There are tens of thousands of researchers in Europe right now. They're producing knowledge, citing each other, following policy debates, and attending conferences. But as a community, their work is scattered across walled knowledge silos. A handful of first-generation Atmosphere apps could change that, without requiring any researcher to do anything they wouldn't want to do anyway.
This post sets out how.
(This is an early draft, written using the permanent versions pattern. Details and version control in the footer.)
The starting point is Bluesky, the most mature ATProtocol app. The building blocks here are ones most Atmosphere users already know: Starter Packs, Custom Feeds, and maybe Labels.
Each EU research project has the building blocks of community: participants, a topic, and a challenge. Bluesky's primitives map onto this naturally.
Each project can create a Starter Pack containing all the researchers in the project, so anyone interested can follow the whole team in one click. The Pack also provides a feed of everything those researchers post... including snaps of pets and holidays.
Hence the starter pack can also highlight up to three Custom Feeds, each driven by a Bluesky List and a hashtag.
If the project has an official Atmosphere account, then its Bluesky account can provide the official project news, while all three custom feeds can be high-value community conversations. I describe these in detail in How newsrooms, scientific institutions & governments can best use Bluesky), but the key idea is that a well-designed Custom Feed is not a broadcast channel. It is a community space.
a well-designed Custom Feed is not a broadcast channel. It is a community space.
A broadcast channel pushes the organisation's official content at followers. A high-value Custom Feed is where the organisation's experts hang out with each other and external experts they trust. Posts appear in the feed not because they're official, but because they're relevant and they come from accounts the project team respects.
This pattern has several benefits that compound over time:
For research communities, this allows a project to use Bluesky not just to announce outputs, but also to convene a conversation with trusted experts around a research area — including voices from outside the project who are doing relevant work.
Note: while Starter Packs can only host three custom feeds, a project can have as many as it likes.
Individual projects are the unit of work, but interesting things can also happen at a higher level. So consider Project Clusters - a group of projects relevant to a specific research strand of a research workprogramme, or a policy area.
On the Atmosphere, a cluster can be managed by a single Cluster Account Manager, creating economies of scale: no need for each project to have its own official account, as the cluster manager handles all starter packs and feeds across all projects.
Moreover, s/he adds cluster custom feeds on top. These operate across all projects in the cluster, surfacing their best content for audiences — policymakers, non-scientific stakeholders, interested public — who want outcomes, not day-to-day project updates. As long as the researchers use Bluesky and remember a hashtag, everything happens automatically.
The incentive for researchers is real: post about your project, get enough reactions to be picked up in the cluster feed, and reach an audience you'd never reach otherwise. And, as set out below, researchers can be connected via another Atmosphere app (Sifa) to their organisation, so those benefits transfer to the researchers' organisation as well.
A single labeller can serve many purposes across the research community. Of course, the only people who see the labels are those who choose to. But if you're a researcher in one project, it is useful to see that the researcher you're replying to on Bluesky is actually involved in another project in the same Cluster.
However, it is difficult to see each individual project setting up a labeler right now. Given that each researcher is in their project's Bluesky list (used for the custom feeds), something at least semi-automated should be possible. When that does become possible, then Cluster and Global labels should also become possible: imagine every scientist working on EU-financed environmental research being labelled "EU-funded" and "Environment science" automatically, making the whole community more discoverable, filterable and legible.
Bluesky makes the community visible. Apps like Sill, Semble and Margin can make its knowledge accessible, while Sifa knits the social network together by providing community member profiles.
A few caveats before going further:
Sill surfaces new content based on what people in your social graph are sharing, liking, and engaging with. For individual researchers, it's already valuable. Collectively, it becomes a knowledge discovery layer for the project.
If I was doing this right now, I'd ask everybody in a project to use Sill for their own benefit (I certainly do), and use Sill's bookmark feature to highlight content they think could be useful to the rest of the project and its community. That's because Sill bookmarks are written to the researcher's PDS (unlike Bluesky bookmarks, which stay on Bluesky's servers), so the resources bookmarked by every researcher in a project list can be aggregated into a Project Library, and the resources bookmarked by the wider project community can become a Community Library.
In the future, when cross-lexicon groups become available, Sill will roll out Sill Groups, a premium feature Tyler demoe'd at my AtmosphereConf 2026 workshop last March in Vancouver. If everyone in a project or community is part of a cross-lexicon group, they can then be easily added to a Sill Group, giving all of them a shared discovery tool for project collective intelligence.
Both Semble and Margin let researchers organise knowledge they discover online — saving, tagging, collecting, connecting. They interoperate extensively, with complementary strengths: Semble lets users connect resources to each other, for example, while Margin enables more granular organisation plus inline annotation, with notes presented on the original knowledge.
For the purposes of this post, the underlying idea is the same: if each researcher in a project uses Semble or Margin, whoever manages the project or community can aggregate across their PDSs to produce a shared Project Library or Community Library of everything the team has curated, collected and commented on.
SIFA is building into the space currently occupied by LinkedIn. In fact, you can import your LinkedIn data into SIFA to create a starting point for your Sifa profile. Unlike LinkedIn, however, SIFA is on the Atmosphere, so:
SIFA accounts are therefore important connective tissue, providing the core product all communities need: community member profiles.
discovering people through knowledge, and discovering knowledge through people
This is vital, because communities are all about discovering people through knowledge, and discovering knowledge through people. Those connections happen through community profiles - if you see something from somebody on any Atmosphere app that you find interesting, you can visit their Sifa profile and discover everything they find interesting.
Each of the above apps present the slice of the community's knowledge the app was used to create. But all this content is freely available on researchers' PDSs. And we know — via the project lists — who is in which project, and hence which project cluster.
That means all of the following can be brought together:
Collected into a database (or via a live interface querying the PDSs directly — the architecture is an open choice), this content can be searched and filtered by:
The same sources can be used to create all sorts of project feeds:
While these databases can be localised to the project websites, there's nothing structurally different between doing this for one project, a cluster of projects, or entire research programmes. The tools just need to know the IDs of the relevant project and community lists.
In Europe, where so much research is funded by the EU programmes, it becomes feasible to create something genuinely new: the living, searchable collective intelligence of thousands of EU-funded researchers. Not a static publication database, but a real-time record of what the research community is discovering, reading, debating, and connecting, right now and in the past.
The above is already valuable, but principally for scientists. But the content assembled at Cluster level, in particular, creates further opportunities by providing useful raw material which can be transformed. In particular:
Researchers don't need to do anything to provide their knowledge - by using the project-level tools set out earlier, their work can be picked up and used by these communications specialists. But it does make them and their work more directly accessible to everyone, which cannot be a bad thing.
None of the above requires researchers to do anything they wouldn't already want to do.
Custom feeds are useful whether or not anyone is aggregating them. Sill bookmarks help you remember what you found. Semble and Margin help you organise what you know. SIFA gives you a better professional profile than LinkedIn.
The added value policymaker/public-friendly knowledge layer is built on top by specialists using the knowledge created by researchers using tools that serve them individually. Everyone does their own thing, and they can because the knowledge flows seamlessly rather than being locked away in walled gardens.
the knowledge flows seamlessly rather than being locked away in walled gardens
What we could build: a living picture of what Europe's research community is discovering, reading, debating, and connecting right now — visible to other researchers, to policymakers, to science communicators, and to the next generation of researchers trying to find interesting conversations. Not a static project database, but a living, interactive community... but with a hugely powerful, enshittification-resistant memory, .
This is one of this wiki's pages managed with the permanent versions pattern described in Two wiki authors and a blogger walk into a bar…