Ignore the Haters. Bluesky is Great.

Numerous headlines, from the Washington Post to Slate, blare about the problems with BlueSky: it is boring, full of scolds, and an echo chamber for liberals. Many critics demand people on BlueSky return to Twitter and simply suffer the trolls, accept they can’t share links, mingle with neo-Nazis, and tolerate the ubiquitous advertisements for erectile dysfunction gummies. In short, they argue, people on BlueSky are using social media wrong.

This thinking betrays a grim belief in our relationship with technology, one that denies our agency: technology decides what we can do with it. This outlook is becoming increasingly familiar, as tech monopolies have grown so dominant that user satisfaction is largely irrelevant. For many people in tech, human agency isn’t something to be enabled by new technology; it is a nuisance to be bulldozed in product design. Advocates for a return to Twitter/X are arguing BlueSky people need to give up agency and accept a thoroughly worse environment and experience. In short, we need to suffer our technology, no matter how awful it is.

But the shift to this way of thinking is recent—and isn’t mandatory. Technology companies have progressively been worsening user experience—a process dubbed “enshittification”—precisely because people haven’t opted out. We’ve been the frog in slowly boiling water, seemingly oblivious to the option of getting out. The condescending pushback on BlueSky users is a good moment to reverse the issue: what do we want social media to do—and how do we get it to do that?

The Functions of Artifacts

We typically categorize things by their functions: hammers hammer and air fryers fry. There is, in short, a right way to use it and, if we use it right, it does what we want. The digital computer, however, is a radically different beast than can openers and door knobs. It doesn’t have a function; it has an infinite number of possible functions. And it isn’t like the multitool in your pocket, because new functions are constantly being added to it. It is "pluripotent."

With most single-use artifacts, the technology comes with a standard use--but it’s fundamentally a suggestion. The company won’t give me a refund if I ruin my graphics card by frying an egg on it—but they also won’t stop me from wrecking it (though they will contend I've voided the warranty). But software products often lean into the pluripotency, encouraging users to "mod" them to do new things. Early personal computers depended on users doing both, with simple operating systems like DOS requiring users to engage in a bit of programming to reconfigure their systems to run different programs (usually by re-writing start up commands to allocate RAM memory in the right way). A lot of software at the time was open-source and, as such, people regularly tweaked and improved it, offering the updates for free over BBS forums.

The exception was Apple: it recognized most people hated the steep learning curve associated with programming a machine and tweaking software. It instead offered a streamlined experience, where users didn't need to consult the manual or browse the online bulletin boards. It took away a lot of agency in exchange for simplicity (best exemplified in the Apple mouse having only one button compared to the PC standard two). The experience was narrower and more rigid, resulting in a small, enclosed ecosystem without many programs or options. But what you got always worked with your machine, and a lot of the old machines are still running because of their simplicity (something that can't be said of, say, Windows 3.1).

The Rise of Apps

Although Microsoft drove Apple to the brink of bankruptcy with its success in computers, the Apple model has dominated phones. Their success in phones, moreover, left its mark on the design of "apps," the (typically) phone-specific programs. Pre-phone social media websites, like MySpace and early Facebook, still permitted a good deal of agency to users, who could decide what information they wanted to share and who could see their profiles. This allowed communities, like a university, to have private networks for on-campus students largely separated from peering adults (like nosey parents). MySpace, for its part, encouraged users to design their own sites.

As phone apps became more common, however, they started to adopt the Apple model and began streamlining their service. Streamlining, in these cases, usually meant reducing users' agency: maximizing engagement demanded users become more passive consumers in the experience—scrolling more and encountering ever more engaging, salacious things that kept them hooked. TikTok became a hit precisely because it demanded nothing of the user, providing a constant stream of short, entertaining clips without any need for a reaction.

Most people barely noticed the loss of agency on the apps--and some welcomed it over the tedium of building one's own profile. People took it as normal that they’d occasionally encounter unwanted trolls, insults, or nudity. There was no requirement of this. Reddit, for example, enabled user agency by permitting moderators to limit unwanted and irrelevant posts that made people's experience worse. But other apps became increasingly mindless, a simple time-killing experiences with little rhyme or reason, because this worked and most people accepted it. The experience became more unpleasant--more ads, more trolls, more inappropriateness and toxicity--but it also became less demanding.

The Fall of Twitter

But the strange COVID years revealed that the shittiness of apps is a design choice--something optional that companies can turn off if they want. Limiting COVID misinformation, banning Trump, and increasing fact-checking revealed that companies could aggressively moderate toxic users and, as a result, create a less unpleasant experience.

For right-wingers, the Twitter ban also provided an opening: they realized they weren't required to stick with Twitter. Many jumped to Parler or Truth, seeking their own space that wouldn't be censored. The purchase by Musk brought some back, but also encouraged others to flee to BlueSky and Threads. The resulting "X," for its part, grew its own audience—almost entirely male, tech-obsessed, and endlessly contrarian--that drove out a lot of other users.

As the marketplace for social media apps has grown, speciation has occurred. The right-wing apps have limited censorship and permitted the spread of conspiracies and hateful memes. BlueSky, for its part, focused instead on enabling user agency, much like the early internet and social media apps. The algorithms for content are simple and not optimized for engagement or for discovering new contend. They require users to decide what they want to see and build up their own feed--which means it is a lot of work. And work isn’t fun, so many users abandon the app because it doesn’t make itself useful and enjoyable. And, of course, many people, for their part, instead remain on Musk’s app, finding it more relaxed than its prior, moderated iteration.

The Return of Agency

The competition between apps is a sign of (more) healthy user agency: BlueSky requires each user cultivating an experience, whereas X has embraced a top-down structure with users receiving an increasingly homogenous experience of high-engagement tweets. There is nothing in the design of either that is "better" or "worse" in an absolute sense, though there are reasons different groups might prefer one or the other.

For people at legacy publications like the Washington Post or folks with contrarian Substacks, their paycheck depends on engagement. In their case, X is better because it can push content on people who won't like it and don't want to see it. If you're hoping for a rage-tweet to help you get views, no better way than some well-placed hate. For many of these folks, getting folks away from Bluesky is necessary for keeping a steady readership. If they stop being a topic of conversation, they stop making money.

But if you'd rather lop off a limb rather than see another "contrarian" article by Chait or Stephens, it is easy to set up BlueSky to avoid it. If all you want is philosophy articles and recent work in cog sci, your whole feed can be built around that. The app doesn't push other stuff--even if you might find it interesting, thought-provoking, or whatever. You'll miss a lot of spontaneous content because you've built a "bubble." But you're allowed to build bubbles: if you just want to look at dog videos, it's your life. You can decide how--if at all--to use this technology.

A common rebuttal is that Bluesky permits ignorance of unpopular ideas and their defenders. To be sure, there are principled reasons why it is a bad thing to shut ourselves off from the marketplace of ideas. But these principles only apply in a frictionless universe; in our world, where we have limited time and opportunity costs associated with reading one tweet/article over another, we need to be judicious. So picking what ideas to give space to is a normal and necessary part of life.

I spend no time reading Creationist scientific literature, or climate change skeptical pieces, or even stuff in my "field" that seems silly, like article on pantheism or Less Wrong pieces on superintelligence. There already aren't enough hours in the day for family, friends, and my own interests. Am I missing out on potentially life-changing information from my disinterest? Possibly. But it's a calculated bet based on my background beliefs and assumptions that it'd be mostly garbage that wouldn't move me.

A different rebuttal is that if lefties and liberals want to reach a broad audience, they need to be where the people are. This seems right--but it includes two dubious assumptions. First, not everyone wants to reach a broad audience. While some may want that, others are content cultivating their own little garden, and many people are passive consumers of social media rather than producers. For many, depth is simply more valuable than breadth.

Second, and just as important, is the false assumption that X is "where the people are." This is silly, as anyone who has seen the site in recent years can attest. The site is absurdly skewed in terms of demographics, with the paid "blue-check" content producers especially unrepresentative of the population as a whole. Pretending that this odd subset of the population is "the people" is an act of marketing for those selling their wares, rather than something backed by the data.

Bluesky is Great because it isn't for everybody

The upshot is that we've actually hit a good spot with the fracturing of Twitter in recent years. Many software products are becoming increasingly anti-user, as seen with Google's interminable ads on YouTube and useless AI search. But Twitter never had a full-on monopoly that allowed it to enshittify without consequence. The rise of Bluesky--as well as Threads, Truth, and the other clones--allows space for people to make the technology experience they want.

It is no surprise people want to undermine this; user dissatisfaction is always a threat to business models dependent on a lack of alternatives. If people can simply shop elsewhere, businesses can't make things worse without losing customers. Bluesky is great because it isn't for the contrarians, trolls, and paid promoters. That may limit its appeal--but the limitation is its chief selling point: there are a lot of jackasses in the world, and Bluesky lets us avoid them.