Most of these relate back to previous posts I’ve made, but I’ve found them to be interesting/complementary to them – or I want to talk about them in a way that’s too long to go in a thread somewhere but too short to be in an actual post.

“Mastodon is easy and fun, except when it isn’t”

Erin Kissane did some informal research on why users bounced off Mastodon, using a sample of BlueSky users who were willing to answer the question as to why they left.

Most of these reasons are quite congruent with what I suggested in my post on the failed “Mastodon Migration”; and the segments about poor discoverability and a culture that focuses on hostility to outsiders and lecturing ring true to me. She also posits some answers as to how these things can be fixed – how likely these are to ever be implemented is however something I would be dubious of, since (as noted in my piece) the users and main developers of Mastodon presently are those who like it already and don’t want it to change.

I did quite enjoy this quote from a bounced-off Mastodon user, which rather sums up how I feel about how poorly decentralisation actually works for users and how badly Mastodon’s functionality is oversold:

I was told picking a server didn’t matter. Then it turned out it actually mattered a great deal for discoverability. Then I’m told ‘migrating is easy’, which is just a straight up lie.

I do have to wonder how people keep talking about migration being easy and painless when you lose all your past posts and media in the process. The usual reply to this seems to be “why would you want to look at all your old posts?”; a typically condescending tone-deaf “works for me” response. What if I do want to, dickhead?

“The ecosystem is moving”

That leads me neatly into this great post by Signal co-founder Moxie Marlinspike, which goes over the considerations of why a federated protocol is potentially of less utility than a centralised platform – using the example of XMPP. His suggestion is that a protocol becomes stuck in a moment in time because of the need to maintain consensus between its users, and this cripples its ability to respond to new user demands and needs.

I discussed this phenomenon briefly in reference to XMPP in my piece on Meta using ActivityPub, and agree. As I say there, XMPP did not fail to gain any consumer foothold because of nefarious ne’erdowelling on the part of Google, but because there was no real reason to use it – and even less after competing products outclassed it in terms of features. XMPP is a great instant messaging platform, if you like talking to absolutely nobody using a feature set that is somewhere behind AIM circa 2008.

The same also applies, although Marlinspike doesn’t directly say it, to ActivityPub, the protocol principally used by Mastodon. ActivityPub now underpins a whole variety of different services, only a few of which it is actively suited to, and so its feature set is broadly limited to a subset of what Twitter offered circa 2012. Things like quote posts are not part of the spec, so implementation of them is restricted to clients, where they are essentially unofficial extensions of the protocol.

Meanwhile, Meta’s Threads (which has faltered a bit after an initial strong start – albeit landing a hell of a lot ahead of the Fediverse in terms of active users) has iterated on itself quite quickly, albeit maybe not quickly enough to recover its initial momentum.

The other side of this is that while Marlinspike talks about how a lot of protocols ossified in the late 90s, a lot of fediverse users and developers seem to actively want to go back to what they see as a halcyon era of the Internet when it was all protocols – probably, quite literally, the late 90s. And again, they’re the ones in charge. That feels dangerous when the supposed driving idea is to try and revolutionise the social media of 2023.

The Internet of the late 90s and early 2000s sucked, incidentally. Yes, Meta might put tracking cookies on your computer, but I’ve also not needed to run Spybot Search and Destroy for the best part of two decades, and in general today’s Internet is far more useful and simply has far more fun in it for the average person. Nostalgia for those times is very misplaced indeed.

“Meta just proved people hate chronological feeds”

An interesting article from Wired on Meta publishing some research showing that people tend to get bored of social media without algorithms very quickly. As a short summary, users start to scroll more at first to try and find what they might find interesting, then discover that it’s more trouble than it’s worth, and then they go somewhere else to get the fun they were looking for.

This is something that could have gone into my Mastodon piece, had it, you know, existed when I wrote it. Mastodon offers nothing but chronological feeds; but given its famously terrible content discovery and its lower levels of activity overall, this just annoys users more, since they have a fairly large pool of activity to sift through but no means of easily surfacing things that they might be interested in.

And, of course, its users resist even the availability of content-surfacing algorithms to other people, even if they aren’t made to use them themselves – sometimes on such tenuous grounds as “I want control over how my content appears” (something which is laughably naïve in the context of how little meaningful control ActivityPub actually gives you over your content).

For all people complain about Internet users just wanting “dopamine hits”, at the end of the day, people get dopamine hits when they enjoy things. It’s a very strange attitude, peculiar to Mastodon, that you shouldn’t easily enjoy the process of using social media – that you should have to work for a long time to try and find things you might be interested in, and if you don’t then you’re somehow doing it wrong. The problem is that Mastodon does not exist in a vacuum, and this “eat your greens” approach to the Internet only really appeals to, well, Mastodon diehards. If people don’t like it, or want algorithms, then there are other places they can go that have them. Meta’s research proved that that’s exactly what they do.

And realistically, people use social media for leisure. For fun, over limited periods of their days. “Making fun hard work to obtain” feels like a pretty shitty goal to have, in general. Not to mention counterproductive if you want people to actually use whatever it is you’re offering. When I go to Tesco to grab a meal deal, they don’t tell me to fuck off and make my own sandwiches in case I get addicted to the dopamine hit of getting food easily; and if they did, I’d probably go somewhere else for lunch.