This message is mainly for the leaders and enthusiasts of the broad-based movement toward decentralizing content, but especially social media. I’m not trying to start a new project or organization—after all, decentralization is what I am encouraging. I’m partly trying to start a conversation among individuals, to get them thinking and talking—but on a massive scale. But I’m also trying to inspire people to action, to come together and go the last mile to achieving robust and extremely widespread decentralization.
Larry Sanger is the chief information officer of Everipedia and a co-founder of Wikipedia.
I’d be championing decentralization, and I’d be up in arms about where the social media giants have been taking us, especially in recent years, even if I weren’t CIO of Everipedia, which is decentralizing encyclopedia writing. Like many of us, I’m incensed at Big Tech for their increasingly bold and arrogant incursions into both our privacy (which puts our information security at risk) and our free speech. As power has come to be concentrated in the hands of Big Tech corporations, they have increasingly posed a threat to our rights. So I’m impatient to see decentralization happen; only with the same decentralization on which the Internet itself is built can we hope to secure our rights to privacy, information security, and free speech.
I recently wrote a proposal on how to decentralize social media, and it got quite a bit of traction and discussion. The response amply underscored two facts: First, there is a huge amount of support, latent or explicit, for the idea of decentralizing social media; second, there are plenty of very smart people already at work on various aspects of this vision.
But both of these things have been true for a long time, and yet the vision hasn’t come together. That’s a problem. In their replies to me, several project representatives maintained that what I proposed has already been done. And while I am a great fan and supporter of their projects, the job clearly isn’t done: Social media isn’t decentralized yet. Part (but only part) of the source of the confusion here is about what I mean by “decentralized.”
What Is Decentralization?
An essential question that we should ask more is, What is decentralization, anyway? This thought came to me again and again as people responded to my original post with, “We’re already doing this on X,” or “You’ve perfectly described Y.” But if you say this system that you’re very excited about just is the decentralized web, then we might not mean the same thing by “decentralized.”
There are seven components to a fully decentralized, open social media network, as I’m using the term:
- 1. Open, common standards and protocols. There cannot be a decentralized social media network unless there are rules that are held in common among an arbitrarily large, open group of publishers and readers—for example, standards for types of content and protocols for transmitting and displaying it. The network is defined by these standards and protocols. Email is an apt example. By contrast, Facebook is a good example of a giant network that is centralized partly because it lacks an open standard. (I’ve quit Facebook permanently.)
- 2. Multiple publishers. A wide variety of (not just one or two) completely independent websites, apps, individuals, companies, organizations, etc., should be able to publish to the network. For example, the RSS-driven “blogosphere” extends well beyond any one blog publisher such as WordPress (I have a WP blog), Blogger, or Medium, etc. This eliminates the centralized “walled gardens” of corporations like Google, Microsoft, and Apple.
- 3. No central content repository. Not only should there be many publishers, there should not be any “master” database of the content—for example, no central database that all copies are expected to stay consistent with. Content should be either duplicated the same everywhere (as in the case of blockchains and Usenet) or else assembled on the fly from an arbitrarily large number of sources that one subscribes to (as in the case of RSS). That eliminates Twitter and Quora (another one I quit), among many. While Twitter has an API, it maintains the master copy of all tweets and will not serve tweets hosted elsewhere; although you can publish first elsewhere and make copies on Twitter, Twitter treats its copies as the canonical tweets. Similarly, Quora aims to be the closed and central repository of the best of our questions and answers. Surely the decent thing would be to support the inclusion of questions and answers found elsewhere. I can’t imagine Quora doing that, though. Can you?
- 4. Open to all publishers. There are no special requirements, beyond strictly technical requirements, for a publisher to use the network. Anyone who wants to set up a service that distributes microposts, pictures, videos, etc., that are published on the network can do so. Codification of living standards and protocols and technical direction by groups like ICANN and W3C is generally fine. This prevents any organization or association from taking editorial control; so there could not be any network-wide group of fact-checkers or moderators such as Facebook has assembled. It also prevents central coordination by a privileged group of publishers.
- 5. Multiple readers; equal access to the entire network. It should not matter which reader you use to view other people’s content, and it also should not matter where the content was published. You should be able to locate all the same types of public content on all (or many) of them, just as you can use any browser to locate anything on the open internet, and you can use any blog reader to read any RSS blog. This eliminates Medium (which I’ve also left). Medium, despite using publishing on RSS, doesn’t (as far as I know) allow its users to incorporate blog posts from outside of its own network, not without cohosting the posts on Medium. Despite being a public resource and volunteer-driven, they even require that you have an account just to read more than a few articles.
- 6. Open to all users. You should not need any specialized skills, and participation should not require any special payments or permissions. For example, the web is easy enough for almost anyone to use, and the only cost is the price of your internet connection. WordPress’s latest redesign, to make editing similar to Medium, is a good example. This April, Everipedia will be launching one of the easiest-to-use blockchain-based editorial tools. This opens the network beyond those who have special permissions, qualifications, or abilities. It eliminates subscription services, “pay-to-play” websites (like many Google services, which are built on user contributions) and blockchains, academic or industry groups, etc. It also, in my opinion, eliminates networks that ordinary users don’t have a chance of setting up. If you have to be a programmer to be able to participate in the system as an ordinary user, it’s not really decentralized for that reason alone. It is centralized, or focused, in the hands of geeks.
- 7. Individuals control their own content. You should be able to fully own and control the distribution of your own content, just as you control your email, your blog, or your website. The network should empower no one to block or censor it at the network layer (or only for technical reasons). The http and https protocols and the RSS standard are excellent examples. There simply should not exist any central authority that you must satisfy, other than DNS and web hosting companies, the means to pay them, and government regulations. Almost all websites apart from individual blogs fail this test, no matter how much they like to talk about how decentralized they are. Only a standard or protocol (and things made out of them, like blockchains) can credibly satisfy this requirement; only an entire network of websites, run by neutral, technical standards and protocols, can actually guarantee individual control. Of course, even extremely widespread adoption of such standards would not foreclose the possibility that Facebook, Twitter, and YouTube would continue to block certain publishers on their platforms. The difference I want to see is that such speech restrictions should not prevent others, who do want to view such blocked content, from being able to view it as part of the same network.
According to these requirements, there are various ways in which social media is not yet decentralized:
- The mere existence of some well-developed standard is not enough; it must actually be in use.
- It doesn’t suffice that some website is using and promoting a standard. Until several fully independent websites are doing so, it is not a robustly open standard.
- If certain necessary tools do not support the standard (by enabling content to be exported in a standards-based feed, by importing and incorporating content from different publishers, etc.), decentralization still has not happened.
- If the tools for participation are usable only by people with significant technical skill, that’s not robust or strong decentralization.
- More generally, until many more of the billions of people toiling in the centralized digital plantations of Big Tech have switched to decentralized social media, the job isn’t done. A relatively small network can be decentralized in a perfectly good sense, to be sure, but its availability does not mean that social media in general has been decentralized.
Why Is Decentralization so Important?
This is a rich philosophical question.
The essential reason to care about decentralization is freedom. For one thing, if social media is centralized, that means there is a concentration of power—of the twin powers of publishing and censorship—in the hands of a few. There is no way to ensure that this power will be exercised responsibly. If we want to participate in social media most effectively today, then we must put ourselves at the mercy of centralized authorities like Facebook and Twitter (or of relatively small groups like the volunteer editors of Wikipedia). Our right to publish, to speak freely to our followers, depends on the agreement of not only those followers, but also of a third party that assumes ultimate control of the entire proceedings.
A closely related but distinct reason lies in the more fundamental value of independence, or autonomy. The centralization of social media means we are dependent on its owners; if something happens to it—say, a company’s servers are attacked or hacked—then our content and personal information is at risk. Our dependency also means we must accept whatever terms (legal, social, and otherwise) they dictate, or leave. Of course, that means leaving our followers. Our relationship with our followers is ultimately dependent upon the permission of the central authority. This isn’t to say we wouldn’t have security risks if we hosted our own content. But we are simply not able to assume responsibility for our own information security and privacy and relationships and more, if our content (or our ability to deliver our content to our followers) is ultimately in the hands of some corporation or other authority.
That we should not trust central authorities with our freedom and autonomy has been amply demonstrated in recent years. Corporations like Facebook and Google have demonstrated that they are perfectly willing to sell our privacy to advertisers, and according to rules that they establish. Centralized data collection and management (as well as social media logins) creates honeypots for hackers, creating the opportunity for massive data breaches. It’s no wonder that so many people are increasingly worried about their data privacy and security. If all of those accounts were scattered around the internet, as websites, email, and other decentralized services are, massive data breaches would be less frequent. And while some of us are clamoring for ever-greater controls of speech, others clamor precisely to the contrary for freedom of speech, and thus freedom from paternalistic organizations that wield the power to control our speech.
But if others have been at work on decentralization (as they have), and haven’t succeeded yet (as they have not), why haven’t they?
Why Hasn’t Social Media Been Decentralized Yet?
The most incisive answer is that most of us, even most of us who live and work online, have complacently accepted the centralization of our social media activity in the hands of Facebook, Twitter, and a few others. I would include Wikipedia here, even though it isn’t often called “social media,” because it has centralized work on free encyclopedia articles in the hands of the relatively few people who are willing to work on Wikipedia.org. These organizations either stopped participating in, or never properly adopted, open standards that would have effectively made the founding website or service just one among many in a larger, encompassing network. Twitter, for example, shut down its support of standards even as the massive network effects were kicking in.
Also, our relationship to what we now call Big Tech has changed a great deal in the past decade. Ten years ago, MySpace was still bigger than Facebook; Facebook was still the new hotness. These companies were once the height of cool, and for many, they could do no wrong. That’s changed. Now they are enormous and powerful certainly, and impressive and useful maybe, but cool? Not so much. Also, the fact that we’ve been using the phrase “Big Tech” more and more (see Google Trends) in recent years is telling. Many of us have awakened to the failures by Facebook, Google, Microsoft, and Apple to protect our digital privacy and, with it, our digital security. The concentration of power in the hands of a few makes it possible for them to make outrageous decisions that would doom smaller companies. Many of us were confirmed in our suspicions of the growing arrogance of these companies when they admitted to increasing the amount of censorship on their platforms, and even openly colluded, and then started explaining quite seriously about how they are on a highly moral mission to shut down speech they dislike. Such arrogant dismissals of concerns about privacy and free speech are very far indeed from the liberal roots of the internet, in which concern about privacy and vigorous and sometimes harsh debate were the norms.
There’s also not a little concern about the unearned wealth of Big Tech’s oligarchs. It’s unearned because, although they did fantastic work in creating their platforms, they didn’t create the content or the network effects that made their companies so huge. That was the doing of (i.e., was constituted by) their users’ participation. They have quite literally exploited their users.
Now, finally, a lot of us are quitting, or thinking of quitting, or wondering how to quit. I don’t think there was sufficient will for this before. But there might be now.
There’s another sort of reason social media hasn’t been decentralized yet. Efforts in this direction so far—which have been substantial—have mainly, though unintentionally, been by and for geeks. That’s understandable. After all, to be sure, that’s where it has to begin. Geeks will always be the inventors and early adopters of cutting-edge technology. But it is more important than ever that we geeks bear in mind that we are developing our tech for people who aren’t geeks. After all, we’re talking about social media, which is meant to appeal to the masses. So excellent design, UX, and convenience features are not just nice-to-haves; they are absolute necessities. Otherwise, our friends and family who aren’t so technically inclined will be stuck with privacy- and speech-controlling stuff they are able to figure out.
Also, some in the open source community have a certain kind of geek pride, even geek snobbishness, about usability (i.e., the relative unimportance of it). There are certain skills that good programmers must learn, and so what seems perfectly usable to them is, for others, inconvenient at best and totally impenetrable at worst. Developers take pride in these skills, which, though understandable, has the perverse result that usability tends to take a back seat, as long as most of the users are their fellow developers. These observations, however, aren’t true of frontend developers—the people who specialize in the parts of systems that nontechnical users interact with. So I would like to ask the backend, devops, and network engineers of decentralized social media: If you aren’t doing so yet, will you please think about actively recruiting awesome frontend developers to your projects, and even project managers who know the product area? Frontend developers can help your project go the last but absolutely crucial distance to real usability by everyone.
A Development and Adoption Strategy
While some speak as if it’s too late, as if the massive power wielded by Big Tech is an unchangeable force of nature, I have seen absolutely no reason to think this is true. And the executives of these companies are fools if they think it’s true. The largest empires and corporations in history have fallen, and internet projects and startups have been especially evanescent.
It’s a canard to think we must accept the current configuration of social media and other tech giants. We can force enormous changes.
I’m not saying it’s easy, though. How can we get from here to there—from the existence of some developed standards and some small but operational projects to their mass adoption? I have come up with a to-do list:
- Figure out the standards. People at work on different content projects should convene large, diverse bodies of people to figure out social media standards (“living” standards a la HTML5) that answer to the habits and preferences of ordinary social media users. If these correspond to any existing standards, all the better. But the conversation must be broader than can be found in any one group, if the next step is to happen:
- Get standards broadly adopted. Secure mass adoption of the standards. Endorsements by internet investors, famous developers, and standards experts, as well as actual use by leading alternative social media apps will, in time, determine whether they’re adopted.
- Write awesome export/sync/storage tools that use those standards. Create or adapt extremely user friendly tools, such as browser plugins or desktop applications, for all major social media platforms (Facebook, Twitter, YouTube, Wikipedia, Instagram, LinkedIn, Pinterest), enabling ordinary users to easily and efficiently export, or better yet, to sync changes to, their content according to the standards. For purposes of mobile computing, it is important to get such tools added to Chromium and Firefox so nontechnical people, who don’t often install plugins, can use them more readily. Exporting data implies that there would be another data store somewhere, apart from the social media platforms; this could be a local machine, a cloud service, or a social media service that specializes in just such data storage. If export and sync are built into browsers (which might well be the best approach), just make sure that Google, which maintains Chromium, and the Mozilla Foundation, which maintains Firefox, don’t become the default repositories of personal data—because they might very well want to be. That would re-centralize the data. Instead, we should pressure browsers to give people lots of choices up front of where they want to store their social data (including locally).
- Create publishing and republishing tools. Whether or not they are incorporated into the aforementioned export/sync tools, we also need ways to help our friends find and use our feeds. Some work has been done here, in the form of various Twitter, Facebook, etc., competitors that support standards-based posts. But a lot of work needs to be done on UX and on incorporating content from across many sources. All the tools mentioned in the definition of decentralization above and in my earlier proposal are apropos. But we’re not there yet in terms of usability: It needs to be competitive with the likes of Facebook. That’s a tall order, but the open source community has created great things before. We can do this!
- Decentralized feed registries. It isn’t enough to have an assortment of “alternative” social media sites and open source projects. It wouldn’t even be enough to settle on a standard or protocol. If you really want to create a massive movement, there will also have to be search engines devoted to surfacing accounts or feeds to follow. Perhaps the standards developers will create a social media account registry, not unlike DNS (the website, or domain name, registry). The hope is that different social media “nameservers” will, like domain names, rapidly propagate across the internet, so it won’t matter too much which one you use.
- Make privacy guarantees part of the standards. Lack of privacy is one of the biggest complaints about Facebook and others. So if end-to-end encryption and other privacy technologies are built into the protocols, this will make a decentralized system much more attractive to users who care about privacy, who will naturally demand this feature.
- What else? I haven’t thought of everything, I’m sure.
Thoughts Toward a Manifesto and Revolt
So far, I have been discussing mainly the general technical requirements for decentralizing social media. But perhaps the biggest hurdle is not primarily a technical problem at all, but a social one—what academics call a coordination problem. If all your family, friends, acquaintances, and colleagues are on a social network, then fun or otherwise valuable social interactions are happening there. You might all agree that the social network has become awful and that it would be better if you could all move elsewhere. But unless you all do move elsewhere en masse, nobody has adequate incentive to move. If you do move, you’ll miss out. Interestingly, it isn’t unwillingness that is the problem. The problem, rather, is a lack of knowledge of others’ intentions, or a lack of agreement about plans—in short, a lack of coordination.
So if we’re going to decentralize social media and defeat the fear of missing out, we must strategize together about how to solve this coordination problem.
Let me make a two-part proposal. First, we should articulate a set of requirements for a decentralized social media system that we all can agree upon. This can be expressed in a manifesto or statement of principles. Second, we should promote a massive call to action.
As to the first part of the proposal, what I am convinced is necessary is a sort of virtual constitutional convention, but with the limited focus of enumerating the most basic principles we want a new, better social media system to follow. We should ask some distinguished internet thinkers and doers to collaborate on it and discuss its various points. When we have arrived at a rough agreement, we will release it and invite the general public to share, discuss, and sign their names to it. If we want the greatest number of signatories, the manifesto will have to be relatively short but extremely well crafted. It needs to be written and marketed in a way that permits an enormous show of support—not hundreds or thousands, but millions of signatories. Let us demonstrate to each other that we are absolutely on board with the idea of decentralization.
The effect I hope this will achieve is to light a fire under developers—who are, collectively, the essential linchpin of this revolt—to immediately start building the many sophisticated but easy-to-use tools (described above) that will help make our shared vision of decentralized social media into a reality. This demonstration of a willingness to abandon repressive social media companies if they do not change dramatically should also free up needed capital to pay for these tools. This is important, because, after all, we are talking about small companies and open source software projects going head-to-head with giant, wealthy corporations employing lots of the best internet developers in the world. We the people can do it—but it will require a lot of, indeed, coordination.
This leads me to the second part of my proposal. It isn’t actually enough that we, collectively, demonstrate a willingness to use a system of decentralized social media. I believe we should give people the opportunity to commit to such a system, to declare their intention to use these tools as they become available. In other words, a call to action should either be part of the manifesto or it should accompany the manifesto.
There are at least two creative and potentially powerful ideas we could try. We could message the plan for a massive social media strike for one or two days:
Please join the social media strike which is scheduled for xxx until
yyy. During this time, please do not either post on or otherwise use (even just to read) social media apps such as Facebook or Twitter.
Instead, you can use one of these tools [there would be a page with a
list of various desktop and mobile apps and browser extensions] that
will automatically post for you variants of text like “I am on strike
against [name of social media network].” We also encourage editing
your profile text and picture to show that you are participating in
The idea of a strike would demonstrate, in a dramatic way that no one on social media could ignore, just how much latent public support there has been for decentralized social media. If you combine a social media strike with a principled commitment to decentralizing social media, I think this could absolutely devastate Big Social Media as a whole. It would, potentially, be a historic event that would kick-start a desperately needed worldwide discussion of what social media should really be like. But it also seems to me that a strike implies unions, and unions can be another source of centralized power. If we strike for decentralization, it seems like an obvious contradiction to organize and centralize our power in order to do so. So I would rather have a decentralized strike, a grassroots or organic strike with no official, managing organization, if we decide to do that.
But perhaps the most compelling idea is this: Members of the group that drafted and initially signed this Declaration of Digital Independence will individually spearhead discussions of what the very best social media networks and standards are. We will create a way of polling a very large, diverse group of verified internet influencers about their top picks. Then we will ask everyone to descend, en masse, upon the top vote-getting networks to try them out and see what they would be like at scale. In other words, we will try to solve the coordination problem by explicitly coordinating some collective tryouts of various alternative social media services. To be considered, however, the services will have to have already made enormous and credible progress toward implementing the principles of the Declaration, and they will have to be on record as wholly endorsing those principles. We name no names at this time, but we are very well aware that there are some social media websites, apps, and projects that have implemented the principles of the Declaration. It is time to free up the resources and to build the user base needed for those projects to thrive and replace the old centralized web.
There’s strength in numbers, and the more that small startups and social media companies band together behind the idea of decentralized social media, the harder it will be for everyone—not just the social media giants, but ordinary users—to ignore what’s going on.
We can solve the coordination problem, and by solving it, we will also regain and preserve our rights to information privacy, security, and free speech. All we have to do is demonstrate to each other, and to the world as a whole, our unwavering intention to rebuild social media—and by extension the internet generally—in a more fully, robustly decentralized way.
More Great WIRED Stories