Matteo Della Torre/NurPhoto via AP
One bright spot in the bleak year of 2024 was the rise of Bluesky. As someone who relied greatly on Twitter for news and my career—OK, I may have been somewhat addicted—before Elon Musk bought it and turned it into a snake pit of neo-Nazi filth, it was nice to see a Twitter-like replacement rise to relative prominence.
I joined in April 2023 as about the 47,000th user. Today, Bluesky has about 26 million users, and seems to be growing healthily. It actually has some notable improvements on Twitter, like the “starter pack” function where users can put together a group of accounts that one can follow at once (here’s the starter pack for Prospect writers, incidentally), or the “nuclear block” where if one participant in a conversation blocks the other, the entire conversation is zapped. This greatly cuts down on Twitter’s culture of aggressive pile-ons and abuse.
Unlike any other big platform, Bluesky does not censor posts with outgoing links. Indeed, it does not have any proprietary “for you” algorithm, instead defaulting to a traditional reverse-chronological feed, and allowing users to pick from algorithms that can be developed by others. This has major implications for publishers: Despite its modest size, The Guardian reports that Bluesky traffic has already outstripped that from Twitter, and here at the Prospect Bluesky traffic now regularly matches Twitter and is many times that of Facebook.
This ability to share outside the platform is proving so popular that Facebook’s Twitter clone, Threads, has belatedly altered its algorithm to include more posts from accounts you follow in an attempt to compete. And this disruption is being done on a shoestring budget—Bluesky has just 20 employees and about $23 million in funding, as compared to Meta’s 70,000+ workers and $156 billion in annual revenue.
It’s strong evidence that there is a large unmet demand for internet systems outside of the control of Big Tech monopolists. I don’t know if there can be a similar option for every walled garden on the internet—it’s hard to dislodge a giant—but there’s no question that there’s a lot of pent-up demand.
THE HISTORY OF HOW THE BIG PLATFORMS got big is not that different from the history of any kind of market rollup. Back in the 1990s when the internet was first taking off, there was an explosion of millions of different websites—home pages, forums, blogs, retailers, you name it.
This started to change as technology evolved. In the ’90s, any reasonably clever person could learn how to code a website after a few nights of studying HTML. But as websites became more complicated, creating one quickly outstripped the technical skills of all but the most dedicated hobbyists. It became much easier to simply use someone else’s tools—first site development services like Blogspot or WordPress, and then separate platforms like Facebook and Twitter, where you could establish a presence simply by maintaining an account.
Then, as billions of people got online, network effects started to take hold. Once a critical mass of users ended up on one platform or another, most people joined up to be in the same place. This growth was critically dependent on the American legal structure, particularly Section 230 of the 1996 Communications Decency Act. When you are hosting your own website, you are legally liable for the content you post on it. A television news program or magazine like the Prospect can be sued if they publish false and damaging claims about someone. (That’s why we fact-check our reporting.)
Section 230 stipulates that if you own a platform where others post, you are largely immune from lawsuits. Ironically, the idea was to actually encourage content moderation, because at the time there was a fear that websites that did moderate would be legally vulnerable as compared to those that took a totally hands-off approach. In reality, Section 230 enabled the growth of gigantic platform monopolies, which do as little moderation as they can get away with.
This remained true even when the platforms started implementing black-box, attention-maximizing algorithms that control what their users see. Feeding people content by algorithm is indisputably publishing by any realistic definition—it’s not much different from self-selecting a bundle of stories and putting them into the morning edition of a newspaper. Yet Section 230 still protects the platforms.
Gradually, most of the vast ecosystem of independent sites died off, replaced by platform accounts. Fast-forward to today, and a handful of monster platforms dominate the internet: Facebook, Instagram, YouTube, TikTok, Snapchat, Amazon, Google, and so on. As a rule, they do everything possible to keep you on the platform and scrolling, so they can sell more ads or products.
The results have been a catastrophe for humanity across the globe. The best way to grab and keep attention is inflammatory, anger-inducing content—above all, hate speech and conspiracy theories. As Steve Randy Waldman writes, “Section 230 has created artificial and destructive economies of scale.” In the worst cases, as when Facebook barges into poor countries offering free access to its platform but not bothering to hire many (or any) moderators speaking the native languages, the result is an instantaneous explosion in racism and hate crimes. In Myanmar, this proceeded to genocide.
But even in rich countries, social media platforms are associated with a rise in bigotry and violence. In Germany, Facebook use seemed to cause a spike in attacks on refugees. South Korea’s very president attempted a coup and nearly touched off a nuclear war, reportedly because he cooked his brain on right-wing, Korean-language YouTube.
There are many potential solutions here. As Waldman suggests, we could repeal Section 230, at least for big algorithmic platforms, forcing them to moderate much more heavily so they restore some level of internal responsibility. Or as my colleague David Dayen suggested years ago, we could ban surveillance advertising, and hence make the targeted, algorithmic feed impossible.
Any such change is very unlikely under a Trump presidency. But what Bluesky proves is that the platforms have gotten fat and lazy. Facebook today is a buggy, slow mess with a bewildering user interface and a feed full of AI slop and other garbage. Facebook is even inserting AI-generated chatbots into user feeds, which has spawned mass outrage. Meanwhile, Twitter/X is a fascist hellscape. YouTube is somewhat better, but plainly has some work to do in its non-English-speaking regions. So the minute a tiny bridge over the platforms’ moat of market power appeared, a flood of users rushed over it.
The old internet wasn’t perfect, but it used to feel like a place of almost magical potential, where you might stumble over a fascinating new site at any moment, instead of watching in horror as your aunt is driven mad by flat-earth conspiracies. This internet did not die of its own accord; it was murdered by a handful of mega-billionaires so they could sell ads for chukka boots and washing machines.
Even with the platforms’ overwhelming market dominance, there is an opportunity for a company to establish itself as the nexus for an old-fashioned internet of independent sites. Bluesky may or may not end up there, and at a fraction of Meta’s user base for Facebook and Instagram, it’s not going to topple the leviathans by itself. But it’s the most promising new development online in decades. I hope it develops a monetization strategy that’s not dependent on surveillance advertising; I for one would happily pay a monthly subscription to be the customer rather than the product. But the fact that a company this microscopic has Big Tech behemoths flailing around to compete for the first time in years is a good sign.
