Life on the internet can feel like a proverbial family reunion dinner. One relative makes an off-color remark, and all the kids leap half out of their chairs, screaming at one another.

If you’ve ever sat through a dinner like that, you know how things usually end: someone throws their napkin on the table and storms off. Everybody nibbles on their casserole in silence.

Online, however, it’s a lot harder to leave. Instead we’re stuck at the table, forced to make conversation with our racist uncle.

The past few years marked an explosion of outrage, and we’ve only just begun to make sense of the carnage. Many have eagerly put forth their theories about why we’re all so angry. Maybe outrage is caused by tribalism, or a lack of context, or information overload, or companies riling us up for advertising clicks.

Here’s another one: maybe outrage is dissent without an outlet, a byproduct of platform lock-in with nowhere to hide and nowhere else to go.

Exit as an alternative to voice

In the 1970s, Albert Hirschman proposed the following economic theory: faced with deteriorating quality of goods, consumers have two choices: exit (leave our environment) or voice (stay and advocate for change).

Our environment, in this case, are the platforms we use to communicate with one another, like Twitter or Facebook. We can’t exit those environments, despite horrors like violent livestreams on Facebook, rape threats on Twitter, and disturbing children’s videos on YouTube, because these platforms have worked so hard to make sure that we don’t. Acording to Hirschman, when people can’t exit, they resort to voice, which looks a lot like outrage.

Right now, I see a lot of people advocating for voice. Users are frustrated because they don’t have a say in the governance of their platforms. We see this play out clearly in the case of Twitter’s efforts to combat spam and abuse.

Government is getting involved. Europe has long been ahead of the game on data protection and privacy laws, but in the past year, even U.S. lawmakers are taking menacing steps towards the so-called “Frightful Five”.

And platforms are responding. In mid-2016, Mark Zuckerberg famously scoffed at the idea that Facebook played any role in the U.S. presidential election. Several months later, Facebook, perhaps realizing that user loyalty was at stake, issued a swift turnaround to tackle misinformation, including allowing users to vote on news they find valuable.

Everyone is trying to stay together and make things work. But in an odd way, regulation implicitly strengthens our ties to these platforms. By investing in voice, we’re doubling down on a bad relationship instead of exploring alternatives.

Off-platform exits

What does it mean to exit? Exiting, in its most literal sense, is migration: leaving a platform in search of another digital land to call home. (In software development, this is forking.)

Taking a page from Rousseau’s Social Contract, there are two kinds of sovereignty. Platforms have legal sovereignty: they’re the watering holes we’ve decided to converge on. We’ve entered a tacit social contract with these platforms, but our true, natural sovereignty lies in the network itself (in other words: us).

Twitter has no inherent value, except that it’s where we agree to gather, find, and talk to each other. Uber has no inherent value, except that it’s where we agree to find, vet, and ride with each other.

Platforms have a responsibility to allow users to leave. Software isn’t the limiting factor here, but data. People should be allowed to export their data, and that data should be treated as a public good, managed by a neutral entity. Estonia, for example, has taken a crack at this.

These ideas are usually met with much scoffing and resistance, because it would require platforms to give up their biggest financial asset. While being able to export one’s data seems important, I imagine it will take government-level regulation, a catastrophe (just kidding), or building an entirely new data set to make it happen.

Without owning the data layer, network effects make it difficult for us to simply go off and start a new place to gather. Just ask Diaspora, App.net, or Ello. And perhaps that’s why we’re so focused on voice right now: because we see the graveyard littered with bones of those who tried to leave.

But while the old “let’s start a new Facebook!” rally-cry rings a bit hollow, it doesn’t mean that in-fighting is our only alternative. If the goal is to diffuse outrage, there are also on-platform exit options.

On-platform exits

Returning to the family dinner analogy, another possibility is making it easy for people to “exit” the conversation, while still remaining on the platform. Let the cousins wander off to the living room to grumble among themselves, even if nobody leaves the house. (In software development, this makes me think of sharding: breaking up data into smaller chunks to make it more manageable.)

Facebook isn’t a community. It’s a landowner: the wide plains and open skies to which communities flock. Your “community” is made up of the people you’re friends with, whose updates you see, and whom you interact with.

In that sense, platforms have a responsibility to let people organize themselves horizontally, on the same plot of land, until they find a configuration that’s sustainable. This is also how Mastodon designed their network, using a federated model. If platforms treat “exit” as a feature rather than a bug, they can retain their users by offering spaces for smaller, context-rich interactions.

We’re already seeing people do this:

• More conversation and activity now takes place in private messengers (FB Messenger, Whatsapp, etc)
• Blocking and muting tools are increasingly popular and effective (e.g. on Twitter), because they let you create the community you want to see
• Groups (ex. subreddits) are ways for people to break up their communities. Facebook has made Groups one of its top priorities as a form of “very meaningful” interaction
• Stories are becoming a preferred way to share content over newsfeeds — not, I suspect, because it’s a “better format” but because it feels more private and curated

Maybe the path to scaling social interactions is not starting over elsewhere, but making it easy to break up big communities into smaller ones.

This puts a hopeful, if somewhat dystopian, spin on the “let’s start a new Facebook” idea. If Facebook’s getting you down, just start a new one…but you’ll probably still do it on Facebook.

Challenges with federated communities

If we embrace the idea of “sharding” communities when they get too large, what are some potential challenges?

Influencers, by virtue of their audience and often polarizing viewpoints, are disproportionately affected by today’s mob mentality. That means more of the interesting conversations are moving to private messaging.

If that trend continues, I wonder if we’ll end up in “knowledge ghettos”, where influencers no longer contribute their best knowledge to the public sphere.

The analogy here would be a poorly-funded public school, where all the rich parents in the area send their kids to the nice private school, instead of investing in their local public option. If social media is a public space, history tells us that public spaces tend to become under-resourced, partly because the elite can afford better options.

I doubt that influencers would disappear entirely. After all, they still need their audience. Instead, the public would still benefit from their knowledge (ex. writing, podcasts), but it’d be more of a one-to-many relationship. There could also be a two-way mirror relationship, where influencers read whatever non-influencers publish, but you can’t see their reactions unless they choose to engage privately. This would take us back to the pre-internet days of communication, where most people are consumers, rather than co-creators, of high-value content.

Whatever the case, it’s in a platform’s best interest to retain influencers, because they bring their audiences and help maintain the quality of public conversation.

---

Embracing the filter bubble

One theory behind why social media isn’t fun anymore goes like this:

The internet was fun in its early days because there weren’t that many people using it. Now that everybody is online, it’s just as tedious as the real world.

We’re currently in the process of bringing our physical world online, and that transition is painful. We can see fascists marching in Charlottesville, Virginia. We can see activists rallying in North Carolina. Everyone is trapped at the party, sullenly sipping beer in the corner, eyeing the crowd and thinking “Who invited those people?”

In the halcyon tech days of 2010, Eli Pariser warned us of the dangers of the “filter bubble”. Given our scale today, maybe bubbles aren’t such a bad thing. Maybe they’re just realistic. If so, letting people break off into their own conversations, instead of trying to keep them in the same place, is the world we should design for.

Platforms have to choose whether to provide exit or voice, but as the de facto governments of the digital world, they also can’t sit on the fence and do neither. To do nothing is autocracy.

(Cross-posted from Medium.)