The last week or so has seen a flurry of announcements relating to running independent infrastructure for Bluesky. All of a sudden it is now possible to use the Bluesky network without using infrastructure owned by Bluesky PBC. People can use a separate client, with that client using various other pieces of infrastructure that are self-hosted and not under control of Bluesky PBC. One question that interests me is: why did it take more than a year to get to this point?

Bluesky officially opened up the network in February 2024, and with it was an open protocol that people could build their own apps on. There have been a number of other projects and apps on ATProto, but they mostly are their own application that has only limited interoperability with the Bluesky app. Furthermore, development has mainly been focused on building applications for users.

One of the things that makes ATProto interesting however, is that it takes the software that runs a social networking app, and splits that up into separate components. These infrastructure components (relays and AppViews, in technical terms) can be independently run, and be reused by other parties. 

Up until recently, there have been a few low-key experiments with running independent infrastructure for Bluesky, but that has mostly been contained to people experimenting for themselves, and not making the results accessible to the public. These projects also needed other infrastructure projects in order to be valuable.

What changed in the last week or so is that there are now multiple pieces of independent infrastructure that connects these separate pieces. Apps like Deer are useful in their own right, but in order to add some new features to the app they needed another open backend application (the AppView). It also was the first time when it actually was possible to select another AppView. At this point it actually became feasible to run independent relays and AppViews to get to a point where you can use Bluesky without using Bluesky infrastructure.

This does not answer the question of this all happened now. All the individual components were largely there already for quite some time, and this all could have happened earlier. I do not have a clear explanation for this. But what I do find interesting is how dependent the trajectory of open social networks is on choices made by individuals. It is a small number of people who decide to actually build something to show that something is possible who can meaningfully alter what this next generation of social networks look like. And that is pretty exciting to me.