I'm sure some are like you describe.

But what I keep seeing is stuff like this:

An account which seems like it was a normal account for like 10 years. Maybe it leans a bit in the direction of the content it is currently posting. But, it was mostly dormant for a few years. Then, it started posting a high volume of partisan content. It otherwise seems to be a real person, just it one day started posting hardcore partisan materials all over the regional groups (usually multiple). It's fishy.

Another one: An account which superficially seems to be a real account, except that it's got a generic enough name you can't pin down which person with that name it might be. Nothing about their profile identifies them as anyone in particular. They may have a lot of friends, but they don't appear to be people who know them IRL - just randos. Their profile feed is a mix of normal stuff and partisan stuff. What they post to groups is 100% hardcore partisan stuff. Their profile is very light on photos of themself, and always kind of low quality. There's never a photo of a place you can pin down to a real location. These don't tend to be accounts that have been around more than a year or three. Again, it's fishy.

Like I mentioned, in both cases, you'll see the accounts posting at odd hours for where they supposedly live.

They could be real people, or sophisticated trolling operations. TBH, the reason I lean more towards the latter is because Facebook is so goddamned lax about any kind of enforcement. Do I think someone could set up a bunch of fake profiles and operate them for years, all from the same Chinese or Russian IP address with locations all in the US, and Facebook would never ban them despite the accounts getting flagged? Yes, that is extremely easy to believe. If it drives engagement (and troll accounts do!), it's likely to be allowed to continue.

I'd be the happiest man on the internet for a split second if someone who works for Meta tromps on in here with a burner account and tells me I'm being an idiot, that they're fighting hard against troll armies and bots, that it's just a super hard problem because the troll/bot networks are so sophisticated. Hasn't happened yet.

Most all recommendation algorithms definitely do not care about locality of reference.

If I watch some local-to-me news stories about libraries, the algorithm decides that I like news about libraries, and begins to suggest library-related videos from all over the USA. It does not care that I only cared about libraries in Phoenix and San Diego.

Likewise, there is this one English teacher who is absolutely gorgeous and I only watch her videos to see her talk. But of course, YouTube got the idea that I like English instructional videos from anyone, anywhere, and the algorithm slowly fills my feed with videos from instructors who are not the lady I like. It's crazy, I tell ya.

I'm not on FB anymore. But when I was, I got interested in Irish topics, and it was often family research-related, and so I'd be studying some esoteric topic in a very small town in a specific county of Ireland, and of course, the recommendation engine would blossom with completely irrelevant Pages and Groups from every corner of the Republic, for no reason except to bait me into engagement.

And that's all it is, engagement bait. They on the other side of the algorithm, they know for sure that you don't actually care about those topics or locations. But they are simply trying to dangle stuff in front of you like a largemouth bass, and eventually something will enrage you or you'll need to correct someone's spelling, or make an angry face emoji on a post, and then they have your engagement, and that's all they really wanted from you.

Not to inform you, or entertain you, but to engage your shit until you absolutely lose your temper. That's what they want on social media.

Is it? Its a glorified ECard service. The entire thing could run on a single machine.

That goes against Facebook policies and the account would get banned. Not to mention that’s a gigantic strawman.