This post writes up some threads I wrote around the time of the ‘Great Migration’ in early November, after Musk’s takeover of Twitter prompted blind panic. It explains how I think even if Twitter dies, the services it confers will live on.
I’m finding it tricky to get good statistics on social media membership and use, so take the figures I cite with a pinch of salt, and give me better ones if you have them.
The Migration, which referred to the new accounts created on Mastodon, seems to have petered out, as most users have got used to Musk’s new era, and the pace of policy change/policy reversals has slowed. The Verge report that active Mastodon users [undefined] jumped from 300k to 2.5million between October and November 2022 after Musk confirmed that his purchase had beeen expedited. [There was a much smaller jump in April when the offer was first accepted]. The latest toot [Mastodon for ‘tweet’] from the bot ‘MastodonUserCount’ shows total users [as distinct from active users] rising steadly to just over 9 million this week. Another user [link when I can find it again] has data quantifying ‘active users’ [undefined] and here the picture shows that this has been falling over the last month by a few hundred thousand. Those who migrated have eyeballs that did not stay [noticing, like me, that others are not using it to write and interact] even if they left a legacy or a lifeboat of an account.
A much smaller migration [of the order of 100k, including a waiting list of about 300k - I have not found good figures for this so let me know if you have] took place to Post.News, a beta of a website founded by ex Waze CEO Noam Bardin, with investment from Andreasson-Horrowitz. Post.News has a slicker user interface currently than Mastodon '[although some independent Mastodon clients are being developed that are getting close]; and it has functionality to implement paid content including tipping/micropayments.
The migration nevertheless still leaves the Twitter network mostly intact, with about 330m active users according to this report. [Other figures exist with different definitions for what qualifies as an active user].
There are many problems with the Musk ownership of Twitter. Musk has brought far-right, populist and anti-democratic values to content moderation. This has antagonized advertisers that don’t want to be associated with that kind of content because they think it is bad for their brand. Musk compounded the problem by attempting to shame them. He has made eratic policy decisions about Twitter Blue, increasing the number of bogus accounts impersonating real ones through verification; promising to promote these accouns above non paying users. He has sacked a large portion of the workforce particularly those in moderation, compliance and policy functions, increasing fears that content moderation will worsen, no doubt worrying advertisers even more.
These actions seem likely to have left his ownership financially unsustainable. [This Washington Post piece is a reasonable summary.] Twitter was rarely making money before he took over. Although he has shed about 2/3 of the workforce, advertising revenue has presumably fallen a lot [that WaPo piece includes a forecast of ad revenue, which used to make up 90% of total revenues being 40% down. And the company now has the burden of servicing $13bn of debt used to finance Musk’s purchase into a period of rising interest rates [as the Fed tries to bring inflation down, a policy Musk not surprisingly finds inconvenient and regularly criticises on Twitter]. An estimate of that debt service cost here has it at $1.2bn annually.
The panic over Twitter’s possible demise prompted a few eulogies for the webiste. Twiter has for many either begun or fuelled new careers; matched them with a partner; or just given people more of a voice than they had before, or made them new friends or some combination of the above. A nice example of the eulogy genre is this one from John Elledge. Here is another in the Atlantic - published in 2014! Here is Allison Johsnon in The Verge. There are more. [Send me your vote for the best].
There are divergent views on Twitter death. It is to be mourned because it would leave a hole in life, and the polity. Or it is to be welcomed, since time stolen by Twitter addiction would be returned. Or we should shrug it off as part of the inevitable internet life cycle or march of progress that had consumed Usenet, Friends Reunited, Myspace, blogging and others. A variant of the thesis of Twitter death was Noah Smith’s post about how centralized social media had perhaps proven a failure, something we did not want. On we should go to decentralized, niche social networks.
I want to advance a note of optimism. There are a few key components to Twitter; first, the invention of microblogging. Second, the proprietorial code and set of choices made by the Twitter owners to implement microblogging and other bundled functions for users; third, the actual networks of users; fourth, the outputs of microblogging - how people felt about it, what they got out of it, views they had about whether it was useful for them, in fact more entrenched than that, features of their life that may depend on it. [Eg a business model for advertising one’s own output and finding commissions].
If you break down ‘Twitter’ into these key components, it’s clear that the only thing that dies if the company Twitter disappears is the propietorial code and the most cosmetic of choices made about implementation [whatever you need to avoid copywright infringement, which judging from the competitors seems to be the brand name, colour scheme and some of the nouns used to describe actions and functions].
Importantly, you cannot uninvent microbloggging. This was invented and we saw that it was good! [Leave aside the irony in writing about the benefits of microblogging in an extended blog application]. People now know that they like to read microblogs, and they like to write them, or both. Along with that goes the demand for succinct insights from lay readers, comissioning editors and other writers for short comments on every issue from potentially unknown and non-credentialed or credentialed but previously unknown to mainstream media. If Twitter dies, this demand will look somewhere else for satisfaction. Somewhere microblogs will get written and read.
The code and the functionality of Twitter is now almost redundant. The emergence of Mastadon and Post.News makes that obvious. These other sites work as well and look as nice. [Prominent Mastodon refuseniks should go and try independent clients for that network like Elk.Zone or Pinafore.Social and explain why 2023 Twitter is any better]. Truth Social, another example, is actually a modfication of the open source Mastodon code, showing how Mastodon can be made cosmetically and functionally identical to Twitter if you want to. [It took the threat of a law suit from Mastodon’s developers to get Truth Social to own up to the fact that they had used the Mastodon code and deposit their version in a public archive, as required under the license agreement.]
The other thing that lives on somewhat - though not perfectly, and no doubt with deterioration over time - is information about individual users’ networks.
On a purely mechanical level, if you migrate/lifeboat to Mastodon as I have, you can choose from a few free apps that will search your Twitter follows for mention of Mastadon handles, find them on Mastadon and follow them for you. Obviously this only preserves your network in so far as it has already lifeboated to Mastodon. Your ability to reproduce your network somewhere else depends on the enthusiasm that those in it have for reproducing it themselves.
Quite a few others live on in your memory as a user. In my case this comprises writers, broadcasters, commentators, academics and snark experts that feature in my timeline a lot, that I am likely to remember and reconnect with once the network demanding herd has decided to where it is going to stampede [more on that below]. And, in my case, it also comprises a smaller number in my emails, whatsapp groups and my diary as I interact offline with people from Twitter that have become friends or like colleagues I meet and encounter regularly, who I will search out or will search me out once new handles have been set up, and help in the mutual reconstruction of our networks explicitly [‘have you seen if X has gone to Mastodon?’].
In short, each of us have our own mini networks, which we know pretty well, and want to reconstruct, and provied there is a place to go and do it we will make an attempt, and, if we all coordinate on one place to do the reconstruction, the aggregate effect of us reconstructing our mini networks will be to reconstruct an approximation to the Twitter mega network.
This brings us to the question of whether, even if you accept that we will all want to reconstruct our networks, and find somewhere new to keep refreshing them, and we will have online applications that give us an opportunity to do it, whether those alternatives will be good enough and whether we will be able to agree on a common destination.
This is difficult to predict, but I am optimistic. Here are a few reasons why.
I think Noah Smith is wrong to conclude that Twitter’s dysfunction proves that we didn’t want a centralized internet, or more specifically, a centralized text based social media application.
Twitter’s dysfunctions before Musk were that it struggled to make money; part of that was probably inherent to the business and part of it was probably because Dorsey and others were not very good or interested in turning the influencial societal phenomenon that had also made them rich into something that generated cash reliably. [I don’t think that observation is a contradiction; the fact that the shares were worth something was because enough people were convinced that with the right attention profits could be made]. These dysfunctions don’t lead us to conclude that we as users didn’t want a centralized text based media product; they don’t even lead us to conclude that our wants and how we acted on them could not be harnessed into generating cash.
A related argument, not the one Noah made, is made by the Mastodon and other decentralized internet fanatics. They want the centralized, proprietorial social media companies to die, and to be replaced with an open-source and decentralized ‘Fediverse’, comprising text based social media like Twitter, Instagram-replicants, and more. But the decentralization is not one that means there would be hard boundaries between components across which you could not communicate. The idea is that anyone can set up a server and host users; that the servers federate if they want to - and largely do. That you are not therefore at the whim one group’s management, commercial acumen, or content moderation. Instead the population of servers gives users some robustness. If one server admin proves unreliable or too tolerent of the wrong sort, users migrate to another and perhaps that admin gets ‘defederated’. So the communication network is potentially still operating as widely as before [ie on Twitter] but the network infrastructure is not centrally controlled.
Another reason for optimisim I think this that productive networks of users have a natural tendency to emerge; stable diffusion of them across competing networks is probably unstable. Each tiny improvement in the network [through more and better users] depletes the competitors and makes the new focal point even better, and so on, enriching the target and corroding the networks the users came from.
Some evidence that we would coordinate on a common destination comes from the Great Migration itself. The overwhelming ‘destination’ [in quotes because most people were not leaving, they were lifeboating, creating a parallel presence to which they could escape if Twitter sank, rather than leaving for good] was Mastodon. Although there was quite a bit of talk amongst established writers I follow on Twitter about Post.News, the numbers there and going there are tiny. Objections that Mastodon was too tricky to navigate the early hurdles, because of the server structure, are somewhat refuted by that mass migration. Somewhat, because although of those that migrated the vast majority went to Mastodon, obviously most Twitter users went nowhere at all [note the 330m figure above].
The alternatives to Twitter don’t look or feel quite like it yet, and so it might encourage people sniffing around to think that they won’t deliver what Twitter delivers now. But in my view this underestimates the likelihood of reinvention. Most of the reason why the alternatives don’t feel like Twitter are because most people are still on Twitter. Even those with accounts elsewhere are not that active - because most of the people they interact with are not there! The user interface and functionality differences are highly likely to evolve as developers respond to what new migrants want. [See, for example Elk.Zone compared to the original clients]. Mastodon seems to have its own cultural norms which might be off putting for Twitter exiles [or not]. For example a culture of politeness. Of not quote tweeting [you have to do this by embedding a link in a Tweet as there is no QT function]. Of putting conent that might be triggering behind ‘content warnings’ - common examples being some story about politics. But in my experience just in the last two months these cultural norms have changed. Virtually no content warnings now in my feed. And now dialogue about content warnings. Mastodon originals have either given up or slunk away to practice their cultural norms separately.
To say all this briefly: we have seen Twitter and we know we want it. The microblogging, the list of contacts, the voice and access. Developers are going to give it to us, whether Twitter itself lives or not.