@MisuseCase On Bluesky, the Trust & Safety team is swamped with the influx of users, and they're really not paying attention to these ones. I think it'll come back to bite them.

The political bots are overwhelmingly scambots. There's no profit for them to be fash - most Bluesky users would block them on sight. So they pretend to be leftist, so their marks - that's you - feel a kinship and might "help out a friend".
The success rate is minuscule, but it's very, very cheap to run bots.

I'll end with my regular reminder: 90% of the bots are commercial, not political. At the moment, almost 100% of the Bluesky bots are "blue".
Most of the fake follower fodder - building up the lefty grifter accounts - are "blue dots in red states". This is exactly the same shit played on MAGA supporters previously on Twitter.
The bots are commercial, not political.

Equally troubling - and it goes from the CEO monarch butterfly to the lowliest Serie A funding investor - is all the money behind Bluesky is crypto.
If the cryptobots move in, I'm not sure they'll be shown the door.

What's troubling, too, is the "tech will save us" approach taken by the devs. Nothing in the history of social media suggests this is remotely true.
I follow most of the devs. They're really nice people and open to chatting with users.
But there's a lot of breathtaking techbro naivety in there. They really don't know how little they know about social media and how people behave.
Take, as an example, domain names. It's similar to creating your own instance, but Bluesky is far less a fediverse than you think. Bluesky, the company, controls a lot more than Mastodon does.
And they sell domain names. So I could switch to be @OutOntheMoors.satan.social as a joke (don't hate on them, they're fine), but there's a lot of ways I could get banned from the entire site without doing anything wrong.

The final new bot wave I've identified alarms me greatly because I'm not even sure it exists.
I've been "bot-hunting" for years - and I'm no longer sure I can spot them all now. The AI is getting really good.

It takes high-level language skills to identify them, and the apps and the sites persist in throwing more tech at the problem.
They're mostly Reply-Guy chatbots at the moment.
The problem is exacerbated by users who keep engaging with these bots, significantly blurring the lines.
It's all shits and giggles now, but I suspect these will become very high calibre trollbots - and hellishly difficult to spot. They could overwhelm a smaller site rather quickly.
Their objective, like that of most bots, is fake traffic metrics. The platforms are lying to the influencers, and the advertisers, and the investors. They're all pretending to offer "eyeballs" (that's you and me) - and the new bots are passing the Turing test.
We all laughed at the Twitter egg-avi influx, and we're still paying the price.

Clear your delusions: SkyNet is human and it's not sending the T-1000s back one at a time.

I noticed at the time mastodon.social became the default target for bots in the fedi that there was a lot of pentesting from bot networks. Essentially, it seemed they were looking at what would work in an open-source environment.
I'm not suggesting that the Bluesky bots are fully exploiting that route yet, but they're certainly looking into it as they shift focus.
It's marching in lockstep with the great move to AI.

The Brazilian bots are "old friends". They're almost exclusively fake follower-fake amplification networks, with sidelines in political manipulation, porn and scamming.
However, they've been around for ages and most sites have pest-control measures.

The second recent Bluesky bot wave is a big one. My god, there are a lot of them; probably around 5M by now.
It's whack-a-mole for the devs (more on the point of this later).
What's interesting is how the bots have evolved.

The first group is the Brazilians. This is not the Lula supporters who joined when he did. They're the bots that arrived when Elon refused to pay petty fines and Twitter was banned in Brazil.
Brazil largely controls two types of bots: pretty picture (scenic stuff) ones built up to sell with inbuilt followers (and the fake followers themselves), and fan pages (with lots of fake followers, too, but the business model is different).

Unlike the Lula crowd, the Brazilian bots trend pretty fashy. But most of the Brazilians - real and not - have fucked off back to Twitter, leaving their accounts open but dormant.

I've been back on Bluesky for a month, and I can barely tear myself away from the bot nonsense.
You probably know it fascinates me, and there are just so many new schemes and fiddles that it's like the first day of school.
It's in three main categories, and I'll talk about them as separate groups though they frequently overlap.
They are all indigenous to, but not exclusive to, Bluesky.