On 11 December 2019, Twitter CEO Jack Dorsey announced, via the social media platform itself, that Twitter would be funding an independent team of researchers, to develop an open standard for social media, of which Twitter would ultimately seek to become a client.
Twitter is funding a small independent team of up to five open source architects, engineers, and designers to develop an open and decentralized standard for social media. The goal is for Twitter to ultimately be a client of this standard. 🧵
— jack 🌍🌏🌎 (@jack) December 11, 2019
Confused? Compare email with Facebook: email runs on an open source protocol called SMTP. Although most people access the protocol through tools provided by Google, Microsoft, Yahoo, etc., these tools are merely clients, with no control over how the protocol itself works. Your Hotmail address can send an email to a Yahoo address, and Google has no say over how the email itself can be used. Facebook, on the other hand, is closed. You cannot send a message to a Facebook account, or make a Facebook post, from any other social media service. Likewise, Facebook has total control over what can be done on its platform. Twitter is currently like Facebook, but Dorsey’s announcement indicates an intention to transition to something more like email. In an attempt to evoke progress and promise, the project will be called Blue Sky.
This all generated a minor buzz—mostly confined to Twitter itself. There was little commentary in either the technology or financial media. A week or so later, everybody seemed to have moved on. Given the widely acknowledged importance of Twitter and the high profile of Dorsey, I found this surprising. This announcement is probably a harbinger of a dramatic transition in how Internet services are produced and consumed, and might well be a pivotal moment in the history of technology. It has potential short-term implications for free speech online—and wider implications for technological progress in general.
One of the most common responses to the announcement was to point to ActivityPub, an existing open source protocol for social media that seems to involve exactly what Dorsey was describing. In a reply tweet, Dorsey acknowledged that the team may choose an existing standard, if deemed better than creating one from scratch.
The amount of work already invested in ActivityPub’s ecosystem gives some indication of what Twitter might be aiming for. With over three millions users, Mastodon is the most popular implementation of the ActivityPub protocol. This statistic is more complicated than it might seem at first, however, as Mastodon isn’t a discrete service:
Mastodon encourages users to set up their own instances by forking the open source code of an existing instance, amending it to reflect their own preferences for an online community, including if and how to allow communication with other instances, and self-hosting. Much of the early interest in Mastodon was driven by a 2017 Vice article by Sarah Jeong, who describes the initial design inspiration as a “kinder, nicer, version of Twitter … without Nazis.” Jeong writes that mastodon.social, the original and now third largest instance administered by Mastodon’s creator Eugen Rochko, explicitly bans “content illegal in Germany and/or France, such as holocaust denial or Nazi symbolism.” Twitter does not necessarily ban such content, so mastodon.social may be preferable to users who value this.
The genius in Rochko’s overall design is not to try to take on Twitter directly, by presuming to have stumbled upon the perfect community guidelines, but to allow users the flexibility to encourage or discourage whatever they want for themselves—without forcing them to build an accompanying communications medium from scratch. They neither have to petition a central arbiter to enforce their preferences, nor silo themselves in protest. As Jeong writes, “a bannable offense in one instance might be completely acceptable in another.” Perhaps presciently, she also comments that, “for now the instances are on friendly terms with each other, coexisting peacefully in federation.”
Two years later, chaos ensued. After years of quiet growth, Mastodon was thrust back into the tech spotlight in 2019, when social networking site Gab announced its intention to transition to a Mastodon instance. Gab alleged that the decentralized nature of Mastodon would allow it to overcome the fact that it has been blacklisted by Silicon Valley infrastructure providers ranging from web host GoDaddy to payments processor PayPal. Depending on who you ask, Gab is either a haven for alt-right trolls or a bastion of free speech. Gab describes itself as First Amendment absolutist, and will only take down content that is illegal in the US. This obviously contradicts the policies of mastodon.social but, crucially, not those of Mastodon and ActivityPub at large.
Whether you think Gab provides a networked space immune from monocultural corporate thought policing, or that it is disingenuously hiding behind free speech to harbour and even encourage openly hateful content, is irrelevant to the more fundamental issue. Within ActivityPub, you are free to make and act on decisions for yourself. You can join an instance with the community guidelines you most desire, or even start your own. As Gab has controversially demonstrated, this does not involve the Herculean effort of bootstrapping a centralized Twitter rival, because all the instances and the underlying communication protocol are open source.
This model of content moderation and dispute resolution is decidedly superior to Twitter’s in almost every respect. A few months into Gab’s move, Mike Masnick wrote in TechDirt that,
the fact that most of the major Mastodon instances have said they don’t want to federate or deal with Gab is an expression of preferences, and in many ways a better overall system than one in which a single company is making the decisions. And it’s much better than politicians telling companies what they need to do.
Users no longer have to vie for zero-sum administrative power, but can build a boundless, uncensored positive-sum federation. This is localism in the space of virtual geography.
Many critics of Dorsey’s announcement have argued that Twitter is abandoning its responsibility to effectively moderate its content. Such arguments entirely miss the point: they reflect a view of the Internet that technological progress is increasingly making redundant.
As I have discussed elsewhere in this magazine, social media platforms tend to become vast monopolies due to the network effects inherent in the service provided. At the time many emerged, the most effective mechanism for funding their maintenance and improvement was monetization via targeted advertising, gleaned from centralized hosting of user data, and served up among the content consumed. As Sarah Jeong puts it, “social media feels like the hellish extreme of late capitalism, Faustian bargains where consumers consume themselves.”
This commercial reality clearly clashes with social media’s desired use as a communications medium in multiple ways. The monetization model of historically unparalleled surveillance is generating increasing pushback, and tolerance of the unilateral norm enforcement this model inadvertently requires is also waning. Gab was started because its creators perceived Twitter as overly prone to censorship; Mastodon deemed Twitter not prone enough. There was no way to settle this disagreement beyond going their own way and creating their own spaces. That Gab and Mastodon exist is due to the logical impossibility of Twitter’s accommodating both desires.
If Twitter pushes content moderation towards the edges of the network that would be a very good thing. It would validate the important work of ActivityPub, Mastodon and free and open source software as a whole. That it is even possible for a central arbiter, as opposed to a publisher, to moderate content is probably a temporary quirk in the history of technology. Post offices, telcos and SMTP clients do not moderate the content of letters, phone calls and emails, and this may not be the case for mainstream social media for much longer either.
This would be a potentially enormous win for online free speech. Not only in the limited sense of being able to say whatever you want, but in the broader sense of wanting to be part of a community in which everybody can opt in, to hear what is being said. As Christopher Lemmer Webber, one of the co-authors of ActivityPub, has explained, “it’s not censorship to decide not to listen to someone. It is censorship to not allow someone to create an avenue to be able to speak … The complement to freedom of speech is the freedom to filter.”
Mastodon creator Eugen Rochko has articulated a more sophisticated objection to the proposed move: “the Embrace, Extend, Extinguish problem—where a huge company adopts an open source protocol, becomes its biggest user, and then changes the system to lock everybody else out.” This is a concern, but I am hopeful that it will not happen because this is usually the result of a desire to bring a majority of users of an open standard onto a proprietary one. In this case, the reverse is happening: the open standard is considered superior on account of its openness, and the proprietary provider will bring to it many more users. Despite his reservations, Rochko feels that, “Twitter adopting ActivityPub would be a good thing and a victory for the web.” I agree.
As the Internet comes to permeate more and more of our lives, the issue of who really controls its resources will become increasingly pertinent. As Alex Gladstein of the Human Rights Foundation notes in his essay “How Bitcoin Can Protect Free Speech in the Digital Age,” “each time you make a digital transaction, your online footprint grows bigger, more telling, and more profitable.” The dominant business model of social media directly incentivises highly profitable surveillance, while simultaneously owning and controlling the content users generate. One of the subtler implications of Alex’s argument is that free speech online is about more than edgy memes or block lists: it is a test of freedom of every kind.
As our behaviour becomes increasingly mediated by networked software, there will be little to distinguish freedom of speech from freedom of association, freedom of contract, the right to privacy, the right to be free from unreasonable search and seizure, and just about every other freedom we take for granted. If speech and money become the same strings of bytes, then they may as well be strings of bytes we can reliably control.
Twitter is very much alive to this. In his announcement, Dorsey cited an article by Mike Masnick on the contrast between protocols and platforms. Masnick writes,
the response many will have to the idea that the platforms could do this to themselves is to ask why they would do this, since it would inevitably mean getting rid of the monopolistic control they currently enjoy over the information in their system and allowing that data to return to the control of the end user and be used on competing services using the same protocol. However, there are a few reasons to think that some platforms might actually be willing to accept this trade-off.
The broader ecosystem of social media in the 2010s may well represent the high-water mark of corporate control over individual freedom. As Dorsey himself says, “new technologies have emerged to make a decentralised approach more viable.”
But I suspect Twitter wants to go further than becoming a Mastodon instance. I think it will try to leverage the work many have put into developing a network and a reputation to provide an option for self-sovereign identities to enable effective coordination of any and all online activity. This may mean tweeting—or whatever verb comes to mean broadcasting between self-hosted servers over ActivityPub—it may involve sending and receiving bitcoin payments, or more exotic digital assets or contracts built on Bitcoin or in its image. Twitter will try to become a kind of personal DNS of the next generation of the Internet. It may completely fail. We may not need corporate mediation of this kind of behaviour at all, but that Twitter will at least attempt this tells us a good deal about where the puck is going.
I understand the frustration of ActivityPub and Mastodon developers who feel slighted by the way the announcement was made. They have been diligently working on this problem for years. But my claim is not that Twitter will make important strides here. My claim is that the fact that Twitter is moving in this direction at all is the important stride. It’s a symbol, a nod to the flag of technology and the future of free speech online, and arguably even of freedom itself. A blue sky vision indeed.
Disappointing to see the comic strip included here (for visual interest?) makes such an egregious misstatement on free speech. The banners and cancelers are not the full audience listeners. They look to deny the opportunity of others to listen. That is an abrogation of free speech.
That’s part of it, but I think it’s even broader than that. It ignores the fact that free speech can’t ever simply be about formal rules and laws but has to reflect a broader ethical understanding of how society operates and the importance of divergent views being expressed. Under the argument the cartoon makes you could have a scorched earth culture war of cancelling, no platforming and attempts to shut people down and say free speech isn’t under attack because I’m under no obligation to grant anyone a platform. You literally speaking might not be, but if enough people lose the understanding of why allowing divergent views even those you may find repugnant to be expressed is important we’re in real trouble. Certainly making sure what have become our privately owned town squares remain places open to expression of views is a huge part of that.
All this means to me is that more work will be foisted upon end users to curate and monitor their experiences. And then some startup will come along and say “Hey, we know this protocol thing is all complicated and tough to understand. If you let us have your data so we can sell it and target ads at you, we can help you keep track of your friends across all these different instances.”
Create a problem then market the solution; it’s the Silicon Valley way.
Sarah Jeong complaining about Nazis is ironic.
I realise it’s not much to do with the article, but it pisses me off the extent to which that XKCD cartoon has come to be seen as some kind of profound statement on free speech.
It isn’t, it should take people about five seconds to see how deeply flawed the argument is but it keeps getting trotted out like it’s some slam dunk to anyone concerned about the principle of free speech being undermined in the current climate.