I don’t like the clickbait title at all – Mastodon’s clearly going to survive, at least for the forseeable future, and it wouldn’t surprise me if it outlives Xitter.
Still, Mastodon is struggling; most of the people who checkd it out in the November 2022 surge (or the smaller June 2023 surge) didn’t stick around, and numbers have been steadily declining for the last year. The author makes some good points, and some of the comments are excellent.
Re blue sky, is anyone actually federating with it? I don’t know of any other instances besides the official one.
Yeah, right now the way I think of it is that Bluesky is (conceptualy) a single big instance, connected to the rest of the ActivityPub fediverse via Bridgy Fed (which speaks both AT and ActivityPub). Bluesky’s decentralized in a different way, and the broader ATmosphere (apps that use AT protocol) is growing as well, but it deosn’t really have the same concept of instance.
Yes but no. Due to architectural differences, federation under AuthTransfer protocol is simply different compared to ActivityPub. In its own terms it is federated as individuals’ data is stored in personal data servers (PDSs) connected to a relay, which currently is only the Bluesky relay, that roughly speaking connects them to other personal data servers.
You can technically operate your own personal data server apart from those operated by Bluesky, but I think it’s fair to say the vast majority on there don’t. It’s not clear yet, apart from fully holding your own data, how useful it is to operate your own given you only have one relay to use anyway at the moment.
So even in its own terms Bluesky really isn’t federated in much of a meaningful sense yet. The problems are twofold: a major part of their pitch is making federation Just Work™, keeping the underlying tech out of mind to mitigate confusion, but you can’t have your cake and eat it too here. Eventually, if you’re really committed to meaningful federation, you have to teach people about the value of operating their own personal data servers, at minimum, otherwise what was the point in separating it out in the architecture?
Problem is, that goes against their pitch to their audience and spoils the appeal. It’s telling a good joke only to kill it by explaining to the one person that went, “I don’t get it.”
Secondly, they’ve already upfront said that relays may be cost prohibitive for many people to operate, resulting in only a few ever being spun up. If that remains the case and is true, then even if a few were spun up, that’s not any more federated or distributed than the rather consolidated web we see now. How much of a difference would it make if the social web was running on AuthTransfer and the major relays were owned and run by Meta/Facebook, Twitter/X, and Google?
Congrats you have your own data in a personal data server…But are you really the one running it, or did you just opt into the PDS entryway offered by Facebook/Twitter/Google/etc. because sorry, what’s that about a server?
@ColonelThirtyTwo Sorta/mostly. The protocol has a bit of a different model then Activity Pub, and it’s in development so there are some limitations, but it’s been opened and there’s people hosting their own PDSs now (the part of Bluesky that hosts your account).
To my knowledge there’s only two AT relays (the part that aggregates content from PDSs), Bluesky itself and very recently frontpage (a link aggregator). That makes the network fairly centralized right now, although BlueSky/AT has made a lot of progress in the last 9 months in terms of opening up so I expect it’ll be a lot less centralized this time next year. I’m also betting that somebody will make an AT client that pulls posts directly from PDSs instead of going through a relay at some point.