Some (Slightly Biased) Thoughts On The State Of Decentralized Social Media

Last week, Bluesky, where I am on the board (so feel free to consider this as biased as can be), announced that it had raised a $15 million seed round, and with it announced some plans for building out subscription plans and helping to make the site sustainable (some of which may be very cool — stay tuned). A few days prior to that happening, Bluesky hit 13 million users and continues to grow. It’s still relatively small, but it has now done way more with a smaller team and less money than Twitter did at a similar point in its evolution.

I’m excited with where things are trending with Bluesky for a few reasons, but I wanted to actually talk about something else. Just before I joined the board, I had met up with a group of supporters of “decentralized social media,” who more leaned towards ActivityPub/Mastodon/Threads over Bluesky. Even though I wasn’t officially representing Bluesky, they knew I was a fan of Bluesky and asked me how I viewed the overall decentralized social media landscape.

Similar questions have come up a few times in the last few months, and I thought that it made sense to write about my thoughts on the wider decentralized social media ecosystem, just as we’ve hit the two year anniversary of Elon Musk taking over Twitter. Since then, he’s wiped out billions of dollars in value and revenue, turned what had been a pretty neutral open speech platform that fought globally for free speech, into a one-sided, bot-filled partisan platform that only fights for free speech when it disagrees with the government, but is happy to cave if the authoritarians in charge are friendly with Musk.

But the one key thing is that the decentralized social media landscape has been invigorated and supercharged, almost entirely because of Elon Musk. Thank you, Elon.

I previously told the story of my attendance at a conference in New York in October of 2022, where there was a very interesting presentation predicting the adoption of decentralized alternatives to centralized social media with this chart being shown:

As I noted, this chart and the “events that trigger disillusion” in particular struck me as a bit too underpants gnomey:

What those “events that trigger disillusion” actually are becomes pretty damn important. So, I had asked a question to that effect at the event. For years since my Protocols, Not Platforms paper came out, I had struggled with what would actually lead to real change. I didn’t find the presenter’s answer all that satisfying, but little did I know that literally while that presentation was happening, Elon Musk was officially saying that he would drop his attempt to get out of buying Twitter, and would move forward with the acquisition.

At that point, Bluesky was still just a concept of a protocol. It was far from any sort of app (it wasn’t even clear it was going to be an app). But in the events that followed over the next few weeks and months, as Elon’s approach to dismantling basically everything that he claimed he supported with ExTwitter became clear, Bluesky realized it needed to build its own app.

Indeed, it’s astounding how much Elon has become the one man “events that trigger disillusion” from that chart above. With it, he has become a singular driving force towards driving adoption in alternative platforms.

I’m not betraying any internal secrets in noting that people within Bluesky have referred to some of the big influxes of new users on the platform to “EME: Elon Musk Events.” Whenever he chooses to do something reckless — ban popular users, launch a poorly planned fight with a Brazilian judge, take away the block feature — it seems to drive floods of traffic to Bluesky. But also to other new alternative platforms.

Thank you, Elon, for continuing to supply “events that trigger disillusion.”

But waiting for Elon to fuck up again and again is not a long-term strategy, even if it keeps happening. It is introducing more and more people to the alternatives, and many people are liking what they’ve found. For example, well-known engineer Kelsey Hightower recently left ExTwitter and explained how ATProtocol (which underlies Bluesky and enables much of what’s great about it from a technical standpoint) is one of the most exciting things he’s seen in years.

But, the reality is that no one quite knows what is going to really “click” to make decentralized social media more appealing long term and for more people than centralized social media. Many of us have theories, but the reality is that what makes something really click and go from a niche (or dying!) thing to essential is only possible to understand in retrospect, rather than prospectively.

Just as I spent a few years trying to work out what kinds of things might be “events that trigger disillusion,” I think we’re still in the discovery stage of “events that trigger lasting value.” People leaving the old place because they’re disillusioned is a starting point. It’s an opportunity to show them there are alternatives. But to make it last, we need to create things that people find real value out of that weren’t available at the old place.

The key to every “killer app” on a new system, even ones that start out mimicking the old paradigm, is enabling something that couldn’t be done on the old system. That’s when things get really fun. Early TV was just radio with video until people figured out to embrace the medium. Smartphones were initially just tiny computers, until services that embraced native features like location were better understood.

We need that for decentralized social media.

But right now, we don’t really know what that trigger is going to be. I can think that some of Bluesky’s features — things like domains as handles, using standardized decentralized IDs, composable and stackable moderation, and algorithmic choice — are part of what will get us there, but I don’t know for sure what the big breakthroughs will be. And neither does anyone else.

As such, we need more experiments and experimenting, and not all of that should be done directly within the ATProtocol system (the ATmosphere). Because, even while I think it’s extremely clever in what it enables, the choices made in its approach might limit somethings enabled by other approaches. So I don’t so much see other decentralized social media systems like ActivityPub (Mastodon, Threads, etc.), nostr, Farcaster, Lens, DSNP, etc., as competitors.

Rather, I see them as all presenting unique experiments to see where the real value can show up. I think there’s a ton to learn from all of them. For example, I think Mastodon’s focus on local community and the power of defederation is a fascinating experiment. We’re also seeing some interesting new systems built on ActivityPub that challenge the way we think about decentralized apps. I think that nostr’s simplicity that makes it ridiculously easy for anyone to build clients and relays is important. Farcaster has a number of really cool ideas, including things like Frames that allow you to create apps within social feeds.

In other words, there is a lot of experimentation going on right now, and all of that helps the wider ecosystem of decentralized social media, because we can all learn from each other. We already see that Mastodon has been making changes in response to the things that people like about Bluesky. I’m sure that everyone working on all of these systems are looking at what others are doing and learning from each other.

The simple reality is that right now, no one really knows what will “click.” We don’t know what the real “killer app” is that convinces more people to switch over from centralized systems to decentralized ones. “Events that trigger disillusion” are great for getting people to look. But, getting people to stay and eagerly participate requires adding real value.

I’m happy to see all this experimentation going on to figure out what that is. Just “being decentralized” is not a value that attracts most users. It has to be what that decentralization enables, preferably the kinds of things that a centralized system can’t actually match, that will create the next breakthrough.

Since no one can predict exactly what that breakthrough is, the best way to find out what will really make it work is having the wider decentralized ecosystem all experimenting. This isn’t even a “rising tide lifts all boats” kinda thing. It’s more of a “we need lots of folks digging holes to see where the oil is” kinda thing. Letting each of these systems test things out with their own unique approach is the best way to discover what will actually excite and attract users positively, rather than just in response to yet another Elon Musk Event.

I’m enthusiastic about Bluesky’s approach. I think the ATProtocol gives us the best chance of reaching that breakthrough. But I’m happy to see others trying different ideas as well, because all of these experiments will help bring us to a world where more people embrace decentralized systems (whether they know it or not) and move away from old walled gardens. Not because of “events that trigger disillusion” but because what’s happening over here is just that much more useful and powerful.