Paul Frazee, CTO of Bluesky, gives a short talk at ATmosphereConf Seattle 2025.

Introduction

Thank you. So, today I’m going to talk a little bit about how we got to where we are, where we’re going next, and the ways I think we can make this future better together.

I’ve been working in decentralized social networking for over a decade now, and in that time I’ve seen the same ideas rise, fall, and rise again. I think it’s helpful to zoom out and look at the patterns that have led us here, because those patterns can help us understand what comes next — and how we might do it better.

The 2000s: The Lost Decade of Standards

If we go back to the early 2000s, the web was in a very different place. There was a lot of optimism around open standards and interoperability — things like RSS and Atom, which gave us open syndication, and the early work on open identity systems like OpenID.

But that decade was also when the big social platforms began to consolidate. MySpace, Facebook, and later Twitter started drawing everyone into centralized silos. For a while, it seemed like that was just the way the internet was going to be.

What’s interesting is that the technology for decentralization already existed. We already had the tools — open protocols, APIs, standards bodies — but we lost sight of them. Business incentives pulled us toward closed platforms, and we didn’t yet have the cultural momentum to resist that shift.

The 2010s: Rethinking the Web

The 2010s were a period of reevaluation. People started to realize that the trade-offs of centralization were real — not just technical, but social, political, and personal.

We saw the rise of projects like Mastodon and ActivityPub, peer-to-peer protocols like Secure Scuttlebutt, and decentralized identity work like DID and Verifiable Credentials.

It was also a time of frustration. These projects showed what could be possible, but they also revealed how hard it is to build something people actually use. We learned that decentralized tech has to compete not just on principle, but on user experience.

The biggest lesson of that decade was this: decentralization has to be invisible. People don’t want to manage cryptographic keys or configure servers. They just want tools that work, and that respect their autonomy.

The Blue Sky Years

When we started Blue Sky, our goal wasn’t to make “Twitter but decentralized.” We wanted to solve the underlying problem: how do we make a social web that’s open, composable, and still easy to use?

That’s where the AT Protocol came from. We wanted to combine the values of decentralization — control, portability, transparency — with the usability and reliability that people expect from modern apps.

A lot of our work has been about finding that balance:

• How do you make moderation decentralized but still effective?

• How do you give users choice without overwhelming them?

• How do you keep things fast and familiar while keeping data portable?

Those are hard questions, but they’re also exciting ones. Because solving them doesn’t just fix social media — it redefines what social software can be.

Decentralization as an Ecosystem

One way I like to think about this is that centralized platforms are like gardens: they’re well-maintained, curated, and orderly, but you only get to see what the gardener allows.

Decentralized systems are more like jungles: they’re messy, chaotic, and full of unexpected growth — but they’re also alive and self-sustaining.

The goal isn’t to turn the jungle into a garden. It’s to build tools that help people navigate the jungle — tools that give them agency and help them find their own communities, values, and norms within it.

The AT Protocol is designed to support that kind of diversity. It lets many different apps, servers, and cultures coexist, while still being part of a shared social fabric.

Building Beyond Platforms

What excites me most is what comes after the platform era. Once we have open protocols for social data, we can start to imagine new kinds of applications entirely — ones that aren’t owned by anyone, that evolve organically as communities use them.

We can separate identity, content, and algorithms, so that each can evolve independently. Users can move freely between apps. Developers can innovate without permission. Moderation can be community-driven instead of company-driven.

That’s the internet we should have had all along. And now we actually have the chance to build it.

The Road Ahead

There’s still a lot to do. We need better tooling for developers, more scalable infrastructure, and stronger social norms around moderation and trust.

We also need to make sure that as this ecosystem grows, we don’t just recreate the same power dynamics we were trying to escape. Openness alone isn’t enough — we have to actively design for equity, inclusion, and resilience.

But I’m optimistic. For the first time in a long time, it feels like the energy is back — not just in terms of technology, but in terms of imagination. We’re remembering that the internet doesn’t have to be a handful of walled gardens. It can be a living, connected network again.

Conclusion

So, where did we come from? We came from an open web that lost its way. Where are we going? Toward a decentralized web that learns from that past.

And how do we get there? By working together — not as competing platforms, but as a community of builders, designers, and users who believe that social media can be something better.

Thank you.

The videos from ATmosphereConf 2025 held in Seattle, Washington, are being republished along with transcripts as part of the process of preparing for ATmosphereConf 2026, taking place March 26th - 29th in Vancouver, Canada.

Follow the conference updates and register to join us!