ActivityPub 2020: Evan Prodromou

To kick off the ActivityPub conference we had Evan Prodromou, who is perhaps the father of Federated social media due to his creation of and He was also the Chair of the W3C’s Federated Social Web Community Group, served on the W3C’s Social Web Working Group, and is a co-editor of ActivityPub. His keynote talk was titled What We Need For the ActivityPub Network.

He begins by making a very pertinent point, which is that Federated Social Media does not need to, and indeed should not, replicate the media of surveillance capitalism, with just some added “free and distributed” sauce. Those media take their form for specific reasons that aid and abet the business models of the companies (e.g. Facebook, Twitter) to the detriment of our communities. This is why they have to promote “engagement” above all else, because that is their funding source. And if promoting fake news from Russian hackers and QAnon will promote engagement, they will lean towards doing it.

Federated social media, in contrast:

  • Does not have that business model
  • Does not need to scale to 1 billion people
  • Does not have the same legal requirements
  • (Usually) does not need to worry about app stores.

And if we think back to the early days of social media, we might remember that Facebook and Twitter had APIs that let other people write apps using the platforms in interesting ways. Of course, once the platforms reached critical mass they closed the APIs and killed all of those other apps. Federated social media has no need to do this. We could build within ActivityPub an API to let people connect and use information. And when you add the idea of a server you can trust it gets better. (Remember, you can at any time run your own server for any federated app. Or you can join with others on a server you can trust.) You can then have social media that meets your needs, not Mark Zuckerberg’s. And no one is mining your posts looking for hooks that they can use to guide your thinking or sell you something.

These are what he thinks of as his guiding principles, and they make sense. But of course there is more to do.

What we need

Evan starts looking at the needs by focusing on the fundamentals. Social media is about receiving, storing, and re-distributing posts. Evan suggest we need a lot more independent servers doing this because a diverse population is more robust. But he would separate the back-end from the lights and lipstick. He wants to see these servers just focus on the three tasks of receive, store, and re-distribute. Then let other layers in other places work on building the Web and mobile interfaces, of which there could be many. In a way this seems similar to the Web, where the basic plumbing of HTTP is common to every use, but on that plumbing you can have different Web sites, different Web browsers, and so on. On Facebook, they have rolled out a new interface recently which most people hate, but it doesn’t matter what people think if it helps Mark Zuckerberg. They will get it anyway.

To get there we need to have a standardized API within ActivityPub for Client apps, such as:

  • Games – We had a lot of games on social media, such as FarmVille and Scrabble. My wife loves to play word games online with her friends, which is after all what the social in social media is about. We could have that in Federated media, but without needing to monetize every eyeball.
  • Different Media – We could have apps that helped us to record audio and video that we then share with our friends.
  • A Yo app – This basically sends a text to your friends saying “Yo”.
  • Suggesters – Finding people when you are on a new network is difficult. I now have connected to a number of my Hacker Public Radio and Open Source friends on Mastodon, but it took a while.
  • Reposters – ActivityPub is great for this. I can post a photo on PixelFed and have it automatically re-post to Mastodon. But we could do better, and Diaspora is so far out of the loop because it does not use ActivityPub.
  • Blockers – You might subscribe to a block list. I am seeing some activity like this on Diaspora as Neo-Nazis and QAnon types come in and troll people. But on Diaspora each person has to actively manage their own block list.
  • Trimmers – Pare down your list of “friends”. If you haven’t exchanged messages with someone for over a year, maybe they don’t need to be on your “Friends” list.

Think small

Since we do not have a business model that requires monetizing every possible eyeball, we don’t need to have monolithic servers that “bring the whole world together”. We can focus on affinity groups of various kinds, such as:

  • Churches
  • Neighborhoods
  • Cities
  • Families
  • Schools
  • Universities – This is how Facebook began, remember.

Groups like this frequently are already a social network, they just don’t have all of the infrastructure. I see in my own family how awkwardly we try to do it with e-mail. And anyone who has ever belonged to a Church, Synagogue, Mosque, etc. knows that these are definitely social networks.

Along with this, Evan envisions a whole different type of server architecture, like zero-config, zero-maintenance boxes that are essentially home appliances. You would buy one, plug it in to power in your house, it would find your WiFi and connect to it, get an address from Dynamic DNS, and then you could connect with your phone, laptop, or tablet. This is where you can really begin to talk about distributed social networks. “Here’s a box, plug it in, it works.”

Federation of Things

We have all heard about the Internet of Things, and how it is awful and unsafe in so many ways. But happens if you have super-simple smart devices that report back to a server you control. We can start to see a different possibility here because the focus is on our control. You could have sensors all over, but instead of reporting to some company far away with its own interests at heart, they could be reporting to your own server in a way that is safe, secure, private, and respects your autonomy. It will take a lot of work to get there, of course, but ActivityPub opens up possibilities we did not have before.

Quantified Self

I wear a Fitbit, and it reports back to the Fitbit servers. The data is for me an opportunity, but having it on someone else’s server is a liability. Who else might access that data, and what would they use it for? As someone with several health issues, I have begun to live by the numbers. I test my blood sugar daily, I keep track of my exercise, my weight, my sleep, my blood pressure, my heartbeat, and so on. These are all ways I stay healthy, but I don’t necessarily want to share them all with the world. As with the Federation of Things, imagine what happens if instead the data only goes to a server I control, and I can decide who gets to see it. In my case, my wife and my doctor would get it, but the point is that I get to decide. I am gathering cool data now which I sometimes share on social media, but that is through companies who may use my data in ways I would not like. Having an ActivityPub Server would give me more control.

Personal Digital Archive

How about a server that you control that has your life history. I do a lot of that right now with Web sites and Web services, but the degree of control varies. I own a few domains and have set up Web sites on them. Right now they are hosted by a hosting company, but at least I get some control over the content this way. There are costs involved, but the obstacles to running my own server right now are not trivial, and there would be costs there as well. My photo account on Flickr is not free. I have some audio stored on the Internet Archive, a video or two on YouTube, and those I have somewhat less control over. But family history is valuable, and worth maintaining if you can do it safely.


The Social Media we have exists to engage us continuously for the purpose of making money (“You won’t believe what happens next!!”) That it is peddling lies and making us less happy is not the point, exactly, but it is an acceptable by-product of maximizing engagement. With Federated, decentralized social media we could instead devise platforms that make people happier and more satisfied. We could optimize for connecting people in a good way, instead of feeding conflict. Instead of doom-scrolling, we could move to awesome-scrolling.

So instead of replicating social networks that never served us well, we can look to building networks that serve us the way we need to be served. We should strike out in a new direction.

Evan Prodromou is clearly someone who has thought both long and deeply about social media, and this was a great opening keynote for ActivityPub Conference 2020.

 Save as PDF