOps pipelines is not the same part of my brain qualified to forecast British civil unrest. I’d like to think I know the difference. Musk, evidently, does not, or has decided it doesn’t matter because enough people will nod along regardless.

Here’s the genuine tension, and I’m not going to pretend it resolves neatly: he’s also probably not wrong that AI and robotics are going to reshape work and daily life faster than most of us are emotionally prepared for. That part isn’t the delusion. People much more careful and much less online than Musk have been saying versions of this for years. The delusion is the specificity, the twenty years, the ten years, the tidy causal arrows between AI singularities and continental unrest, delivered with the same tone you’d use to read out a bus timetable. Confidence is not the same as being right, and I think we’ve collectively gotten worse at telling the two apart, not better, despite having more information available than any humans in history.

What actually got me, more than the Britain stuff, was the pettiness underneath it all, the fake gaming leaderboard business someone brought up in the thread. Not because it matters, in the scheme of things, whether a very rich man cheated at a video game. It doesn’t. But it’s a small, low-stakes window into character, and small low-stakes windows are usually more honest than the big performed ones. If someone will quietly pay a team of people to inflate a leaderboard ranking in a game nobody was forcing them to play, competitively, in front of no one who mattered, that tells you something about how they operate when the stakes are higher and somebody is watching. It’s not proof of anything. It’s just consistent with a pattern.

I don’t know what happens with any of this, honestly. I don’t think civil war in Britain is coming in ten years, and I don’t think universal post-scarcity abundance is coming in twenty, and I don’t think the man saying both of those things in the same breath has a coherent model of the future so much as a strong desire to be seen as someone who does. That’s a very human impulse, wanting to be the person who saw it coming. It’s just more alarming when the person having it owns a chunk of the satellites keeping half the world’s internet running.

I’ll keep an eye on the actual robotics and AI developments, because that part is real and worth taking seriously, on its own terms, separate from whoever happens to be shouting loudest about it this week. The rest of it, the timelines, the wars, the leather jacket, I’ll leave for the comment threads to sort out. They seem to be enjoying it more than I could.