Severian posted this last week, but I’m only just getting caught up now:

I was wrong about Musk buying Twitter. Lot of that going around — the Z Man got a whole podcast on “avoiding error” from his misread of the situation. It’s well worth a listen. I, too, had a “hot take” on Musk’s offer — not that it was particularly hot, as most folks on this side were saying it, but I too thought it was a stunt. After all, Musk, like Bezos and all the other “new commerce” billionaires, don’t exist without massive government support. I figured his “offer” was stoyak — he’s got something in the works in the Imperial Capital and needed to play hardball with somebody.

But I was also working off my longstanding assumption that Twitter, Faceborg, and all the rest are essentially CIA / NSA fronts. When I first heard about Facebook, my first thought was “Wait, don’t we already have Friendster? What does this bring to the table?” My second thought was the first one I’d had about Friendster: “That’s clever, I guess, but how on earth is this going to make money? Even if they saturate it with ads, to the point where it’s unusable — which will happen in about two weeks — they can’t monetize your personal data any farther. People are pretty set in their habits — once the algorithm figures out you’re the kind of guy who likes anime and New Wave music, any further data is useless.”

Being a much more naive, trusting sort back then, I figured it was just stupidity. You know, Pets.com level stupidity. The VC boys were trying to get another dotcom bubble inflated, because if the first one proved anything, it’s that people are dumb and will keep falling for the same obvious scam over and over. I could hear them in the board rooms: “This time, instead of sticking ‘cyber’ in front of everything, we’ll call it ‘Web 2.0.’ Cha-ching!”

Obviously that didn’t happen. So I went with the common explanation that was floating around in those days, that “social media” sites made their money by selling your data to advertisers. But that doesn’t pass the smell test either. For one thing, as I said above, your habits don’t change very much. For another, as anyone who has any experience with them knows, those algorithms really suck. The other day, for instance, I was listening to some old music one of the streaming music sites. And I mean really old. Nothing I’d played the whole morning had been composed after the 17th century, but the service’s algorithm was convinced that what I’d really like to listen to next was some rapper.

Indeed, the whole point of the ads on Pandora, Spotify, whatever seems to be: To annoy you to the point where you pay for their premium service. Pandora, for instance, either really really really believes I want a Surface Pro 8 and some Taco Bell, or they’re just playing those ads every two songs to annoy me into buying the premium service (which is every ad that isn’t Surface Pro or Taco Bell). Which is just bizarre, because I haven’t had Taco Bell since college — which was 30 years ago, and I paid cash — and this essay right here is the first time I have ever even typed the words “Surface Pro 8”, much less looked at the product.

I really wouldn’t be surprised that the “algorithm” is reading itself. Hey, this guy sure has seen a lot of ads for Taco Bell and Surface Pro! He must really want some!

But the algorithm for companies whose entire business model is e-commerce is no better. Amazon seems to have gone to a “push” model — they must be selling their suppliers on the idea that they can push you stuff, which is why they always pimp the same four or five items in the “Amazon’s Choice” recommendations, no matter what you’re searching for. And these again are laughably wrong — the only things I get off Amazon are used history and philosophy books, and stuff for my dog. Based on this, they have concluded that what I’m really looking for are chick lit and beach gear.

Given all that, I came to the conclusion that “social media” (and Amazon too, probably) really only have one customer, who really does have a use for your data, and that customer’s initials are CIA. It’s a giant domestic surveillance operation.

And why wouldn’t it be? The Regime has had a legitimacy problem for a long time, and a “feedback loop” problem for longer than that. Even if we assume no ulterior motives whatsoever — fat chance, but let’s stipulate — the fact remains that public opinion polling, however you want to define it, has a similar problem as psychological studies. Since the vast majority of study participants are college undergrads, what you get is WEIRD — that’s Western, educated, industrialized, rich and democratic, and also in a very narrow age range. Psych studies that purport to be universal are, at their very best, snapshots inside the head of the BCG.

If you haven’t encountered the Basic College Girl, he provided a thumbnail sketch here.