Last week I spent an energizing, educational, occasionally infuriating week in Vancouver around the ATmosphereConf, a community event gathering developers, investors, data scientists, the odd academic, and a large contingent from Bluesky PBC themselves to discuss what we call the ATmosphere.
The ATmosphere is the group of projects emerging around ATProto, the protocol powering Bluesky. I wouldn’t call it an ecosystem yet — Bluesky continually calling it a community is telling of how they are approaching the space. It is currently more of a primordial soup out of which an ecosystem may eventually emerge.
It’s a weird, vibrant space. There is a glut of skeuomorphic projects (I’m guilty of creating Winesky.app myself, which is in no small part “Vivino, but on ATproto”) but we are starting to see a few native lifeforms appearing.
Shaping a taxonomy
At the event, I articulated my ATProto investment razor, flipping around an issue that I realized applications like Graze had when talking to outside investors.
ATProto investment razor: if a team can easily explain their project to an investor coming in cold, they are not building something native.
ATProto is weird. Any native life-form will look weird to an outsider.
🧵
— Ricardo J. Méndez (@ricardo.bsky.social) March 27, 2026 at 4:30 PM
A simple razor is a start, but not enough — it helps us see what may survive, but doesn’t tell us why, doesn’t help us build categories — buckets into which we may slot new entrants and project their behavior.
We need a taxonomy.
So I used that razor to cut away all non-native projects (kill your own Winesky darlings!), and what was left fell into three general application categories:
- Symbionts,
- Offshoots, and
- Cuckoos.
These are applications, specifically — things that end users will end up interacting with. Each category is defined by a different survival dependency: what it needs from the ATmosphere to thrive, and what it will eventually need to grow beyond. There will be a layer of ATmospheric infrastructure that will cut across all three, but I won’t get into that today.
Symbionts
Symbionts are projects that complement Bluesky while benefitting from the growth of Bluesky’s own user base, and help fill in the main platform’s gap with their own expert knowledge. Their survival hinges on two things: Bluesky’s user base to grow into, and Bluesky’s restraint in not competing with them directly.
Two key examples are:
- Graze, a custom feed builder created by a team with long expertise in machine learning, and which has so far done work in reaching out to organizations to figure out their needs when it comes to real-life feeds, and
- Leaflet, a long-form writing platform that integrates seamlessly in the ATmosphere while bringing in a team with knowledge of what publishers are looking for.
This is a category that I would have considered a fertile ground for native applications, until Bluesky came on stage and made clear to anyone paying attention that, as Trezy put it, they were all too happy to eat their young — showing an experimental product that directly competed with Graze.
Symbionts are thus in a precarious position: since they are so unique to the ecosystem, they will run afoul of my razor above and have a hard time explaining their nature to investors; while at the same time they have just learned that they will need to keep nervously eyeing their main partner.
It would be tempting to tell yourself that the expert knowledge is the key differentiation factor and that Bluesky, as a rational player, will stay focused in their core area.
If you are trying to convince yourself of that, consider that Bluesky’s Discover feed has been widely panned for being useless, while the For You feed, run by user @spacecowboy17.bsky.social as an experiment out of their home, is vastly superior.
If Bluesky’s failure in building a personalized feed didn’t stop them from encroaching on a community member focused on custom feeds, why are they going to stay away from your area?
Offshoots
Evolutionary offshoots are platforms that look a lot like Bluesky, but have their own user base and ideally a fundamental differentiation. They start with a dependency on Bluesky’s identity and social graph — but their whole project is to escape it.
Blacksky is the foremost example.
They run Blacksky.community, which reads like a Bluesky client re-skin. However, Blacksky is going well beyond it behind the scenes: they have their own moderation and T&S team; their own rules about which accounts to keep around, ban, or relay; their own growing parallel stack (which will be great for general resilience); and are even building products for their community like Blacksky Cash.
Eurosky is another, more recent example, and at this point likely larger than Blacksky by sheer account numbers, after they got over 9000 signups during their first week. Eurosky currently differentiates on jurisdiction — if you are a European social media user, you might be worried about keeping your posts and social media identity with a US company, under the capricious thumbs of their increasingly deranged administrations.
And then there’s Cartridge, which aims to be a blend of IGDB, Steam, and GoG — not only letting users keep control over their own reviews and comments, but giving publishers direct access to their fans as well as control over their games and distribution.
(Cartridge may end up developing as more of a Cuckoo, but more on that below)
What all these share is that they may start by piggybacking on Bluesky’s identity, interacting with its community, and writing to and consuming the app.bsky.feed.* lexicons — maybe without even as much as creating a parallel social graph — but they all aim to build their own community of shared views and interests which can, with enough critical mass, survive separate from Bluesky itself.
Their permanence will come from their unique services — Blacksky Cash, Eurosky’s jurisdictional arbitrage, Cartridge’s single-minded focus on gaming. They get to play a positive variant of the old Microsoft playbook: Embrace, Extend, Escape.
Cuckoos
Cuckoos ostensibly integrate into the Bluesky ecosystem, using ATproto identity and storing most of their data on the PDS, while building services and platforms that they get to charge for. They are, in practice, playing a smart game of infrastructure arbitrage — one that depends not on Bluesky’s user base, like Symbionts, but on ATproto’s infrastructure, which no single party controls. Unlike Offshoots, where their infrastructure build-up cost will scale at least linearly with the users, Cuckoos get to flatten the curve by heavily leveraging the existing build-up that already serves the ATmospheric user community.
Cuckoos are orthogonal to Bluesky’s core “messaging and writing” focus. They may initially scan skeuomorphic — Tangled, for example, could easily be dismissed as ”GitHub, but on ATproto”.
I expect every Cuckoo will look distinct from each other, while at the same time scanning initially skeuomorphic. Tangled and Germ are very different beasts, and as I mentioned above, Cartridge may end up being a chimera as well.
This is also the category in which I’d add the burgeoning group of data researchers and collectives, which are currently experimenting with using ATproto to create distributed datasets that no single research organization controls.
A Cuckoo’s survival depends entirely on three factors:
- how non-threatening it looks at first, both to investors not well-versed in ATproto and to other ecosystem participants (including their hosts);
- how well they can leverage the arbitrage opportunity; and
- how far away they are from Bluesky’s core area of interest, to avoid getting cannibalized.
Just the start
This is just the beginning. It’s too early to tell how things may develop and, when I revisit this piece at the next ATmosphereConf, things may have changed significantly.
So there it is — an initial field guide to how this ecosystem may coalesce. Expect it to be thoroughly annotated during the coming year.