I’m sure we’ve all seen the news that Donald Trump has been banned from Twitter, Facebook, Twitch, Shopify, YouTube, Instagram, Snapchat, and TikTok. Pinterest, Discord, and Reddit have similarly cracked down on Trumpy content. Parler, a social network for people that find Twitter insufficiently extreme, has been banned from the Google Play store, and Apple has fired a shot across Parler’s bow that will probably end with the app being banned from the App Store. (Thanks to commenter Wyatt Salamanca for the above link and the reminder to write this post.)
Is this good? That is a difficult question. It isn’t hard to whip up a slippery-slope argument. It also isn’t hard to say that Trump’s social media accounts present a truly unique problem at this time, which justifies a unique response. But is that special pleading? It’s easy to argue the importance of free speech, just as it’s easy to ask what the true goal of enlightenment principles were, and whether slavish adherence to them will get us where we want to go. But is ditching principles for two weeks going to weaken them? Should it? Reasonable people can disagree (not that all of the people disagreeing are reasonable).
I’m interested in a different question, though. How did we find ourselves grappling with this problem in the first place? Much as I’d like it to be Donald Trump’s fault, it really isn’t. This is an inevitable result of how the Web works 2021: walled gardens and closed protocols have concentrated informational power in the hands of a few companies, companies whose every action affects the structure of our press and our democracy. As a wise man once said, “Every system is perfectly designed to get the results it gets.”
So, what’s to be done? I don’t know, but I do know one way we could have done it better. So, if you’re interested, join me below the fold for a discussion of distributed social networks.
What is a social network? In a nutshell, it’s just people, forming connections, posting content, and reacting to said content. We know these mostly as centralized systems like Twitter, where a single organization controls all of the content and user accounts. It is completely predictable that such an organization, of sufficient size, will find itself in a constant, morally fraught fight against extremism, harassment, and illegal content. This architecture will always have a serious problem via Masnick’s Impossibility Theorem: “Content moderation at scale is impossible to do well.”
So what is a distributed social network? Let’s talk about the most popular one: Mastodon. Mastodon is, essentially, a series of communication protocols that are woven together to link decentralized communities.
One protocol you might know is HyperText Transfer Protocol, or HTTP. Anybody can set up an HTTP server, and if you send it the right text, it sends you a web page, or whatever. You can link websites together, form informational networks, and so on, very easily with this protocol. This is what the Web is, or was, back in the day: a mostly-open collection of linked documents encoded in HTML, an open standard. We would later call this Web 1.0, and it was very different from the walled gardens of today.
Mastodon occupies a space somewhere between Twitter and Web 1.0. Anybody can set up a Mastodon instance; think of it as their own personal Twitter. Mastodon instances can form voluntary connections with one another, and all instances, user accounts, and posts (“toots”) share an open data format and communication protocol.
The key word here is interoperability. It’s a little like Reddit, where people run subreddits with their own rules, except in this case, there would be nobody who actually runs Reddit–it’s like a community of linked subreddits.
Here’s a toy example. Let’s say I run a Balloon-Juice Mastodon instance, and Scott Lemieux runs a Lawyers, Guns, & Money Mastodon instance. We both have public timelines, and can post replies to each other, etc.
Now let’s say that the people from some troll or Nazi instance discover ours, and we want to get rid of them. We can block individual users or even their entire instance, preventing them from reading (and therefore replying to) our posts. Their Mastodon account is still around–it’s just formatted data, after all–it just won’t do them any good here.
In a system like this, you deal with a Donald Trump not by deleting his Mastodon account–which you cannot do–but by banning that account from the respectable instances. He’s still welcome at any instance that does choose to deal with him, which in this case would be something like Parler (or for a real example, Gab, which is a Mastodon fork). We’ve reduced the moral footprint of this decision dramatically, even as we’ve achieved a similar result.
I hold little hope that we will end up in an open, decentralized informational utopia any time soon–it’s just too hard to monetize, compared to the alternatives–but a man can dream.
Update: I just remembered that Twitter is funding a small, dedicated team to explore this space and either develop a new standard or bring an existing one to the next level. So, good for them!