(Expanding on this a little.)

You may have seen a confusing sponsored post by the American medical insurance company Aetna:

There’s no particular reason for any human being ever to reblog this except maybe to make fun of it (see: me). It conveys no useful information of any kind.

But there’s also no reason for a bot to reblog it. Bots like stuff that 1) contains plain text that search engines can read, and/or 2) is tagged such that it looks appropriate to their ostensible theme

The post presently has over 2,500 notes, many of them reblogs. Maybe there are a lot of self-improvement-themed spamblogs, or education-themed ones? Getting tricked by the tags #mindfulness and #educate?

Nope, most of the reblogs look like this:

Q:

Why would large numbers of gay porn blogs reblog this incoherent medical insurance ad?

A:

Because Yahoo promises Tumblr advertisers a minimum amount of exposure, partly to be measured in number of Likes and Reblogs.

I don’t immediately see any reason for pornbots to reblog this post. If you look through their content, it’s clear that they’re generally sophisticated enough to avoid reblogging stuff other than actual porn. And why would you advertise for a company that’s not paying you for it?

But it’s good for Tumblr’s bottom line, because they can show their advertisers note-count and reblog-count - give them the idea that a lot of real people are really engaging with and sharing their ads.

The advertisers most likely do not know how to look at the notes and see how many of these notes are bots. Aetna’s PR department doesn’t have TagViewer.

If I’m right, Tumblr isn’t getting rid of the pornbots because Tumblr is running a lot of the pornbots.