Late in 2019, Twitter CEO Jack Dorsey floated "Project Blue Sky," a plan for an interoperable, federated, standardized Twitter that would let users (or toolsmiths who work on behalf of users) gain more control over their participation in the Twitter system. This was an exciting moment for us, a significant lurch towards an interoperable, decentralized social media world. We were especially excited to see Dorsey cite Mike Masnick's excellent Protocols, Not Products paper.

It’s been more than a year, and Twitter has marked its progress with an “ecosystem review” that sets out its view of the landscape of distributed systems. The world is a very different place today, with social media playing a very different role in 2021 than it played in 2019. Locked down, with the legitimacy of the US political system under threat, with social media moderation policies taking on national significance, the question of how a federated, decentralized internet would work has become even more salient...and urgent.

Wise Kings vs Autonomous Zones

It's a safe bet that no one is thrilled with the moderation policies of the Big Tech platforms (we sure aren't). But while we can all agree that tech has a moderation problem, there's a lot less consensus on what to do about it.

Broadly speaking, there are two broad approaches: the first is to fix the tech giants and the second is to fix the Internet.

The advocates for fixing Big Tech are implicitly saying that the net is a fait accompli, with the tech giants destined to reign over all forever. Fixing Big Tech means demanding that these self-appointed dictators be benevolent dictators, that the kings be wise kings. To that end, fix-big-tech people propose rules and structures for how platforms relate to their users: clear moderation policies, due process for the moderated, transparency and accountability.

By contrast, the fix-the-Internet people want a dynamic Internet, where there are lots of different ways of talking to your friends, organizing a political movement, attending virtual schools, exchanging money for goods and services, arguing about politics and sharing your art.

We're fix-the-Internet people. Sure, we want virtual spaces to be well-managed, accountable and transparent, but we also want there to be other places for users to go when they're not. Wise kings are great in theory, but in practice, they are just as fallible as any dunderhead. When the platform gets it wrong, you should be able to pick up and leave -- and still reach your friends, show and sell your art, and advocate for your causes.

That's where interoperability comes in. The problem isn't (merely) that the CEOs of giant tech companies are not suited to making choices that govern the digital lives of billions of people. It's that no one is suited to making those choices.

Switching Costs: Tear Down That Wall

Twitter has lots of competition today: there's Facebook, of course, but also the thousands of Mastodon and Diaspora instances that make up the "fediverse." If you don't like the moderation choices at Twitter, you can certainly take your social media business elsewhere.

But it's not quite that simple. If you like someone else's moderation policies better than Twitter's, you might still hang around on Twitter, because that's where all the people you want to talk to are hanging out. What's more, the people who you want to talk to are still on Twitter because you're there. It's a kind of mutual hostage-taking.

Economists call this a "network effect": the more people there are on Twitter, the more reason there is to be on Twitter and the harder it is to leave. But technologists have another name for this: "lock in." The more you pour into Twitter, the more it costs you to leave. Economists have a name for that cost: the "switching cost."

Network effects create lock-in through high switching costs.

But here's the corollary: lower switching costs reduce lock-in and neutralize network effects!

Let's unpack that. Berlin was once divided by a wall. People on the eastern side of the wall weren't allowed to leave. If you tried to leave, you had to abandon everything and everyone you loved, and if you made it, you might never see or talk to them again. The Wall was a big deal, but it was only the visible part of a much larger system that bound the people of East Germany. The soft controls -- the social and material costs of abandoning your old life -- were just as much of a disincentive to leave.

By comparison, if you want to move from Berlin to, say, Paris today, you can take the train to Paris and check it out. If you like it, you can have all your worldly goods moved there (even your appliances!) -- and you can video-conference, text, and email with your Berlin friends. If you're lucky enough to have a spare room in your Paris flat, your friends can come for weekends (if not, you can put them up on the sofa). You can tune into your hometown radio and TV channels from your screens, and read the daily Berlin papers at the same moment as they're made available in Berlin itself. And if it doesn't work out in the end, well, you can move back to Berlin.

These low switching costs mean that many people have experimented with living somewhere else. Some people move away for university and never come back, or try a work-abroad scheme and fall in love with a new city. Lots of people change their minds and go back home. They have choices. They can try a new life out before they commit to it, and even after they commit to it, they don't have to write off the life they lived before. They can have one foot in the old place and one foot in the new place. They can swing through the jungle -- but not let go of the vine they're holding onto until they've got the next one firmly in hand.

But once you commit to a social media platform, you're behind a digital wall. The platforms only recently started to allow you to bring your stuff -- your messages and address book -- with you if you decided to leave (and they only committed to this tiny bit of freedom once the EU started to regulate them).

One thing they definitely won't let you do is hop over their wall and then lob messages back over it. Facebook has repeatedly sued and threatened others merely for daring to create an alternate interface that aggregated messages from Facebook with messages from other services -- the idea of allowing Facebook users to talk to their friends on their own terms was too much for the company. It's hard to believe they'd tolerate a world where people who left Facebook could still talk to their friends and read messages in their groups.

The parallels between walled gardens and the Berlin Wall don't stop there: the East German government maintained that the Wall wasn't there to keep people from escaping; rather, they said it was there to stop westerners who longed for the East German lifestyle from pouring across the border. Today, Facebook insists that it blocks interoperability to keep privacy-plunderers out of its service -- not to trap its users inside.

Fix the Internet, Not the Platforms

Facebook has a point. Not a good point, but a point. After all, the company is under pressure from all quarters to fight bad speech on its platform -- not just illegal material, but also disinformation, misinformation, harassment and other unpleasant or frightening communications. The more ways there are to connect to the platform, the more ways there are to make mischief.

But we need to ask ourselves why people stick around on the big social media platforms despite the harassment, hate speech, and other unpleasant experiences. It's because of the mutual hostage-taking of lock in and the high switching costs of taking your social media activity elsewhere. You use your social networks because your friends are there. They're there in part because you're there. The cost of losing the connection to your friends (and, for artists, you audience; and for businesses, your customers) exceeds the benefits of fleeing to a rival.

Interoperability reduces that cost and makes it easier to switch. Specifically, it lets you switch to smaller platforms whose administrators can enforce policies that suit you, personally: platforms whose definitions of "hate speech" or "harassment" or "rudeness" meet your own standards. And critically, it allows you to make the switch without losing contact with the colleagues and friends you collaborate with, nor with the inspiring or amusing (or enraging) strangers you follow, nor with the fans who follow you to be inspired or amused (or enraged).

In an interoperable world, we still need to advocate for the platforms we use to adopt good rules and then uphold them -- but we have alternatives when the platforms let us down. If you can fix the Internet (by restoring choice), fixing the platforms isn’t nearly so urgent.

Interoperability's Downsides

There's a downside to interoperability. When the massive, monopolized platforms that dominate today's Internet wield their power for good, good things happen. For example, if a platform adopts strict privacy rules (like transparency reports, limits on third-party access to user data, bans on selling or mining user data, high legal standards for official content removal demands, and so on), and then actually follows its own rules, hundred of millions -- or even billions -- of users reap the benefits.

In a decentralized, interoperable Internet, it's much harder to enforce policies in ways that affect billions of people at once. That's the point, after all.

And if users can easily switch platforms without giving up access to their social circles, then platforms that countenance harmful or undesirable speech will accumulate users who enjoy that sort of thing.

But we need to distinguish between undesirable speech and illegal speech. The U.S. Constitution's First Amendment means that, in America, most speech is lawful (even odious speech, such as racist diatribes), but some communications (for example child sex abuse materials, or nonconsensual pornography) are not.

The present-day centralization of online communications means that even speech that is lawful can be removed from much of the Internet. If Facebook or Twitter ban a word or phrase or link, many people will never see it. If you agree with Facebook and Twitter's judgment, then this is great news. One thing that is increasingly apparent, though, is that most people -- irrespective of their political orientation -- are not fond of Big Tech's editorial judgment.

A decentralized, interoperable Internet is an explicit tradeoff: you lose the power that comes from convincing the platforms to eliminate the speech that disturbs or upsets you, but you gain the power to move to an online community where speech policies are to your liking...and you still get to talk to people who want different speech policies.

After all, most of us can't convince the platforms to bow to our demands for new speech policies.

But what about illegal speech? Fraud, nonconsensual pornography, serious incitements to violence? Well, these remain illegal, and courts and prosecutors (as well as private persons, in many cases) have the legal right to punish people who use platforms to spread this unlawful speech. What's more, depending on the kind of speech and the platform's complicity in it, the platform itself may share liability for criminal speech.

In an interoperable world, the decisions about what speech can't appear in a given community are made by the community members. The decisions about what speech can't appear anywhere are made by democratically elected lawmakers (not giant corporations that bow to some form of public pressure).

Likewise, in an interoperable world, we can’t rely on the platforms to protect our privacy. But they haven’t been very good at that so far. Instead, we will need to rely on privacy laws -- such as a federal privacy law, which is long overdue -- to protect us.

Everyone has some flavor of speech -- "icky speech" -- whose very existence is a source of distress, even if they never have to encounter it. Just knowing that it's out there is unbearable. When a broad public consensus emerges about that speech -- as with child sex abuse material -- it is prohibited by law. But if no consensus emerges, then each of us must bear the unbearable: we can avoid the speech, we can shun or criticize the people who engage in it, but we can't reasonably expect that the speech will cease, no matter how badly it makes us feel.

Twitter's Unique Position (Number Two)

Twitter has 340 million users. That's a big number.

Facebook has 2.6 billion users.

As Mark Zuckerberg has reminded us, Facebook's moderation staff exceeds Twitter's entire headcount. No wonder that Zuckerberg has called for a set of rules for what speech he should block on Facebook: even with his army of moderators, he'll fail to live up to those rules, but everyone else will fail much, much worse, clearing the field of Facebook competitors and miring us all permanently in Facebook's choices about what is and is not acceptable speech.

Twitter is big, but they're an order of magnitude smaller than the market-leader. They have a lot of resources, capital and users -- but they could easily be swallowed up or forced out by the giant that dominates the field.

That puts Twitter in a unique position: big enough to do something ambitious and technically expensive, precarious enough to take the risk.

Thus far, "Project Blue Sky" is just a lofty goal. It remains to be seen whether Twitter will live up to it. But while optimism is in (justifiably) short supply in this fraught moment, an interoperable social Internet has never been needed more. Social media companies make errors, just like all of us, but their vast numbers of locked-in users makes these companies’ errors more consequential than any blunder any of us might make. It’s fine to demand better judgment from the tech companies - but it’s far more important to reduce the consequences of their inevitable screw-ups, and to do that, we need to take away their power. Interoperability moves power from corporate board-rooms to tinkerers, co-ops, nonprofits, startups, and the users that they serve.