Recently in the Atmosphere there were several contentious conversations happening, and I believe many people either lost the plot or didn’t even know what the plot was to begin with.
I didn’t really weigh in on these conversations because every post I began to write felt woefully inadequate when confined by the short-form medium of Bluesky, so I’ve spent the past couple of weeks writing my first long-form post in months. I plan to begin writing in this manner more frequently, so I joined Leaflet Pro and am launching an email newsletter you can subscribe to.
dame.is blogging
The contentious discussions from that week were in one way or another a continuation of the recurring “Bluesky is dying” discourse that has been popping up at least once a quarter for some time now.
First, there was a new Attie update. If you aren’t familiar, Attie is Bluesky’s experimental AI platform. Reactions to this update were mixed; many users eviscerated Bluesky by resurfacing posts the company made in 2025 where it dunked on other tech companies for introducing unwanted AI features. While those misguided brand posts didn’t age well and look hypocritical on the surface, I think people overreacted and also unintentionally made some false equivalences.
Graber > Gruber
The Attie update came on the heels of a widely circulated (and widely criticized!) John Gruber blog post in which the creator of Markdown praised X and bemoaned the shortcomings of Bluesky:
“Fascinating new apps are being shown and discussed on X and nowhere else.”
“...there’s little action on Bluesky or, especially, Mastodon. There’s a lot of activity on Threads, but little action. There is action on X.”
There’s a lot that can be said about Gruber’s words, but my main rebuttal in the moment was: “There isn’t action on X! There’s engagement grifting, linkedin posting, & talk. You can’t build. Elon prevents you from acting.”
Many people, including John Gruber, are operating under an incorrect assumption that Bluesky’s goal is to build a better version of Twitter/X/Threads in order to replace those apps. I don’t really fault people for having this belief, but it misses the atmosphere for the blue sky.
All the way back in 2019, Jack Dorsey described Bluesky’s objective this way:
“Twitter is funding a small independent team of up to five open source architects, engineers, and designers to develop an open and decentralized standard for social media. The goal is for Twitter to ultimately be a client of this standard.”
A couple of years later, when Bluesky spun out as an independent public benefit company, Jay Graber stated clearly what its mission was:
“Our mission is to develop and drive large-scale adoption of technologies for open and decentralized public conversation.”
And to this day, when you visit Bluesky’s company website, you’ll see the following:
“We're building a foundation for connection that gives users a choice in their experience, creators independence from platforms, and developers the freedom to build...”
Bluesky’s original mission was not to create Bluesky the app, it was to create the Atmosphere: an open and interoperable ecosystem of social experiences that was resistant to billionaire capture and big tech oligarchy. The Twitter-like app most people know as “Bluesky” is merely the current biggest and best example of what can be built in the Atmosphere.
Here is what Bluesky’s CEO, Toni Schneider, had to say that same week:
Atmosphere update: atproto apps are growing 20% month over month, and last week we passed 1,000 daily active apps. My goal is to get more people building on atproto. We'll keep integrating more apps into Bluesky. Earlier integrations like Germ, Streamplace, and Standard.site have done well, and they make the Bluesky experience richer while showcasing the people building on the protocol.
I’ve been on Bluesky since basically the very beginning and have been building things on the network full-time for the past 2 years. When I first started, there were hardly any formal apps on the network other than Bluesky. Today, there are too many to keep track of!
AI and atproto
We got incredibly lucky that AI technology found its footing around the same time as Bluesky, because the dance between the two has facilitated the beginnings of a Cambrian explosion of social experiences that could fulfill Bluesky’s original mission.
If agentic software development had not come along when it did, I think Bluesky’s mission would likely have failed. At the very least, it would have been much slower and faced far greater obstacles, more likely than not fizzling out before it could reach escape velocity. This is ironic when you consider that a large portion of Bluesky’s user base is deeply anti-AI — the thing they hate is, in many ways, a key part of what could make the mission work.
And so, when we talk about Attie and the current state of the Bluesky app (especially as it relates to its so-called competitors), we have to talk about them with our eyes on the real prize: the Atmosphere flourishing and taking over the social web.
Attie
Attie was announced less than a year ago on stage in Vancouver at ATmosphereConf 2026. I was in the audience and one of the first people to try its early prototype. In the following hours and days, I watched as the community reckoned with what it meant and the feelings of dissonance it created. At first it seemed as if Bluesky had sherlocked part of its builder community. Thankfully, that didn’t happen. But it’s been about 7 months now and neither our worst fears nor our best hopes about Attie have come to fruition... yet.
The first 3 months of Attie’s existence were pretty quiet, leaving some in the community to wonder if it had fizzled out. It became one of the most blocked accounts on the network due to so many people being strongly opposed to AI. The features Attie offered at this time were half-baked (it was a prototype after all!), so most folks, myself included, didn’t use it much beyond those first few days of testing. Eventually Bluesky hired some talented folks from the community to work on Attie full-time, and since then we have seen them shipping more experiments and enhancements on a regular basis.
I don’t use Attie much yet, but I’m glad it exists and that Bluesky is dedicating resources to exploring the intersection of social media and AI. It would be a generational mistake for Bluesky not to do this, because AI will be central to the experience (and growth) of the open social web. AI is, unintuitively, part of the escape hatch from big tech.
During Attie’s ongoing development over the past few months, the Bluesky team has at times given me early access to some of the things it is experimenting with. Through my own work building in the Atmosphere, I’ve come to know a lot of the folks who work at Bluesky, and I’ve had the opportunity to collaborate with them at times on things unrelated to Attie. My relative “closeness” to them informs my perspective of course, but what I’m sharing in today’s post is not based on or influenced by any “insider info”. I don’t know more about Attie than anyone else at this point.
I see Attie as Bluesky’s test kitchen for AI experimentation; it’s a place where they can try things and mess up in an environment away from the flagship Bluesky app. This is important to note because, as I said at the beginning, many people felt Bluesky was being hypocritical about AI features and falling into the same frustrating pattern we’ve seen play out in big tech. In reality though, Bluesky’s approach is quite different.
Instead of forcing half-baked AI features and product whims into the main app, Bluesky established a separate space for this that ordinary users could completely avoid: Attie. Additionally, thanks to the nature of the AT Protocol, if Bluesky did begin introducing AI features into its app that people dislike, those people could simply switch to a different client and/or provider. These two things make Bluesky’s approach to AI extremely different from what Mark Zuckerberg and Elon Musk are doing.
Bluesky should not be criticized for this. It should be praised.
Action
Thanks to the AT Protocol’s openness, I’ve been able to play an active role in imagining what the intersection of social media and AI could look like. Through my independent work on Anisota, Aturi, and Poe, I have found some promising ways AI can give everyday people control over their online lives that wasn’t possible before, enabling a more accessible and pro-social future.
Anisota, my experimental social interface for Bluesky that helps people be offline more and online better, now has multiple “AI” features that are intentionally scoped to key pain points many of us experience in our day to day lives:
- 1.
Posting Adviser: helps align your posts to your values/goals before you publish
- 2.
Alt Text Generator: drafts quality image descriptions so more posts are accessible
- 3.
Feed Summaries with Content Control: catches you up on your feed or provides you with a peak without having to get into the weeds, with control over what gets included
- 4.
Reflections and Patterns: helps you notice patterns in how you use social media (and how you could use it better)
I built each of these features because each met a practical need in my own digital life, and they are geared towards cultivating a healthier Atmosphere.
One of my other projects, Aturi, has a hosted MCP server (a standard way for AI tools to plug into outside data) that makes the Atmosphere more traversable to AI agents.
And then there’s Poe, a modern take on Clippy for the Atmosphere. It’s a digital potato that can answer nuanced and complex questions about the ecosystem by tapping into the Aturi MCP as well as its own infrastructure like the Atmosphere Weather Station.
Each of these is an experiment in how AI can make the social web healthier and more accessible than in the past. Bluesky is also engaging in this kind of experimentation with Attie.
I don’t think Attie should worry much about its UI or form factor right now. It isn’t a finished product, and at this stage its job is to change quickly, which means avoiding anything that would slow down pivots.
So when I see Attie and the team try something to see what happens, I don’t fault them for not always getting it right or living up to some extremely high bar that we’ve come to expect from a fully polished product. We should expect them to try things that end up not working! Playing and experimenting is how you discover what comes next, and in the context of Bluesky PBC, that’s best done in an environment like Attie, away from the main Bluesky app.
And for the rest of us, atproto and the Atmosphere are rich environments for play and experimentation of our own.
We can be a part of discovering what’s next.
We aren’t locked down in a place like X where its owner actively prevents play. We can actually take action here instead of just talking.
As for me and my work, my goal is to help make it as easy as possible for anyone to take action in the Atmosphere, regardless of their skill level, background, or experience.
The more of us that get involved in shaping the future, the better, and it’s already happening right here in the Atmosphere.
If you’d like to read more pieces like this, subscribe to the newsletter.