The biggest surprise for me was the https://hexbear.net count, an instance I hardly interact with.
| Community Count | Community Subscriber Count | |
|---|---|---|
| beehaw.org | 6 | 133450 |
| hexbear.net | 33 | 663204 |
| lemdro.id | 1 | 17052 |
| lemmy.blahaj.zone | 1 | 15907 |
| lemmy.dbzer0.com | 1 | 53006 |
| lemmy.ml | 14 | 356460 |
| lemmy.one | 1 | 16257 |
| lemmy.world | 39 | 851950 |
| lemmynsfw.com | 2 | 33586 |
| sh.itjust.works | 1 | 16006 |
| sopuli.xyz | 1 | 14093 |
The data this is based on comes from https://lemmyverse.net where you can just download a full json of the data they have (I excluded all communities marked as "suspicious")
EDIT: The data if you sort by active users last month:
| Community Count | Community Active Month Count | |
|---|---|---|
| awful.systems | 1 | 2616 |
| feddit.org | 2 | 7363 |
| feddit.uk | 2 | 5289 |
| hexbear.net | 1 | 2952 |
| lemdro.id | 1 | 2898 |
| lemm.ee | 3 | 8898 |
| lemmy.blahaj.zone | 1 | 11422 |
| lemmy.ca | 3 | 14910 |
| lemmy.dbzer0.com | 3 | 13752 |
| lemmy.ml | 10 | 54949 |
| lemmy.world | 57 | 338384 |
| lemmy.wtf | 1 | 3602 |
| lemmy.zip | 3 | 12020 |
| mander.xyz | 1 | 11469 |
| sh.itjust.works | 5 | 37365 |
| slrpnk.net | 3 | 10897 |
| sopuli.xyz | 2 | 10070 |
| ttrpg.network | 1 | 4107 |
Community Count:
Community Users:
Lemmy.world has no lock in on their "power". They have the most volunteer labor, money, and infrastructure. That's makes them stable, so people aren't worried about their data suddenly going offline (like kbin) and they don't worry about the service being flaky.
The same can be said about gmail and it is the same kind of problem here. Yes lemmy.world is not a profit orient it giant, but it is still a problem when one actor has this power over a federated network. (the scale of the problem is of course a lot larger with gmail)
Technical issues with Lemmy are, I think, still driving people to larger instances.
The big one is that if I make a community on a smaller instance, and gain ANY amount of volume and traction (which is not all that easy to do in the first place) and that server vanishes, shit's just... dead. It's gone and not coming back, because you can't move a community from a dead server to a live server.
Which means using one of the big, established, funded, stable, working instances is the only rational choice, but that also means I'll probably just make an account and post exclusively from there, and thus you end up in this cycle of everyone just going to one of the larger instances in preference to any of the smaller ones.
Everyone goes on and on and on about account portability being very important (which, I suppose it is: I don't think we need account portability but rather distributed identity independent of the specific platform you're using, but that's a whole different technical mess) but for something like Lemmy, being assured that the community you're working on will survive servers vanishing and a means to "take ownership" in a way that lets you port it to another home if and when your instance dies - because, for the most part, it's going to at some point - is far far more needed.
The size difference between Lemmy.world and lemm.ee could still be improved
It's that, plus the next largest instance being practically unusable due to hyper aggressive tankie censorship. Getting banned from .ml for not sucking Stalin's boot hard enough is practically a rite of passage at this point.
Is it good or bad that I had to think if you meant hexbear or lemmy.ml, and even after doing so I'm still not sure?
I agree in principle that .world containing most of the fediverse’s activity kinda isn’t great for the idea of the democratic nature of the fediverse. However, the point of the ‘verse is that anyone can spool up an instance if they dislike it, or start more communities on existing instances. If .world were to disappear it would suck, but that’s part of the problem with any instance in an informal community. Any of them can disappear.
How many instances federate by default? It may be difficult to get your new solo instance into the others.
That's just how federation works out in every federated service ever.
While spreading out is good, this isn't something like cryptocurrency where it's specifically bad if you have over 50% share. Each instance is the source of truth for their users and communities hosted there. It's not like a block chain where something with over half can suddenly define their own truth for everyone. So it's not necessarily a massive cause for alarm.