I can tell DevRel season is starting back up again, because I counted the number of flights I currently have booked in the next 6 months, and came back with 18. Granted, some of these are vacations, and I'm a bit of an over-preparer when it comes to vacations (PLEASE do not start me on a conversation about min-maxing Chase points, I will embarrass us both), but a big part of that is the job! I'm usually on the road a little more than once a month from January-May, and while I think that's a mostly sustainable cadence (and appreciate getting face time with AT heads from all across the world), it does come on pretty strong every year.
Anyway, this is all a roundabout way of complaining/bragging that I've so far had 2 talks accepted since starting in this role, which is a nice little streak! One isn't for many months from now — it'll be at the first North American event of WeAreDevelopers, which is co-presented with Docker, and is I think intended to be a successor to DockerCon. WeAreDevelopers' Berlin events have always been good, so I'm looking forward to this when it happens next September. The other talk is barely more than a month away — at FOSDEM's Decentralised Communication developer room in Brussels at the end of January! This will be basically my first proper AT Protocol talk, and figuring out what points that I personally like to hit is one of the most fun parts of a new job. I know how I write about this project, but that's never quite the same thing as how to talk about a project. Milestones!
I was also already planning on stopping in London on the way to Brussels to check out the new Punchdrunk production (I am a big immersive theater head) so it was nice to hear there's an ATProto London event being planned for the Wednesday before FOSDEM. Technically I guess that will steal the "my first AT Proto talk" spot due to scheduling 🙂.
If I can be overly sincere for a minute, I'm really thrilled to be back in a job for which FOSDEM is a significant event. As you've probably gleaned from me talking about having used desktop Linux on and off throughout my career, I care a lot about open source, but I make a very poor zealot, and have worked several commercial open source roles that have more and less emphasis on the "commercial" part. Put another way, I thrive in workplaces where at least one of my colleagues runs Arch Linux despite it being my job to support slightly more popular operating systems. For a few years before joining Bluesky, I had been doing more Kubecon-style or Reinvent-style Linux work (you know what I mean) than FOSDEM-style work, and I'm really excited to be skewing just a little more granola-heavy again. Who needs a business model anyway? KIDDING. I'M KIDDING.
My colleague Bryan Newbold's exchanges with the Fediverse developer Christine Lemmer-Webber were a major inspiration to me in deciding how to approach FOSDEM. This is an audience for whom AT Proto is simply one protocol among many (in the best possible way), and only good things can come from us continuing to learn from other projects in this space. Fun fact: our internal metrics currently suggest that the number of posts originating from Bridgy Fed, which facilitates cross-posting from Mastodon, is pretty close to the number of posts coming from self-hosted PDSes. And that's not even counting projects like Wafrn that implement that integration on their own!
Actually, that reminds me — just before I accepted this job, I was at the Hackers on Planet Earth conference in NYC last August, which was a similarly granola-y breath of fresh air for me, and it was there that I happened to see my first AT Proto talk in the wild! Go take a look at the abstract, I believe the conference was recorded too — it was a good one, and they're having to find a new venue for next year, I wish them nothing but the best.
Other big news from us: just after I published my blog last week, we announced the release of tap, a new tool that significantly streamlines the process of firehose mirroring and synchronization. This is really welcome news for anyone who actually wants to backfill a significant amount of AT Protocol data (e.g. for doing local data analysis on a many-TB corpus of Bluesky posts) and then continue receiving new data from the Firehose. Prior to this, streaming new data was pretty straightforward, and backfilling was doable if somewhat manual, but the cutover was no fun at all, and this is now a solved problem.
Funnily enough, releasing a tool that makes heavyweight backfilling easier happened to coincide with the release of new lightweight tooling that doesn't require backfilling: I think Slices' new Quickslice is extremely cool, and a great example of how accessible AT Protocol social graphs can be. Communicating the value of tap-style backfilling vs. Quickslice or Microcosm-style backlinking is a fun ongoing challenge for me/us, and community member Orual also wrote a great blog about exactly that this week.
In other news:
Our moderation tooling is getting a lot of love right now; we'll have some new tutorials to coincide with our docs relaunch next month, and the self-hosting story here is much less scary than it used to be.
The Germ Network devs did a great couple of posts this month about their journey so far; private messaging is another topic we're thinking very, very hard about right now, and it's great to see such thoughtful "development in public."
We shipped Contacts import on Bluesky! It's double-opt-in! I was a tester! Little by little we are filling in that roadmap.