The failure of Mastodon to break into the mainstream especially in the wake of Twitter/X's debacles is a good example on how decentralized web is a good idea on paper but won't work in practice. It turns out that a) people will just join the largest instance/community possible for network effects b) internet discourse is bifuricated enough that preemtively banning instances just in case makes discourse even more bifuricated and c) maintaining and moderating said instances is expensive and time-consuming.

Even though Bluesky did break through to the mainstream with ATProto, it's unclear how ATProto is either a functional or competitive benefit, or if its users even know about it.

Mastodon might not have broken into the mainstream, but it's surprisingly viable. I am totally OK with its current state to be honest - feels like the cozy internet of 2000s.

This is, at least in part, a UX problem. Even as someone with a lot of technical experience I found Mastodon quite disorienting at first. Bluesky has solved this much better which is why they've won out over Mastodon.

I agree completely. This isn't the world of BBSes where everyone had a corner or two that they hung out on. As long as posts are gamified with likes and notifications, decentralization will never win. Even HN isn't immune and it has carved out a pretty specific niche of text based and no ads social network focused largely on tech. I don't think another decen-HN is likely to take off or break into other self organizing segments even though this is easy to attempt with some Mastodon servers.

Those sound like reasons why federation doesn't work. Some of us don't think federation is decentralized enough anyway.

I asked people who use platforms like X and Substack (Notes) why they don't use Mastodon and they said they can't discover new people and new content.

I'll post this every time mastodon comes up: to compare mastodon to Twitter, Facebook, bsky, etc. is fallacious in the extreme. Saying that mastodon is "losing" to bluesky is a complete non-sequitur.

Mastodon hasn't lost to these other services because they aren't competing. Nobody on mastodon actually wants to replace Twitter. Mastodon exists explicitly as a rejection of mainstream social media. Calling this "losing" is like saying that a bunch of kids playing basketball in the street are "losing" to pro hockey teams. Not the same game, not the same scale, the only similarity is that they're both playing a sport.

Mastodon doesn't want to compete. That's totally counter to the entire philosophy. Mastodon isn't even remotely playing the same game as twitter and Facebook. It just so happens that mastodon has a text box with limited characters. That's it, that's the only similarity.

Of course a chess club of 10 year olds in Japan isn't going to "win" and displace the entire American baseball league. You'd be laughed down for even suggesting that. And yet, people think that mastodon is "losing".

Mastodon is exactly where it wants to be. The small scale and diffusion are features, not bugs.

Besides that, it's functionally impossible for "mastodon" to go mainstream. To suggest that is a fundamental misunderstanding of what mastodon is, how it works, and how the people involved run it. At absolute most, you could get a few big servers to go 'mainstream'.

When a server gets too big or too popular, the network treats it as damage and routes around. A lot of admins do this purely on principle: huge servers are antithetical to the philosophy and actively bad for the network. The rest block them because the vast majority of attacks and spam originate from such servers.

If mastodon went mainstream, the network would fracture (again) and the majority of nodes would re-form the network in the obscurity they enjoy and value.

This is all extremely intentional and deliberate. Mastodon is an explicit rejection of all the things you think it's "losing" at. Mastodon doesn't want to "win", they largely want to be left alone to continue enjoying a small internet with limited connections.

Mastodon has won its own game. It's firmly established itself in a niche, and is not going to move from that niche to chase fads, engagement, metrics, profit. Mastodon wants to be exactly what it is because mainstream social media is a very bad thing

Discord servers are half of a good counterexample, since people try to avoid the extreme end of network effects and want their chat partners pre-screened.

But... it's all on Discord, not IRC or jabber or anything.

Now that you mention it, I vagely remember something about Bluesky being built on some distributed protocol. So yeah, I guess it's correct that doesn't mean much to it's popularity. Which is sad of course. But yeah, most people don't know about technology, and security and privacy, so yeah, why would they care..

I think there's an element of how decentralized services work that's not sufficiently refined yet. In a lot of ways it's a better way to have a good user experience with decentralized data that we're sort of looking for, and in some ways, it doesn't really fit how the web works.

All of these decentralized social media concepts were/are too technically complicated for too long during the Twitter debacle. Even now, the top results for "how to use Mastodon" have jargon incompatible with mainstream usage.

If decentralized is great on paper and not in real life, then the paper is wrong and thus not great. Additionally, if decentralized doesn’t work, please explain the survival of email.

d) The kind of people who put a little thought into how Twitter could be done differently may come to the conclusion that they don't actually need a Twitter in their lives

Mastodon would be OK but the UI/UX is just terrible... interacting with someon on different instance is utterly cumbersome and I do/could miss a lot of posting in global view if someone didn't already followed someone from other instance...

The entire point and purpose of mastodon is to not be mainstream social media. Almost everyone there signed up because mainstream social media is so awful. The small and diffuse nature is one of the main benefits. Because it's distributed, there's an extremely niche community server for any subject you might be interested in.

Mastodon captures perfectly what twitter was at the beginning. Just people talking to each other and having fun. Not performative like fishing or whatever. No ads, algorithms, endless feeds, AI slop, or spying and tracking.

It turns out that a lot of people think mainstream social media is objectively abhorrent and want to connect with other humans in a more natural and user-driven way.

Frankly, "mainstream" is a dirty, disgusting concept here. A very large fraction of users would put a significant amout of effort to prevent that from happening. If one server did become "mainstream", as mastodon dot social did, the network treats it as a damage and routes around. Many servers just cut them off because the vast majority of attacks and spam originate from the big public servers.

Mastodon won't go mainstream because they don't want to and the system fundamentally cannot operate that way. A few individual servers may go mainstream, but we'd eventually consider it a hard fork. The network would fracture (as it has many times) and the network of small servers will go back to the obscurity we enjoy and cherish.

Case in point: Facebook tried to force their way into the fediverse network through threads or whatever it was called. There was a pretty hard split in the network as nodes that value privacy and safety cut themselves off from a literal hostile invader. Many servers went recursively through the network to cut off any servers who hadn't blocked Facebook. I haven't heard anything about Facebook trying activitypub again, so I guess the quarantine was effective.

> decentralized web is a good idea on paper but won't work in practic

This claim is so vague it's meaningless. I could just as well argue that torrents are the purest example of decentralisation, and work amazing.

That’s a key feature