Welcome.

For those of you who don't know, BlueSky is part of a social media environment called the AT Protocol (also known as ATProto, or the ATmosphere, depending on the circles you hang out with). No, it is not a replacement for X/Twitter in the strictest sense. And a lot of the community issues that people have on the platform are often a result of misunderstanding what the platform IS.

Think of the ATmosphere as a park: it is a set of a standards for social media things like posts and likes and feeds. Those standards are rules that anyone in the park has to follow, and which in turn enable everyone in the park to connect/talk with each other. A PDS, or a Personal Data Server, is the picnic table that one can bring in at the park, to enjoy the stuff AT the park. BlueSky is a PDS. The reason BlueSky gets to moderate your content when you make an account with them, is because you are HOSTING your content on their servers. That means they're directly on the hook for whatever content you put in there. You're also not paying for hosting, so they get to make whatever rules they want as to what content is and isn't allowed.

But the magic of ATProto is that if you don't like their content restrictions, but you do like the vibe of the ATmosphere, you can make your OWN server or PDS, and move to host all your stuff THERE. And suddenly BlueSky's restrictions no longer apply to you. (Legal restrictions will still apply to you, I shouldn't have to specify that, and yet...)

What all this means is that a bunch of nerds got together and said things like "wow, traditional social media sucks" and "I fucking hate it when there's a site migration and I lose track of ALL MY FRIENDS" and then designed an open source standard that allows users to OWN their own data and therefore be able to share it across multiple platforms without having to change services. It's all very technical and very cool. If you give a shit about the free internet, you should be very excited about ATProto. It was built by people who actually care and understand the issues of mainstream social media and actually want to DO something about it.

Anyway, where does the AI fearmongering come into all this? One of the limitations for the ATmosphere is the fact you CAN build custom feeds - basically your very own algorithm, built by you, controlled by you and monetized by fucking no one - but you need a certain threshold of technical knowledge to do it. And unfortunately, as BlueSky has grown, and with it, brought growth to the ATmosphere as a whole, the average technical skillset of the population has gone down.

Listen, Britney who works at Starbucks and just wants a place to see pictures of cute dogs is working for minimum wage during Current Events. Britney does not have time to learn how to fucking code in order to build a personalized feed of ONLY cute dogs. Even though having such a feed would materially impact her overall mental health in a very positive way.

There are hundreds of thousands of Britneys all over the ATmosphere.

Feeds have been a goddamn pain in the ass for everyone involved.

Enter the ATmosphere conference of 2026. In it, BlueSky showcased Attie, an AI feed builder, designed to break down the barrier of access to feeds. Here's what you need to know about Attie:

It is NOT part of BlueSky.

It is a SEPARATE app for the ATmosphere.

It was built BY the team at BlueSky, and it is OWNED by BlueSky, but it is not integrated INTO BlueSky. There are currently no plans or any intention to integrate Attie into BlueSky.

The AI bit of Attie is this: It uses a version of Claude Code to interpret instructions given in Natural Language and converts them into the code required to build a feed. The feed is now matched to your user, owned by you, and therefore it will follow you, persistently, across the ATmosphere, regardless of which platform or service you're using, ready for deployment.

The mostly tech illiterate userbase from Bluesky lost its mind and then immediately spiraled into fearmongering about AI.

And here I want to talk and describe why fearomngering about AI is effectively useless:

The most important one is that fearmongering is not tethered to reality. Therefore, it cannot effectively interact WITH reality. Rie, what the fuck does that mean? It means that people started having meltdowns in the comments and quotes of the announcement, and landed on things like "block the BlueSky profile for the app, that way they can't use your content in their feeds!" being passed on as secret wisdom to fight against the AI overtaking the space!

This is veritable nonsense. This is like saying "if you block the twitter profile of the IRS, you don't have to pay taxes anymore!" Or something equally asinine. This is pure mythologizing as a way to cope with a reality you don't understand. And fearmongering makes that anxiety worse. It actively prevents people from doing things that would actually further their goals, by allowing them to wallow in distraction and feel satisfied that they've Done Something About It, even though they really haven't.

It's also actively doing harm to people who don't know better on how to parse the fearmongering. I saw no less than five people having meltdowns in the quotes of the announcement about having to leave the site because, to their understanding (fueled by fearmongering in their feeds), Attie is basically Grok but Woke. So BlueSky is the same as X/Twitter and they might as well leave before all their stuff is stolen and fed into an AI.

Also, let's not leave it implied: this kind of fearmongering makes you look stupid. It just does. Anyone who knows anything about the tech is squinting at this reaction and mentally dismissing anything you have to say on the matter because you clearly don't know what the fuck you're talking about. Fearmongering makes you look and sound hysterical and irrational, and therefore a lost cause as far as actual productive conversation goes. Any real, valid concerns you might have are now buried beneath the histrionics of everything you're yelling about that is not in fact happening in reality.

And that's my key point: fearmongering disconnects you from reality. It presents to you an easily understandable yet profoundly distressing window into the world and tells you "that's how it is!" and if you do not have any other access point into reality, but the fearmongering space you're stuck in? There's virtually no distinction between you and anyone else who is fully removed from reality and being loud about it.

Fearmongering puts you in a state where no solutions are available to you. No mitigation is acceptable. No compromise is possible. This is by design: fearmongering does not trade in reality, it trades on feelings and the imaginary evil will always be bigger and stronger and grander than anything reality can offer to make it better. Nothing will ever be good enough to stop the crimes of the evil that exists solely in your head.

The kicker is that this reduces you to a loud, irrational minority that can safely be ignored, because the louder you get, the more obvious it is you're not engaging with reality. And if you're not engaging with reality, why would anyone bother engaging with you? And thus the vicious cycle around this: the less reality engages with you, the less you have an incentive to actually engage with it. Fearmongering doesn't feel GOOD in the traditional sense: it fills you with paranoia against impossible adversaries that are coming at you from all angles. But it is comforting in the same way as finally having a diagnostic after years of fighting to have your pain be taken seriously: you KNEW something was wrong, and you finally have a name for it! Even though you don't understand and are discouraged from discussing and better grasping what it means. It leaves you unable to parse reality or make any meaningful commentary about it, because you have so many assumptions based on nothing, that most people just refuse to engage. Trying to dismantle the web of stupidity feels like an impossible task that only someone who truly loves you will see the point in doing: in rescuing you from ignorance and suffering and let you grasp reality again, because they care about YOU and what you have to say.

Here's the kicker: the fearmongering about AI smothered a different conversation that weekend. A much more grounded, more realistic conversation, criticizing BlueSky for bringing AI into the mix.

Because there are teams currently working on solving the feeds issue, without fancy AI. Because feeds are a design issue for the ATmosphere and fixing them is priority to bring in better quality of service across the entire ecosystem.

But those teams got horrifically shafted and pushed aside by the introduction of Attie. It doesn't matter whether Attie works or not: because of BlueSky's size and presence in the ATmosphere, they wield very real power in the space. And their venture capital funding and fancy AI tools are actively pushing out real developers trying to provide solutions that aren't built on ethically dubious technology that is also currently subject of a bubble and all the uncertainty and instability that comes with that bubble.

And those are real, valid criticisms that BlueSky had to actually acknowledge and respond to, as opposed to a nondescript "wow, you guys are upset about AI, we heard you, lol" that amounts to nothing.

We're at the point where the bubble is starting to collapse. More and more people are calling bullshit on it. There are real risks associated to the technology that aren't the same three broken telephone complaints that people parrot out without really understanding what they mean: AI is bad at what it does! AI is bad for the environment! AI is built on stealing!

Here's the thing: I'm not saying those statements are not true, but they're nuanced. And this is a nuanced conversation. What happens now is going to be the result of a nuanced conversation. If you don't open yourself to nuance about this, you're not going to have a seat at the conversation.

If all you are willing to do is parrot talking points you don't understand and partake in the fearmongering, regardless of who it hurts and how ineffective it is: at this point, I don't think you're anti-AI or AI critical.

At this point, I think you're either an idiot or a plant.

We desperately need to have nuanced, difficult conversations about the datasets that exist today. We need to talk about usage and privacy rights and the way AI has gotten actually really good at the one thing it was built to do: code. We need to talk about how early failure is not permanent: the same way we told you that "six fingers is the mark of AI" was not going to work forever, you need to learn how to actually parse current failures and what those publicized failures are actually feeding into, as a general discourse.

We need to talk about AI.

AI is not magic or the devil. It's a tool. It's like a CNC. It's big, industrial, expensive and technically niche. Sure, your average Joe might do some silly or cool shit with it, and if you try to bring your CNC stuff into a handcrafts market and take up a stall, there's going to be problems. But you need to be able to parse it. You need to be able to understand what KIND of AI you're dealing with. You need to be able to handle distinctions and nuances and texture.

Because only by understanding that will you be able to advocate appropriately. If AI is stealing your data but not creating gen AI content, for example. Knowing that difference fucking matters! You cannot demand regulation and policy to be built, if you're not able to engage and describe reality as it is.

We need to be able to talk about AI.

And it starts with you, doing the reading, putting in the time and the effort. Listening to experts. ACTUAL experts, not just randos in the internet. People who understand how the tech works and what the impacts are. People who are actively working to prevent erosion to our rights. People who know what the fuck they're talking about.

They're out there! So maybe next time you hear the dreaded "AI", take the plunge and look for the experts and their opinions, instead of the screeching asshole trying to give you a fucking heart attack.

Sources: