Bluesky is just Twitter, without the burden of everything ruining Twitter.
It shouldn't work. But it absolutely could.
I’ve been on Bluesky for two weeks now. It’s so much better over there. While it’s still way too early to tell these things, I honestly think it’s gonna be the thing that replaces Twitter.
(no invite codes yet, sorry!)
That’s surprising, for a few reasons. Bluesky is just a barebones Twitter clone. It actually began as a Twitter initiative back in 2019, and was spun off as an independent company in 2021. The theory behind Bluesky was “what if we reproduced Twitter as a protocol, instead of a company?” (Give some credit to Mike Masnick here. Jack Dorsey read his “Protocols, not Platforms” essay and decided to give it a shot.)
It shouldn’t work. Twitter’s advantage from network effects ought to be insurmountable. I have ~42,000 followers on Twitter. I have ~500 followers on Bluesky. I have ~5,000 followers on Mastodon. Ceteris paribus, Twitter should remain the default venue for my attention minutes. It should be the place where I share my work, make wisecracks about NYTimes columnists, and think through what’s happening in the world.
If you had told me one year ago that a barebones Twitter clone with no gifs or DMs might replace Twitter, I would’ve donned my finest tweed-jacket-with-patches, dug my best pipe out of storage, and patronizingly explained why that would simply never happen. There would have been diagrams, and a reading list. It would have been exhausting.
A basic lesson from the history of social network sites is that you can only overcome the network effects of an established incumbent by offering a new class of activity. MySpace displaced Friendster by letting people create profiles for bands and alt identities. Facebook displaced MySpace by opening its API, allowing outside developers to create way more stuff to do on your Facebook account than on your MySpace account. That was when the zeitgeist shifted — Myspace become a ghost town, because everyone was spending their time at the other site.
BlueSky, by contrast, might manage to displace Twitter just by giving existing power-users a version of Twitter without any of Elon’s bullshit. That’s remarkable. It’s a testament to just how badly Elon has fucked everything up over there. Elon has turned Twitter into Myspace circa 2008. If there was a Darwin Award for managing a social media platform, he’d be an instant Hall of Fame candidate.
Every time I open Twitter, I’m greeted by some Elon or Ben Shapiro nonsense. Then I reflexively block a dozen paid blue checks at the top of the comments. (#gottablockemall). Then I feel gross and log off. Set aside the moral implications of spending time on a site owned by a shitty racist. There’s also the much more basic question of “is this even any fun anymore?” I had a Twitter-dunk go viral over the weekend. It still felt like a drag.
By comparison, every time I open up Bluesky what I immediately see is a bunch of jokes from my favorite twitter personalities. They’re having fun over there. The people who made Twitter good are all congregating on Bluesky. They’re relaxing and misbehaving. It’s awesome. (Check out Faine Greenwood’s piece, “How I Accidentally Ruined Bluesky with Pictures of Sexy Alf” to get a sense of how much fun everyone is having right now.)
[some background music: “Blue Skies” by Ella Fitzgerald]
That’s also what separates Bluesky from Mastodon. I still like Mastodon, but it’s a whole different vibe over there. Mastodon feels like an interdisciplinary academic conference (which, to be clear, is a high compliment. That’s my favorite type of academic conference, and I really enjoy academic conferences.) I share links to my work on Mastodon, and I get thoughtful feedback from people with interesting things to say.
By comparison, it’s common to read skeets on Bluesky that say things like “is it cringe to promote my work here? Are we doing that? Here’s a link.” And then people go back to being goofy as hell. (Also, they call posts “skeets” over there. Because why not?) Bluesky feels like a party that I don’t know how I got invited to, but I see a bunch of people I know and they seem happy to see me. Mastodon is a conference with great discussions, but I might get in trouble if I started cracking jokes.
Bluesky and Mastodon both run on protocols. But Mastodon’s growth strategy has been centered around the inherent value of protocols and federated networks. Bluesky has focused instead on seeding the network with Twitter’s power-users. They’re building a community on top of a protocol, rather than building a community around the principle that protocols are good. And the people joining Bluesky have spent decades participating in online communities. They are extremely online, in the best possible ways. There’s also a huge contingent of trans shitposters, who seem thrilled to be in a Twitter-like space where they can just make their jokes without facing a barrage of death threats for once. And they’re funny as hell.
Of course the hard part comes later. Bluesky still has less than 100,000 users. It has 2 million+ on the waitlist. One of the main ideas behind making Bluesky a protocol is that it will provide better solutions for content moderation at scale. We won’t have any clue how that all works out until Bluesky actually has to face the problems that come with scale. A reasonable retort to all my excitement is “Talk to me at 100 million, not 100,000, Dave.”
I’m optimistic about the underlying potential of replacing Twitter with a protocol like Bluesky though — not because of what a protocol is, but because of what it isn’t. A protocol is not a publicly-held company, with shareholders and vulture capitalists and all the dynamics that produce what Cory Doctorow calls the “enshittification” cycle. One of the main themes that has emerged from my history of the digital future project is that “big money ruins everything.” There is no Big Money in developing the AT protocol. No one is eyeing the down payment on their next megayacht with Bluesky. It’s basically the polar opposite of Web3.
Charlie Warzel wrote about the Twitter-Bluesky comparison last week, “How to build (and destroy) a social network”:
Twitter Blue may be boneheaded as a revenue scheme, but it excels as a case study in how grappling for status can ruin a social network—especially as a point of contrast to a new social network that the most online among us are flocking to. I’ve spent the past week clicking between two tabs: In one is Twitter, which is visibly suffocating under the weight of Musk’s enormous ego and a series of awful managerial decisions; in the other is Bluesky, a decentralized, invite-only clone of Twitter, which has exploded in popularity in just a few days, attracting celebrities, politicians, and a legion of beloved, extremely online shitposters. Status across social networks is always in flux, but the past few weeks have felt like a controlled experiment in how social capital is won and lost and how online communities respond to upheaval.
I’m also, admittedly, rooting for Bluesky to succeed because it would be the funniest possible outcome. Elon Musk bought Twitter because he said it was the public square, and therefore he ought to own it. But what if it turns out that analogy was all wrong? What if the thing that made Twitter special was the people — the power-users, in particular. What if it was a weird, giant party all along? And then Elon paid $44 billion to own the party. And then one of the original party hosts said “cool, thanks for the money. …Hey, uh, does anybody want to come hang out at the new place I’m buying with some of this money?”
While Bluesky keeps offering the experience of just-being-like-Twitter, Elon’s Twitter keeps circling the drain.
Elon found a new CEO, Linda Yaccarino. She’s a Trumpy ad exec, but Elon’s biggest fans are already furious because she chaired a committee for the World Economic Forum or something. (“Oh no, she’s connected to the economic elites!” they say to the world’s richest man.) Musk will be Executive Chair and CTO, in charge of product, software, and sysops. So, basically, she’s just there to woo advertisers and explain why they shouldn’t mind Elon’s conspiratorial rantings., while Elon keeps rage-tweeting conspiracy theories. We’ll see how many months she lasts.
One thing she’ll get to explain to advertisers is why they should be pumped about their ads appear next to animal torture videos. Ben Collins reports that graphic videos of animal abuse have started widely circulating on Twitter after the company ripped out the autocomplete safety filters that its Trust and Safety team had developed.
“Type-ahead search was really not easy to break. These are longstanding systems with multiple layers of redundancy,” said [Yoel] Roth [former head of Trust and Safety]. “If it just stops working, it almost defies probability.”
Animal torture-porn on main. It’s just so bleak.
Also, Elon’s other business interests ran into his ironclad commitment to free speech and, like a Tesla on Full Self Driving mode, just powered right through (credit to Nilay Patel, who called this on Day 1):
I talked about this with Peter Loge on our podcast this week. From a strategic communication perspective, even if you choose the coward’s route and give in to Erdogan’s censorship demands, the company has to either (a) say nothing, or (b) offer a smokescreen defense about how “complicated” and “delicate” these matters are. You definitely do not publicly announce “The choice is have Twitter throttled in its entirety or limit access to some tweets. Which one do you want?” Because if you say that publicly, then you have effectively just told every authoritarian regime “if you threaten to cut off access to Twitter in your country, we will agree to basically anything.” This is intro-level stuff. My undergrads could manage Twitter’s comms better than Musk.
Oh, and Tucker Carlson is relaunching his show on Twitter. Peter Kafka lays out the strongest case for why this might conceivably work. But, I mean… No. Just no. This is going to be a delightful train wreck. It’s going to fail because Elon has no follow-through and manages product rollouts on the basis of whims and grievances. It’ll fail because Carlson’s core audience are resentful boomers who leave the television on all day and have to ask their grandchildren how apps work (but their grandchildren no longer speak to them. Because of the racism.) Tucker’s audience is on Facebook, not Twitter. They might change the channel for him, but they aren’t going to completely upend how the consume media.
My best guess is that Carlson’s show will fizzle quickly, just like every other “Twitter 2.0” project. Both Carlson and Yaccarino will be gone before Twitter declares bankruptcy (which, by my reckoning, is about four months from now/whenever the FTC and EU announce their fines). And, honestly, good riddance. They deserve each other.
All the news about Twitter is bad. All the conversation happening on Twitter is increasingly bad. And it all serves as a reminder of how much better the Bluesky user experience is, simply by being Twitter-without-the-awful-bullshit.
I don’t have invite codes yet. But I think we might all be on Bluesky eventually. It turns out that the way you overcome Twitter’s network effects is just to build a simple replacement and invite the people who made Twitter fun in the first place.
Bluesky is Twitter without all the Elon nonsense, and without any expectation that it will make big money for investors.
And I suspect that may be enough.
Talk to me at 100 million, not 100,000, Dave.
You're right that that seems like the proper response, because of something you yourself linked to, the now-classic "Welcome to Hell, Elon" post by Nilay Patel, which I just reread since you linked to it (and boy does it hold up). Patel wrote:
"The essential truth of every social network is that the product is content moderation, and everyone hates the people who decide how content moderation works. Content moderation is what Twitter *makes* — it is the thing that defines the user experience. It’s what YouTube makes, it’s what Instagram makes, it’s what TikTok makes. They all try to incentivize good stuff, disincentivize bad stuff, and delete the really bad stuff."
If this is right, then it sounds from what you wrote that Bluesky hasn't been tested yet. It's invite only: it's fun because the trans shitposters can post without fear. That's great. But it's easy to do if you're only inviting people & are remotely selective about it. The question is will they be able to shitpost without fear *after* the doors are thrown open. And nothing you've said gives any indication of what the answer to that will be.
I haven't used Bluesky yet—not invited—and you make it sounds great. I hope it works! I liked (as well as hated) twitter pre-Elon (and you do underplay here how twitter was justly known as the hellsite *before* Elon took over, even as its users all kept coming back and also got things out of it), and if Bluesky can recreate that, great. (I mean, partly great. Its being a hellsite wasn't only do to easily-bannable-by-AI trolls: it was also due to the dynamics of attention and group action and out-of-context tweets and all sorts of things that would necessarily be recreated if Twitter Pre-E was recreated, since they were baked into its format). But I feel like all you're saying so far is that private parties don't have assholes in them. Which isn't much of a test, as far as I can tell.
But I'm not a media theorist, so maybe this is wrong. If you think so, I'd love to hear why.
It seems to me that the current state of BlueSky (a curated group with big names) is not a good representation of what Bluesky will be. If you ask yourself, how is this not going to be Twitter, you end up with a few observations and questions:
1. Is this just Twitter but with your own choice of algorithm?
2. How is this going to be paid for? It is a network of ‘personal data servers’ and ‘big graph servers’+’feed generators’. (see https://blueskyweb.xyz/blog/5-5-2023-federation-architecture) The latter is where all the data goes (firehose), requires a lot of money, and where the algorithms run.
3. If you rely on algorithms (and a free choice of labeling by others) to curate (or select your own bubble), how does that protect the providers against anti-hate-speech and all other societal rules? Will the party that runs the bgs say “It’s not me, guv” and get away with everthing? Because that bgs is what Twitter is doing now centrally: run the algorithms.
4. If there are many bgs’s, how much waste in traffic will there be as they all will want access to all traffic to put into their algorithms. They all want to ‘crawl’ all the providers. Instead of a push architecture, bluesky is a pull architecture, it seems.
These are just initial questions.