IPFS at Eurosky Live Berlin: Highlights From A Bright Future

The Authenticated Transfer Protocol (ATProto or AT), which sits under Bluesky and many other emerging projects, has tremendous potential. At the IPFS Foundation, we're big fans, and not just because AT builds on the IPFS stack through IPLD and DASL (though admittedly building on DASL is one of the coolest things anyone can do). ATProto deals in self-certifying data structures, which removes location-based authority, an idea that aligns directly with our own IPFS Principles. We believe that AT is not just the future of social media but also the future of the web, a web built atop social primitives that reflect how humans naturally organize and collaborate. It has identity as a core primitive, it guarantees that people control their own data, and it prevents any app from centralizing value by hoarding data by baking adversarial interoperability right into the protocol. There's a lot that needs working out still, but suffice it to say: it's hard to overstate AT's promise.

Code and protocols are great of course, but we should learn from previous mistakes. If all you have is code or protocols, you still don't have the means to keep a system open and democratic. I could give you the entire source code for, say, Instagram tomorrow morning but that would not allow you to take over and compete with Meta. I could also give you all the data they have, and that would take you no closer. They control the location and, crucially, they control the infrastructure, the binding and localized contracts, the (technical, legal, and psychological) trust between all the parties. That means they would keep the users and the advertisers.

A similar problem could arise with AT. The protocol can have the loveliest decentralization and agency properties, but if at the end of the day Bluesky PBC owns most of the infrastructure underlying the AT network, then they own AT. Don't get me wrong, everyone I've met from Bluesky has been nothing short of delightful. But we can't stake the future of the web just on people being nice and staying nice, on no future coup happening and replacing them with less nice people. As Bluesky CTO Paul Frazee often says, "the company is a future adversary."

That's where the ATProto community is taking Bluesky PBC up on its founding intent and is getting independent infrastructure running. Groups like Blacksky, Northsky, or Eurosky are working on it. Far from competing with one another, all of these groups understand that they all benefit from each other's successes, and indeed Evelyn Osman presented NorthSky at this Eurosky event. The goal of Eurosky is to operate independent AT infrastructure (in Europe, as the name may have hinted at) to help guarantee that the AT network as a whole can be operated as a commons. Eurosky is a project of the brand new Modal Foundation, good friends of the IPFS Foundation.

On November 19, in Berlin, Eurosky ran its first public event: Eurosky Live, which coincidentally took place on the day after the less concrete Summit on European Digital Sovereignty.

I won't write up a full blow-by-blow of the event — for the curious among you who want it all and want it now, you can watch the whole event on video. (The entire event was streamed live and captured on the excellent Streamplace, a video streaming system built atop AT Proto, which worked without a hitch. Gotta love the dogfooding. What I'll focus on below is primarily aspects that are more relevant to the IPFS community — it was a packed day, I'm skipping a lot. You can read other takes elsewhere, like OpenFuture's Eurosky Dawns, Mathew Lowry's Politics meets protocols in Berlin, coverage in major German network RND that gets syndicated to millions, and highlights from the social web's star reporter Laurens Hof over at Connected Places.

Unbundling Ownership, Coordinating Communities

The week kicked off with an informal meeting of multiple groups contributing to the ATProto ecosystem two days before the event to discuss the many things that Eurosky and allies could work on. (There was also an atproto.berlin event on the 18th, but sadly I couldn't make it.)

One point of agreement in the room was that none of us wanted there to be one organization owning AT Protocol. It's pretty common for programming languages or protocols to be owned by a foundation or a standards organization that acts as both technical authority and public voice (like the PSF for Python or W3C for CSS), but that seems inappropriate in this case. Good governance requires unbundling responsibilities — something that AT Proto itself does particularly well — and we felt that the institutional ecosystem of AT should live up to those principles. It does look messier and it complicates representing AT in spaces in which having a central spokesperson would help, but it sets the community up for longer term success. The IPFS Foundation can keep shepherding DASL, the IETF can house the repo and sync specifications, the community can manage its own lexica, we can have ATProtocol Dev run conferences and organize funding, Modal can host Eurosky, Gander and Northsky can build their own flavors of sovereignty and independence, and more. Democracy looks like a hairball — that's what makes it powerful and resilient.

We agreed that we'd work on shared material to make our hairball easier to comprehend, to support one another in getting funding, and would consider developing some products as a community. Overall, it was lovely simply to take the time to get to know one another when many had only ever met online previously.

Policy Meets Tech

Perhaps the most remarkable aspect of Eurosky Live was that the single-track event had both policymakers and technologists on stage and in the audience, and all attended one another's sessions, often asking questions and participating. It was a surprising but, I thought, effective mix. We had both the likes of Alexandra Geese, Member of the European Parliament, or Heike Raab, Representative of the State of Rhineland-Palatinate and Coordinator of the Broadcasting Commission of the German States on the one hand, and techies like Anirudh Oppiliappan showcasing Tangled or Sebastian Vogelsang talking about Eurosky infrastructure and Flashes. You don't see that every day.

The crossover made a compelling case to policymakers that AT was a thriving emerging area and not just the Bluesky startup. I know that several were impressed with Boris Mann's whirlwind tour of AT projects and with Sill, a multiprotocol reader that highlights news from your trusted connections, gathering their posts across Bluesky and Mastodon. Conversely, it showed to technologists that policymakers have real demand for and real interest in concrete alternatives. From a distance, policymaker attention can appear overly focused on AI Gigafactories but that's not the case. More often than not, they're thirsting to see real-world projects that show what kind of other world is possible — it's a lot easier to build good policy if you know other ways in which things could work — and bring the demo rather than the memo.

Protocols & Infrastructure

A point I made in my own presentation was that tech monopolies draw tremendous power from the infrastructure that they've captured and, simultaneously, that the architecture of a typical platform service today, like legacy social media, is so monolithic as to be ungovernable. We need to both reclaim infrastructure — which comes with its own headaches in terms of funding, governance, and regulatory responsibilities — and to design technology such that responsibilities for the various components that go into building a user-facing app can be unbundled, operated by different stakeholders, and seamlessly glued back together using stable, openly-governed protocols.

I almost wrote "getting that right is harder than it sounds" but let's be honest: it already sounds pretty damn hard. We're tackling similar challenges with IPFS: balancing modularity with usability, content moderation, Bad Bits, gateways, the challenges with supporting the public Amino network, who should fund open source implementations, who should be in charge of standards… But hard as they may be, they're not problems that we can just ignore, and they’re a fair tradeoff to ensure the benefits of a truly open web. One thing that struck me at Eurosky Live was the general acceptance that these are issues we need to own and to solve together, one by one. We'll just roll up our sleeves and do it, one chunk at a time.

None of us working on next-generation decentralized protocols — AT, IPFS, but also Beckn, the Decentralized Identity communities — and the communities that most critically depend on our success, notably for the protection of science, journalism, or democracy, need to work alone. We have shared problems and we can share our solutions. More than anything else, it feels like a new era of collaboration is upon us, a time when the small cooperative mammals can outsmart lumbering dinosaurs. I can't wait.

What's Next?

There's plenty to do, expect many new announcements of cool things from the community in 2026. Our next gatherings (that we know of yet) will be ATmosphereConf (Vancouver, March 26-29), long-time classic Re:publica (Berlin, May 18-20), and Public Spaces (Amsterdam, sometime in June). 

I hope we'll meet there!