“A nearly impenetrable thicket of geekitude…”

Tweets and Toots

Posted on April 27, 2022 at 17:05

On Twitter

By some metrics, Elon Musk is the richest person alive. Admittedly, after his deal to purchase Twitter for an incomprehensible amount of money was announced, he got quite a lot poorer on paper because a lot of his wealth is dependent on the excruciatingly high P/E ratio of Tesla. That valuation has dropped by 23% in the last couple of weeks, 12% of that on the announcement.

It must be nice to be the kind of person who can find $44bn down the back of the sofa, but I’m not going to dwell on that. Instead, my interest in this story is centred around how it will affect me as a Twitter user.

I’ve had a Twitter account for 14 years. In that time, I have made a grand total of 77 tweets. I follow 134 accounts, and am somewhat miraculously but also inexplicably graced by 75 followers. I follow accounts for friends, colleagues, software I use, authors I like, journalists, law professors, and Foxes in Love. Most of my mentions are crypto-currency schemes of some kind, or typos. Mainly typos, I think, but it can be hard to tell.

In other words, I’m not a super engaged member of the Twitter community (I’ll argue below that “Twitter community” is in any case an oxymoron). I’m mainly using the service as an aggregator rather than as somewhere to go for a good argument, despite the latter seeming to be the service’s USP. The outrage and shiptoasting that is endemic to the platform only really affects me when it outrages one of the relatively small number of people I follow, and has in any case reduced substantially in both quantity and ferocity since the banishment last year of a certain prominent political figure. I still find the atmosphere intimidating.

The problem for me is that Musk’s ideas about how to fix what he perceives as Twitter’s problems, if implemented, are likely to revert the experience to at best that of the worst point of the last few years. Claiming to be a “free speech absolutist”, he seems to believe that the platform will be improved if anything that is legal speech is permitted, with no moderation outside that boundary. Improved for him personally, perhaps, but likely not for me.

My opinions align fairly closely with those of John Scalzi on this: even if that happens, Twitter will still be useful enough in its likely degraded form that I will still use it. It might be the start of Twitter’s ultimate decline, but in the same vein as “Markets can stay irrational longer than you can stay solvent” (looking at you, Tesla) I won’t be holding my breath.

It does, however, make sense to be at least investigating alternatives, or additional services in the same space, just in case. Of course we’ve been down this road before; I think an important insight is that all communications systems (except IRC, some would snark) should be regarded as ephemeral and subject to falling into disuse and eventually obscurity over time.

On Mastodon

My record on predicting social media trends and selecting new platforms is not good, but I have been looking at one platform in particular which seems to be getting some mindshare as a possible alternative to Twitter. That platform is Mastodon, which has been around and slowly gathering adherents since 2016.

I said earlier that I thought the phrase “Twitter community” was an oxymoron. To expand: I don’t think that there is such a thing as “the Twitter community”; that term doesn’t make sense for me in the context of a service with 200 million active daily users. It’s clear, for example, that the maximally broad user base participating in the Twitter “community” is at the root of many of its problems: there are no attitudes or interests in common across such a large fraction of living humanity. Moderation becomes a near-intractable problem as a direct result of this: the people who want a forum for rational discussion will never see eye-to-eye with the people whose idea of a good time is to actively offend and outrage. That’s particularly true when the two groups disagree so vehemently about who is taking each role. For the record, I don’t think Musk’s idea of reducing the amount of moderation is a good answer to that failure to agree.

At a technical level, Mastodon splits the user base across many instances, each of which is intended to represent a (real) community of interest. Such instances can have their own community standards as to the speech allowed, and you’d expect that the difficulty of moderation to such standards to drop from intractable to merely exhausting as a result.

This allows a community of users to operate within rules of their own devising; as you might predict, this means that Mastodon instances exist catering to all the varieties of speech you would expect to be moderated on universal platforms like Twitter.

If what we wanted was a community around a single area of interest, Mastodon would behave a bit like a small to medium size forum instance. Obviously those already exist; instances are however visible to each other as part of a global system referred to as the Fediverse; in principle, any user on any Mastodon instance can see and interact with users on any other instance.

As well as being under no obligation to permit unconstrained speech, it’s also possible for each instance’s community to determine whether its own users may access speech from all other instances or to restrict some. Those choices depend on the instance’s local community and not some globally imposed norm.

As an example, the mastodon.social instance’s rules include clauses such as “no incitement of violence or promotion of violent ideologies” and “no content illegal in Germany”. The instance makes content from specific other instances unavailable under rationales such as “misinformation”, “hate speech” and “transphobia”. I personally think those are a good set of rules, so I’m happy to agree to them and set up an account there. If you found some element of the rules problematic for any reason, at least in theory you could set up an account on another instance: after all, the claim is that there are several thousand.

So far, so utopian, right? Maybe it’s time I started talking about issues.

Platform Norms

One place to start is my statement above that rules are local to an instance, and “not some globally imposed norm”. I wasn’t completely lying there, but there are pressures for instances to conform to a more broadly accepted set of rules. I happen to like that ruleset and that pressure, but its existence isn’t obvious at first sight and I think that’s notable in itself.

I’ve already mentioned one of those pressures implicitly in saying that mastodon.social bars content from other specific instances on the basis, for example, of “hate speech”. This obviously limits the reach of users on instances assessed to be in that category; if you want your speech to be able to reach the widest audience then you’re under pressure not to make your account on an instance with this kind of reputation, whether your own speech would be acceptable elsewhere or not.

The second normative pressure is around discovery of the community you’re going to call your home. If you use the widely advertised joinmastodon.org site to begin your journey, only certain instances will show up: at the time of writing, a total of 125 out of the 3000 or so that exist. To be listed, instances must commit to the Mastodon Server Covenant which (amongst other things, like taking backups and giving reasonable notice of shutdown) mandates active moderation of certain broad classes of presumptively harmful content.

I think both the covenant and the specific content rules implemented by the various instances I have looked at are desirable for me, and the covenant in particular seems like a good baseline for healthy discourse. I like these rules and the way that they intentionally exclude certain behaviour (and more indirectly, certain people). It would be dishonest, though, to claim that they don’t represent a significant platform-wide norm that may affect the way the platform grows.

Servers and Sysadmins

Twitter is “free” in that there is no charge for most users to participate in the service. The resources to run the service are paid for, presumably, out of advertising revenue derived through the usual surveillance ecosystem.

Each Mastodon instance’s resources need to be paid for, somehow, by the community the instance serves. In some cases, that’s going to be a single person who is passionate about running a service for that community offering up their credit card to a cloud hosting service. In others, a crowdsourcing system such as Patreon might be used. So far at least, I haven’t seen an instance trying to fund itself through advertising.

The money has to come from somewhere, though, and used to pay for an internet-exposed server which also requires a system administrator: not everyone who is passionate about a community has those skills even though these days it might “just” be a question of spinning up a Docker container. I don’t think there is any reason to believe that any given small community (present company of course excepted) will always contain a member with both the skills and inclination to take this on.

The other major cost, though, is not necessarily financial: an instance which has rules by which content will be moderated needs to have a way of developing and then implementing that scheme of moderation. Anyone who has been involved in trying to herd Internet cats will tell you that this is tricky even with a small number of people and gets very hard very fast as the numbers grow. You don’t have to get to the scale of a Twitter or a Facebook for it to be a huge time-sink for the lucky community coordinator.

The advantage of a small Mastodon instance is that its community can develop its own rules of the road. The amount of cost and effort per user is probably much higher than for a larger instance with more users, however.

The advantage of a larger instance is that fixed costs of all kinds spread across more users are lower per user… at least up to the point where moderation costs — which are definitely anything but fixed — start climbing exponentially.

Is Community What People Want?

I think the ideal of Mastodon — interlinked community-based instances with their own rules, with a set of broader norms against harmful content — is pretty attractive. It’s certainly worth a try for those whose use of Twitter is oriented around a specific community which happens to be represented on Mastodon already (or for which they are prepare to stand up an instance).

It’s less clear to me that the ideal represents anything close to how people think of Twitter today. To repeat:

I follow accounts for friends, colleagues, software I use, authors I like, journalists, law professors, and Foxes in Love.

Which community do you think matches this list? Or, put another way: which Mastodon instance should I set up a Patreon subscription to help pay for the running of, in order to participate in the broader Fediverse?

I think it’s telling that I don’t have a good answer to that question. The people I follow on Twitter seem to be spread out across Mastodon instances, if they are present at all. None of those instances seem to represent a community as such. As a pragmatic first step, I’ve done what I suspect many others will do and just pick the “flagship” mastodon.social instance. This is run by the German non-profit behind the software and is funded by donations through Patreon.

I’ve set up a bottom-tier Patreon contribution which I will maintain as long as this is interesting to me (or until I move to an instance that is a better fit) but my suspicion is that most people won’t bother. At the time of writing, the mastodon.social instance has around 680k users and Patreon donations averaging less than £0.01 per user per month. Long term, that’s going to turn into a scaling problem if it isn’t one already. I’d be surprised if moderation at even the current scale wasn’t a significant issue, for example.

If a significant proportion of users don’t identify primarily as a member of a specific community, and instead opt for a generic instance, a lot of the potential Mastodon has for changing the nature of discourse for the better is going to be lost.


It may take a while for Twitter to get visibly less useful for me than it is today, but it seems inevitable in the long term. That’s regardless of the acquisition by Musk, but he’s more likely in my view to accelerate that process than to slow it.

Mastodon and the Fediverse look like they could be the basis for more humane social networking, because they are (at at least in principle) structured around self-identified social groups rather than “maximum ‘engagement’ and ignore the consequences” algorithms. I’m far from convinced that’s what most people want, though.

Some Notes

This article has fewer links than I’d usually provide. That’s intentional; there are definitely things in this area I have no desire to feed traffic to. Use a search engine.

I’m not linking to mastodon.social because the site explicitly requests that people link to joinmastodon.org instead. There, you’re at least exposed to a variety of possible instances before choosing one. You should definitely start at joinmastodon.org if you’re at all interested in Mastodon.

Do I really have to explain the article title? Twitter calls individual posts “tweets”; on Mastodon, they are “toots”. Because elephants? Trumpeting?

I am @iay@mastodon.social.