We gathered in Vancouver on March 27, 2026, for a full day exploration of AT Protocol for science, education, and open knowledge. The gathering brought together researchers, developers, builders and curious passers-by. ATScience was part of AtmosphereConf, the main global AT Protocol conference, which happened later that week on March 28-29.
Registration to ATScience was maxed out at 125 registrants, and the day was buzzing with energy, inspiration and exciting new projects, with new connections being sparked every which way.
The energy at #atscience is wild. @mariaa.bsky.social gave a wonderful talk about Lea, an AT network app for researchers. So so so inspired by the hope in the room right now. Rewilding in real time! @loriemerson.net, @mariafarrell.eurosky.social and @robin.berjon.com mentioned #atmosphereconf
We had talks, panels and facilitated discussions spanning a wide range of topics - data, AI, social media research, collective sensemaking systems, and new tools for researchers on Bluesky/atproto. Check out the talk recordings here:
Bridge building at ATScience
One of the strongest themes to emerge was bridge-building: between ATScience and the wider atproto developer community, the modular research movement, and broader questions about social media, epistemic environments, and the future of democracy.
We saw this play out in the cross-pollination between ATScience and the broader AtmosphereConf, with themes from Friday reverberating throughout the rest of the weekend:
ATScience <> Society
mentioned ATScience multiple times in her wonderful AtmosphereConf keynote, as a community already demonstrating the promise of ATProto for building healthier information environments:
As the science community is amply demonstrating, we can develop features that cluster and amplify the knowledge people assemble...
However, Erin also emphasized that this work matters not just for science, but for society and our democracies more broadly:
... but if we only do this work for highly motivated specialists, we're missing the chance to bring that fire to everyone else.
During ATScience itself we had a whole track dedicated to collective sensemaking systems: presented ViewSift, a social app to reduce division. presented SkySquare, which uses social media data to provide a social context layer over existing content. presented Semble as a knowledge network for researchers but also as a general tool for collectively mapping information.
ATScience <> Atmosphere
A meeting between (building Roundabout at ) and building ) sparked an ongoing collaboration around schema translation and category theory powered interoperability.
We also had some great followup conversations with of Astrosky and at Modal Foundation/Eurosky that sparked a new initiative to set up a PDS and portal for researchers - stay tuned!
We talked about this in @atproto.science at AtmosphereConf! Running a PDS for science similar to Eurosky @emily.space @tgoerke.bsky.social @byarielm.fyi @mathewlowry.eurosky.social
ATScience <> Research Institutions
's ATScience workshop explored the role of ATProto in supporting research institutions:
your institution no longer has to rely on Twitter, Facebook, or LinkedIn to host conversations about its content. Instead, you can integrate your website directly into the social conversation on the ATmosphere, bringing the discourse back to where the content lives rather than scattering it across platforms you don't control
The workshop walked through this with three complementary contributions: (Leaflet.pub) on publishing via standard.site, (Sill) on collective listening via shared social graphs, and (Semble) on collective curation via Groups.
ATScience <> Open Science / Modular Research
Our keynote speakers were Rowan Cockett ( , ) and Matt Akamatsu ( University of Washington, ). Watching their keynote is like looking into the future of science; they are both leading exciting new initiatives around open, continuous and modular publishing.
Matt Akamatsu on building the future of research labs with Discourse Graphs
Rowan hadn't previously encountered AT Protocol, but on connecting to the ecosystem he immediately saw the potential. He and his team went on to publish a blog post titled "Scientific Documents as First-Class Objects on AT Protocol," which observes that
The properties that make AT Protocol compelling for social networking are the same properties the research community has been asking for
Post-event survey
We were a bit late with our post-conference survey (note to selves for next time!), and so the response rate was relatively low - only 12% (15/125). Nonetheless, the responses still provide valuable qualitative signal and highlight some consistent themes:
- 1.
ATScience 2026 drew a technically engaged crowd: researchers and developers dominated, with most arriving as "informed enthusiasts" already familiar with AT Protocol.
- 2.
Sentiment was exceptionally strong: the average event rating was 4.9/5, 14 of 15 would attend again. And as the Sankey diagram shows, no one left less enthusiastic than they arrived.
- 3.
The most consistent theme in open responses was the value of community connection and cross-disciplinary conversation, while the main improvement request was more unstructured networking time and better unconference coordination.
Thanks Claude for the visualization help!
Notable comments:
“Seeing other engineers so excited about what could be possible was so inspiring to start making applications with love and care.”
“Sometimes the tech lens has us think myopically. AT Science Day lets you grow how you think about tech (AT proto specifically) and problem solving.” — Daryl Ducharme (Open source developer advocate @ Google )
“ATScience was an incredible opportunity to learn more about the protocol and new projects being built for researchers, and to spark and get feedback on new ideas. I left inspired by the support and energy of this community.” — (PhD @ Cornell Tech , Paper Skygest)
Parting thoughts
Most ATScience projects are scrappy, under-resourced and volunteer led passion projects. One participant described a "common phenotype of assistant professors who are being entrepreneurial and need more time to build," and a refrain we heard throughout the event was: "if only I could quit my job and work on this full time." Even so, these projects already demonstrate the promise of ATScience. Imagine what the ecosystem could do with funding to match.
We’re tracking the potential of ATScience in this open Semble collection - feel free to contribute as well.
Thanks…
to the ATScience organizing team who spent 4 months going from 0 to full day conference, to for instigating the idea and to the fantastic and teams for their close support.
Thanks to for event sponsorship and travel grants for speakers. Finally, thanks to all the speakers and participants who made this such a special day.
See you in the Atmosphere and hopefully also next year!
ATScience Organizing Team: