I had something else I was going to write about, but absolutely all conversational oxygen has been sucked up by Elon Musk buying Twitter, so I feel kind of obliged to say something about that. I’m largely in the Kevin Drum camp of thinking this is massively over-covered, but Twitter is enormously popular with people working in journalism and media, the only group whose self-absorption is more tediously worked out in public than that of professional academics (who narrowly edge out Hollywood actors), so it’s inevitably going to be overblown.
But, you know, I also spend way too much time on Twitter, so it’s been hard to avoid reading more than I really ought to about the purchase, much of it operatically stupid. There have been a handful of reasonably good takes, though, in among all the dross. I think Charlie Warzel probably has it about right with the “timeline” he identifies as the most plausible (Muck makes a few largely cosmetic changes, then mostly hands things off to underlings). I found this suggestion to split Twitter into two different businesses intriguing— the idea being that the actually valuable thing here isn’t the app and ads but the social network, so you could make real money selling access to that to people who could then make new front ends offering more highly customizable experiences. Matt Yglesias has a characteristically slightly overlong list of mini-takes, and Ethan Zuckerman has a good thread with discussion of alternatives to Twitter as we know it:
While Ethan’s a really smart guy, though, I think a lot of his ideas run afoul of the same problem as the suggestion that the whole problem is algorithmic feeds, or more broadly the idea of letting people share takes, as Jonathan Haidt argued shortly before we went on vacation. The core issue is that the features being identified as Bad are things many users actually want.
Oddly, the clearest explanation I’ve seen of this comes from Ben Dreyfuss, who’s often a bit of a shitposter supreme, but whose brief history of Facebook jibes pretty well with my recollection of how the current features evolved. I think this is good enough that I’ll do the embed thing (even though nobody clicks through to those, either):
As Dreyfuss notes (and notes that Haidt elides), a lot of the changes that happened at Facebook and later Twitter were driven by users:
The retweet button was introduced by Twitter only after users had on their own invented what came to be known as the “manual retweet.” People wanted to share other people’s tweets so much that it became a convention to write “RT @AshtonKutcher Dolphins don’t seem that smart.”
After years of this, Twitter gave in to the demand of their base and supplied a way of doing it that showed the original tweet in your feed, so you 1) knew it was real, and 2) could decide to follow that person yourself if you wanted.
There were lots of consequences of the native retweet. It led to user growth and lots of other things. But it was not invented or designed by Twitter. It was a response to an organic creation of their users.
He’s absolutely right about the RT function and the “is” at Facebook (I’m old enough that my daily “memories” from Facebook usually include a few that don’t make sense without being preceded by “Chad Orzel is…”). I suspect but don’t really know that something similar is at work with the algorithmic timeline— that is, that people who aren’t the big power users of the service who bitch about it endlessly actually like it.
I say this in part because I’m personally more positive about it than a lot of folks in my feeds. This is a direct result of the way my days are scheduled: I’m most free to use Twitter in the morning hours in the Eastern US, when approximately nobody else is online, and least likely to be on and active in the late-evening hours when most of the cool kids are tweeting up a storm. A purely chronological timeline for me in my peak Twitter-using hours would be dominated even more than it already is by incomprehensible outrage about Australian politics. Leaving the algorithmic feed on lets me see some of the highlights from the previous evening that would otherwise be lost in the flood of the strictly chronological feed.
(This is not to say that it’s perfect— it has a tendency to keep resurfacing a half-dozen tweets that I’ve already seen, sometimes for days at a time. But it’s wayyyy better than the chronological version; I’ve tried the comparison multiple times at this point.)
Now, this doesn’t necessarily mean that these tools aren’t Bad, nor that social media wouldn’t be better if they were eliminated or restricted in some of the ways that people have suggested. It does mean, though, that at the core they’re things that people want, and eliminating them by executive fiat isn’t going to go as smoothly as many tech writers seem to think.
To the extent that these tools are really a problem (as opposed to Just One of Those Things that will eventually get sorted out by new social norms), they’re a problem of people wanting things they “shouldn’t.” Which can, in the end, only be solved by convincing them to want something else. Which is a slow process that ultimately has to be done from the ground up, not something that can swiftly be enacted by a change of top management.
This is a problem that goes well beyond the world(s) of social media, of course. A great many of the hugely divisive political issues of the current moment are similarly about large numbers of people wanting things that other groups of people think they shouldn’t want. And these issues seem incredibly intractable in large part because the people with strong opinions are trying to short-cut the process, to skip past the hard work of convincing the general public to want something different by getting courts or corporations or Congress to simply impose the “better” version of whatever.
So, where I end up is basically the paraphrase of Douglas Adams that I used for the post title: People are the problem. The tools and rules of social media are not entirely an external imposition, but reflect revealed preferences on the part of the users of these services. And that means that the ultimate solutions will have to come from those users changing their preferences, either through the spontaneous evolution of new norms (which Dreyfuss thinks will happen, but having spent the better part of thirty years watching college professors fail to grasp email, I’m less sanguine…), or through a concerted effort at persuasion (that doesn’t seem to be happening yet).
As a result, I don’t really expect anything to get solved quickly. So, to end on a somewhat fitting note, I’ll outsource my final prediction about this whole mess to somebody else on Twitter:
So, there’s the muddled position that everybody expected and nobody demanded. Here’s a button if you’d like more of this:
And here’s another if you’d like to leavea comment to tell me I’m wrong:
"And that means that the ultimate solutions will have to come from those users changing their preferences"
It feels like this is happening to some degree, with people migrating to either other platforms like Tiktok, or to more private/temporary platforms, such as giant group texts or Discords.
So put and excise tax on something excessive re-tweeting, intensive use, etc. and let the algorithms figure out how to get less of it.