As a systems architect, I am compelled to always consider two things among many: system capacity and costs. This translates into two questions: “how does it scale in capacity?” and “how much does it cost, and who pays?” These are two questions that few are asking when they place their trust in federated systems like the Fediverse. Let me be clear, I find it amusing and it’s a place where I enjoy being. But this does NOT mean that I truly believe everything will be that way one day.
If we examine who pays today and how it scales, we have two types of instances. There are the self-hosted ones at home, which are essentially “offered” by a kind of volunteer/patron who enjoys being a sysop, and there are larger instances that ask people to make contributions in order to support the hosting costs. I don’t want to dwell on whether billions of people accustomed to having social media for free can truly start paying for its costs. Instead, I am looking at the other side and wondering how many administrators are willing to personally shoulder the costs or rely on some form of donations.
Indeed, the model of the Fediverse as an OnlyFans for system engineers, where they provide the instance and then ask for donations, does not scale well. First and foremost, attempting to replace global players based on a network of volunteers and donations is simply impossible. You would end up with a rather small network, much like the Fediverse is today. But scaling would be a challenging task.
But let’s even imagine that global consciousness suddenly awakens, and people are all willing to join the Fediverse. Very well. Very well, “BUT.” We’re talking about multiplying the current capacity by at least a thousandfold. (Taking into account bots and multiple accounts for peace of mind and accommodating 4 billion people, the “comfortable” order of magnitude would be in the tens of billions. Today, we’re talking about tens of millions.) And if you think that META or Alphabet’s electricity bill is high, considering that a hyperscale data center is much more efficient compared to our homes, you’ll soon realize that the first question is: yes, but who pays for the electricity? Will it be the volunteers and those who make donations? Do you think that’s a number of people that can scale by a factor of 1,000?
I believe that the Fediverse has reached the limit of volunteers running instances from home, relying on their own internet connection, and even those who pay for cloud hosting are unlikely to grow significantly, perhaps by a factor of 10 at most. But how much will the number of paying customers, who will then join the Fediverse, scale? How many people accustomed to having everything for free will decide that it’s worth paying for their own instance to gain access?
Whichever way I look at the issue of economic sustainability, I always see the same scenario. Increasing costs and not-so-increasing revenues. I highly doubt that the recent waves of Twitter users have brought a donation boom to the larger instances. I believe it’s more likely that they haven’t become rich. Of course, it would be nice if the larger instances published more detailed data on their finances and economy, but strangely enough, I don’t see any of that happening.
It is clear that we will reach a breaking point, and the larger instances will reach it first. While volunteers may be able to withstand it because we enjoy doing it, if we ask ourselves, “What happens if Eugen Rochko’s company fails?” all the answers are extremely bleak. Not so much because we would lose the development of Mastodon (which is not the only software), but because the rest of the network would have to handle something like two million users. There are certainly many instances, but they are in the order of thousands. For them, accommodating one or two thousand additional users would be impossible. As for my instance, for example, it may have around a hundred users. And I certainly won’t be purchasing new hardware to accommodate a group of non-paying users.
Some people say that when the world realizes how cool the Fediverse is, “private individuals” will come forward, while others say that “the government” will step in. The idea of a private individual sustaining an instance without any economic return honestly escapes me. So, either they will make you pay for your presence, or (once again) they will monetize your data using the same old methods. Maybe (again: maybe) might work because the “Joe’s Pub” instance may be interested in keeping a group of loyal customers in contact. Or an instance like “Hardcore Metalheads” may benefit from partnerships with clothing stores, record labels, and the like. But in any case, either you pay for access or you become the product. It doesn’t seem like a groundbreaking change to me.
The idea of the government providing public spaces like the Fediverse honestly scares me. It scares me because “the government” can refer to a political party, depending on the moment. And it’s very easy to upset a political party by saying the wrong thing. A Fediverse under state control would effectively be under political control. Sure, maybe city mayors might want to create an online community of citizens, but it would always end up censoring those who criticize the mayor “too much.” And you understand that this is not a good thing: the notion that the future of the Fediverse relies on the goodwill of the government to allow free criticism seems too naive and optimistic. In short: foolish.
So the problem now is: how do we see it scaling? And who pays? I believe that in the end, we will have a mix. The enthusiasts will still be there, as they are necessary for large-scale software testing. There will still be non-profit organizations because there will always be groups driven by certain ideals that can be funded through donations. I think governments, as they are currently doing, will continue to have their own instances to ensure the identity of those who write: if you are on the government’s instance, you belong to the government. And I think that some private individuals who find it beneficial to have a community (a pub, a football team, a local cultural entity, even a record store or anything with a communal aspect) will join as well.
And I believe that the Fediverse is waiting for a new type of instance: the paid one, where you pay a few euros and get something in return. The problem with these instances is that they involve an economic contract. Let’s set aside the question of how the payment would be made, whether it’s Bitcoin or credit card. In the latter case, the KYC (Know Your Customer) principle would be satisfied. But the issue is, “What do I write in the contract? In other words, what are the rights and obligations of the contracting parties?” We could say that the instance provides a specific type of moderation, tailored to certain interests. In practice, they would be like club instances. You need to have a particular passion (such as a type of music or gardening) to join the instance that suits you. The question is, why pay when you can go to free instances? And the answer would be that the free instances don’t have enough capacity to cover the costs of four billion users.
These paid instances wouldn’t be anything new. During the BBS era, you could indeed join free BBSs on Fidonet, which then interconnected and federated. But you could also pay a small fee and choose Compuserve, which was a similar but private network. Why did people use Compuserve instead of Fidonet? Because it provided reliability to individuals who created their own spaces and often offered their own software to download. If we were to use Misskey, for example, we could offer storage space since Misskey has a storage function. It would involve developing a synchronization client, similar to Dropbox. We could charge, for instance, for the creation of federated groups. But in the end, if we exclude advertising and virality as economic models, the only option that remains is payment for the service or a premium service.
So, if you want to envision a dominant Fediverse in terms of users, what you need to consider is when a private ecosystem will emerge. And the situation is dire: there is no federated shop that can compete with Amazon. This lack is significant, especially when discussing the replacement of giants, including Amazon. A federated shop would, in a sense, have significant advantages over Amazon, but on the other hand, there is currently no software infrastructure for the Fediverse. There is nothing connected to a specific device, such as something linked to a particular brand of smartphones, and there is nothing tied to almost no commercial entity. Likewise, I see job requests and sometimes offers circulating, but there is no federated marketplace for employment.
To conclude, I see the Fediverse as incomplete. It is still in the idealistic phase where everything can be done through volunteers and donations, but it completely lacks (even though it would be possible by writing the code) everything that could transform it into an ecosystem without solely relying on advertising. Sure, we have the equivalents of Facebook, Twitter, YouTube, and even podcasting systems that can be federated, as well as blogging platforms.
But as of today, we cannot use it to find a ride, a rental room, or have a marketplace. We don’t have newspapers, party instances, or instances for professionals.
However, the idea of paid (or business related) instances MUST enter people’s minds, even among the idealists.
