The Atmosphere Is Rebuilding The Web, But Better

It’s now been a whole week since I flew to Vancouver (first time in the city, which is lovely) for the second annual Atmosphere Conference. I hadn’t been able to attend the first but only heard good things. So many people have been writing up their experiences and thoughts, and I thought I’d add a few to the pile as well.

1. The Atmosphere is buzzing. The feeling was palpable. So many people just messing around and building cool shit. It really does feel like the early days of the web. The semi-joking phrase “we can just build things” is real. Getting ourselves free from giant silos and API pricing and permission is opening up a wonderful level of creativity. There was so much of it and it’s only just starting.
2. The Atmosphere is not just Bluesky. A point I mentioned a few times in the aftermath. Almost none of the conversations I ended up having with people were about Bluesky, even though it is certainly considered the kind of 800 lb-gorilla in the space (which is a bit ironic, since it feels like the small guy compared to the Big Tech) companies. But so little of the discussion was actually about the microblogging platform, but about other things that can be done and are being done on the protocol. That’s such a good sign.
3. These are my people. Everyone kept talking about “the hallway track” at the conference. The actual schedule was so jampacked with three separate tracks of amazing speakers (every spot in the schedule I felt like I needed to clone myself to go see different sessions), but I still ended up spending a ton of time just chatting with friends new and old in the hallway or lunch area. It really is a community of people I enjoy being around. Thankfully, Streamplace recorded all the videos of the in-room sessions and is setting up a video-on-demand system to view them. Also, the YouAndMe app, which was apparently coded up just before the conference stole the show and made meeting others even more fun.
4. SO MUCH NEW STUFF. I had only vaguely heard of npmx, Semble, pckt, Spark, Blento, and a bunch of other tools. I was familiar with others like Anisota, Leaflet, Offprint, NorthSky, BlackSky, EuroSky, Tangled, Graze, Surf, Flashes, Germ, and Sill, but getting to learn more about all of them was really cool. I know I’m forgetting others too and I apologize! The story of npmx coming together in such a short period of time and building a new, better browser for npm is truly incredible. The incredible thing is how different most of those apps are. There is some overlap, sure, but so many unique approaches to building tools for an open, connected, social web.
5. The ecosystem is unbalanced. To some extent this is always going to be the case in that some organizations in any ecosystem are going to be larger than others, but while people were excited about the Atmosphere and generally supportive of the concept of Bluesky, you could definitely feel a tension between Bluesky and “not Bluesky.” That’s going to happen, but it could do real damage if it festers. There are a bunch of things that need to happen to fix that, and my sense is that people both inside and outside of Bluesky realize that and are thinking through how best to deal with it.
6. The potential of the Atmosphere is revolutionary. Perhaps building on the points above, we’re in a moment where many, many people (both at the Conference and around the world) are starting to not just realize what the Atmosphere represents, but seeing (and building) its broader potential. The excitement at the Conference wasn’t just “ooh, fun toys to build” but “hey, wait, if this really works, think how the world could be different….” Nothing is inevitable, but I really see this as our next best chance to rebuild the web in a way that puts individuals in much greater control.
7. The money question. Many of the Atmosphere projects are labors of love, or speculative ideas, hoping to turn into businesses. Lots of talk both inside and outside the sessions about funding and business models. I mentioned on Bluesky that I’m going to try to write up my larger “framework for thinking about monetization” on an open protocol, but it’s going to take a few weeks. I have something I wrote for myself from two years ago that needs to be updated. Oddly, some of what I wrote 25 years ago about business models on the open web are even more relevant than what I wrote two years ago. I need to sort it all out and will put it into a Techdirt post at some point in April. Hopefully.
8. The Attie stuff. Bluesky launched Attie as a separate tool and it did not go over well (understatement alert). I know it will be unpopular among many but I think Attie represents something incredibly powerful and I think many people are missing its real potential. Releasing just one small feature of it — one that looks kinda like features provided by existing other players in the space — likely distorted the overall vision. But that’s a messaging failure that will need to be corrected. It’s understandable that many in the Bluesky community are sick and tired of AI being shoved into places they don’t want, and we’ve had decades of experience with platforms enshittifying. So I get the initial reaction. But the intention and setup of Attie is very different. It’s an entirely separate project that has some incredible potential, but is entirely voluntary to use, and is very much designed to empower users to do more with the Atmosphere. I think it’s very much along the lines of why I think agentic coding is our best shot at taking back the open web. But the company will need to do a better job getting that across. My own role in all of this, as a board member, is limited, but I want to try to help smooth some of that over, and I’ve been chatting with plenty of people across the community and within the company. Stay tuned.
9. The other divide. It’s kind of interesting to me to see that the “negative” reaction to Attie was very different at the Conference as compared to online on Bluesky. The reaction on Bluesky among many vocal users was simply “fuck AI, no one wants this.” The reaction at the conference was… “oh, hey, this is cool and we do want it, but we’re scared that it’s Bluesky releasing it and potentially undermining others in the community.” That’s also a messaging problem, though a different kind of messaging problem.
10. Growing pains on the path to a new World Wide Web. Some of these points might read as concerns, but they’re the kind of problems you get when something is actually working and experiments are happening. The tensions are real, but so is the collective awareness of them — and the willingness to work through them. That matters.

The crypto bros sometimes claim that the original sin of the internet was leaving out programmable money. I think it was more in how it left out social. The social layer of the internet turns out to be incredibly important. The internet is made of people (or as Anuj Ahooja likes to say “People, not Platforms”) and enabling an open social standard that can’t be captured by a few giant companies is the key. We use the internet to connect with other people. And we need an open social layer. And we’re seeing the early versions of what that open social future could look like in the Atmosphere.