One question I've had for fediverse people is how you prevent a federated system from centralizing. I am legitimately curious. Email is often given as an example, but imo that a perfect example of a decentralized system BECOMING centralized. Sure, other players exist but the vast majority of people are on gmail, apple, or outlook (which is much smaller than the other two). Things tend to follow power distributions due to the momentum force being critical. In network systems (e.g. twitter,facebook,HN,email,ISPs,Walmark,etc) the utility/value is not linearly proportional to the userbase, but super-linear (this was one of the big problems with cryptocurrencies too. "Gotta have money to make money"). In these systems resources are "attractive."
So with this in mind, how is a decentralized paradigm any different than an attempt to just reshuffle the top players? (i.e. re-centralize but with a different group at the top) I just don't see the mechanism that prevents centralization.
It's certainly a challenge. Mastodon's development tends to prioritize mastodon.social (Eugen Rochko is BFDL of the software platform and also runs mastodon.social) -- for example, the mobile app now signs people up by default on mastodon.social, and functionality that people running smaller instances have implemented in forks hasn't been integrated back into the main line. So there's the weird dynamic that people generally have better experiences on small instances (as long as they're well-admined) but the vast majority of the current fediverse is on large Mastodon instances. So it'll be interesting to see what happens in response to Meta. There's likely to be a partition, and if .social winds up taking a Meta-friendly position, then the anti-Meta region may be much less centralized.
https://heat-shield.space/mastodon_two_camps.html looks at tensions between people who just want a "better twitter" (which tends to lead to centralization) and people who focus more on small communities (a more decentralized solution).
I think you don’t, which is why I think federated systems built on the existing stack are doomed to fail (by recentralizing). You’re right to point out email as an example of this failure.
You need urbit or something like it to fix it, the problems are deeper.
I am continually impressed by how much more relevant Urbit has become (contrary to people years ago arguing it would fade out into irrelevancy). Urbit was designed holistically to solve the problems the "web 2" internet experienced in a structural way, not just apply some fancy "web 3" bandaids to some them. So of course it is a lofty project. But time and again it's proven that it took the right stance on socio-technological issues. Will it ever gain enough traction so that it replaces your text message app? Well that's really a social question, and one can wish. But it certainly solves all the problems people keep bumbling into when trying to "do web3". I wish more people would give it a serious look.
I work at Tlon (the main startup behind urbit, so disclaimer) - it's a lot easier to use urbit now than it used to be thanks to free hosting.
It needs to be a lot easier still (particularly the mobile experience isn't there yet without a fully formed app), but if it's been a while it's worth checking out again: https://tlon.io/
It'll be insanely hard to actually pull off, but it's the only attempt in this space that I think has a legitimate chance of a successful outcome. The others are dead on arrival because they don't actually fix the underlying issues. (Success being widespread adoption of software the users actually own and control.)
> the vast majority of people are on gmail, apple, or outlook
What do apple email addresses look like? I genuinely don't know anybody that uses an address that says to me "provided by Apple"...
I do know a lot of people that use their corporate/organisational addresses for personal email though, which always surprised me.
Edit: just read that "me.com" email addresses are apple-provided - I have at least seen them used occasionally, though nowhere near as much as gmail and hotmail/live.com (outlook).
I couldn't recall ever noticing anyone use an icloud.com address as their primary personal email address, though searching through my inbox it has cropped up once or twice. Actually my sister did use one briefly in 2018.
Are there any stats on how many people do that? It seems me the only way it could be true that Apple mail has far more users than outlook/live/hotmail is if the former tend to be associated with custom domains.
just to muddy the waters a bit, I recently moved my personal domain to use iCloud as the mail provider, so you wouldn't see @me.com/@iCloud.com, but just my domain.
I've seen email being centralized mentioned a few times but I don't think it's true. I email with plenty of people who don't use gsuite... seems to work fine. I can see a power distribution, sure, but that's still a distribution not a central organized cabal that stops newcomers.
Still I agree -- decentralized paradigms that are successful seem like they'd end up with big players like you describe eventually. Still better than actually centralized like slack or something.
The reason there's this confusion between us is that we're using the term monopoly (and consequently centralization) to refer to different things. A lot of people don't use the word to refer to the pure sense but rather the practical.
A pure monopoly (a single story) is rather uncommon and usually associated with government direction. Generally these don't exist without government intervention because if anyone is able to create a company, in any form, and sell the same product/service (in any form/price) then it's technically not pure. So doesn't happen unless it's illegal or some other factor. Whereas an effective monopoly refers to dominance in market share.
The reason a practical monopoly is used is because we need to look at the reason to why we discuss monopolies in the first place. Usually it is because they are bad (but they aren't always). The reason they can be bad is because abuse of power: they are able to dictate the market and force prices that are unreasonable.
So with this understanding, we usually call things like Coke+Pepsi a duopoly despite a combined market share of 68% (48% + 20.5%). Or we talk about oil despite OPEC not even being 40% (followed by Persian gulf, oapec, then US), and all of these are multiple players. Similarly we say the same about ISPs. The problem with these isn't that it's single player control, but rather significant market share and the leverage that comes with that, which includes the ability to reduce competition and effectively cornering the market.
I want to give this explanation so we can discuss on the same page and understand what each other means. Because when I'm referring to email (context clues should tell you I don't mean pure -- I mentioned multiple players) I'm referring to these companies abilities to force aspects onto others. Maybe a more clear example is browsers because Chrome's pressure is frequently discussed here and Explorer's previous history also shows similar influence.
I'm not saying decentralized services will become 100% centralized by a single player (it would be incredibly naïve to interpret my words to mean that), but that people will collect into a small set of servers (likely seeing a power distribution) who then have a clear ability to use their leverage and dictate the format of smaller servers. Effectively this is not too different than the power that Salesforce has over Slack. While not absolute power, it is closer to that end of the spectrum than a fully decentralized system. If you make the mistake of creating two bins -- centralized vs decentralized -- you are missing the entire point of decentralized systems. At least the point that is philosophically argued.
Try to realize that when people are talking they have different priors than you. Language is messy and communication is about the exchange of ideas. So even if you disagree with my definitions (neither of ours are objective) you can still argue my points. Just claiming semantics is problematic because it derails the conversation or creates artificial contention. See 5.11 of the Simple Sabotage Field Manual.
The fight over whether or not to federate with threads might actually cause more decentralization, as people move off instances if they don't like the policies the admins have chosen, and different admins make different choices.
Even "effectively re-centralized federated" will still provide much better ways to keep all that browser fingerprinting stuff and the like away from Meta et al than a closed system. E.g. if for some reason you need to operate a Google email address you could run the day to day read and write behind a forwarding setup never connecting your regular browser or your imap client to Google servers, dealing with the occasional setup/maintenance from a browser properly isolated from your search history and the like.
It's not just that there is a risk of re-centralization from network effects.
Since you mentioned email, there is also a large risk that filtering unwanted content and impersonators from appearing on your instance becomes so expensive and/or time-consuming that only a few instances have the resources to do it, just like happened with email.
This is similar to how urbit’s ID system works fwiw - the ID is tied to the computing node itself 1:1 and you can’t have an account that’s separate from a node.
What's the problem with most people being to Gmail? It doesn't prevent you or me from being on a different server and communicating with them. On the contrary, e-mail is a great example of how a decentralized protocol can be widely adopted.
Let me rephrase. I don’t think the current iteration of Mastodon can scale well enough no matter what hardware you throw at it. It’s not designed for it. You would need to build a new system to achieve what was proposed.
I suspect that the above along with the cost is why it hasn’t happened. Although you do miss a lot of what’s going on by not being on one of the larger instances.
One question I've had for fediverse people is how you prevent a federated system from centralizing. I am legitimately curious. Email is often given as an example, but imo that a perfect example of a decentralized system BECOMING centralized. Sure, other players exist but the vast majority of people are on gmail, apple, or outlook (which is much smaller than the other two). Things tend to follow power distributions due to the momentum force being critical. In network systems (e.g. twitter,facebook,HN,email,ISPs,Walmark,etc) the utility/value is not linearly proportional to the userbase, but super-linear (this was one of the big problems with cryptocurrencies too. "Gotta have money to make money"). In these systems resources are "attractive."
So with this in mind, how is a decentralized paradigm any different than an attempt to just reshuffle the top players? (i.e. re-centralize but with a different group at the top) I just don't see the mechanism that prevents centralization.