[Scoundreller]

What I find funny is when some multibillion company buys tens of thousands of followers on Twitter, but each post gets like 1 like or 1 comment. Like, hello!!!!!

But maybe it's because of FTC rules: fake followers are legally okay, but fake likes and fake comments could be considered a fake review and therefore FTC violation.

[thow-58d4e8b]

Another giveaway is accounts posting multi-sentence comments or reviews within seconds of their previous submissions. How can this get past bot checks is beyond me

[kabes]

For the same reason that on facebook I still get daily friend requests from very sexy women with a brand new profile, 4 friends total, none in common, in a random location in the world with 1 post linking to some dating site. It's because they just don't care. Well, either that or facebook engineers are completely incompetent.

[bigiain]

> Well, either that or facebook engineers are completely incompetent.

Or they're making bank on dating site affiliate fees as a side hustle...

[3np]

They don’t need to. The increased time you spend in the site as a consequence is enough of an upside to look the other way.

[disgruntledphd2]

This really isn't true. Remember that FB has many, many users and (presumably) many, many scammers, so one would expect to see lots of scams.

I too get those (messages lately) friend requests, and almost always by the time I read the message the account has been deactivated (when it says FB user).

So, overall, I think they're doing a reasonably good job on this particular problem.

[kabes]

My experience is different. I have many times reported a profile like this. Exactly as described above, a child can see it's a scam account. And then a couple of days later got message that after review the profile doesn't violate the community standards.

[Scoundreller]

The social networks don't mind the bots if it leads to engagement. Their "anti-bot" policy is a joke. They know who the bots are. They only ban the ones that are anti-engagement (which they can also determine).

[shrikant]

Absolutely this.

Just a couple of days back, something had been bothering me about some new-ish commenters I'd encountered on Reddit, so I did an unscientific check. I realised that about a year or so back, Reddit quietly made it so that when you create an account, they autosuggest a username of the form "Word1-Word2-Number" or "Word1Word2Number", making it hard to tell apart from automated bot/astroturf accounts.

(Try it out -- look at the profiles of any account you encounter that has the above form, and they've been created pretty much always on/after October 2020)

My conspiracy theory is that this was a growth marketing hack to muddy the obvious differences between regular people accounts and bots/spammers.