Distributed publishing in the age of corporate censorship
The suspension of Chuck C. Johnson from Twitter has prompted a lot of discussion about alternatives based in protocol, and I had a thought on the subject.
Everyone became sick of Facebook a long time ago. Feeds loaded up with garbage and ads that shove how much they're stalking you right in your face. Facebook aggressively censors any content it's founders don't like, such as nationalist politics and pro-Palestinian activism. Now that Twitter is a public company and all the good people have left, it seems that it has been similarly hijacked by partisan actors who are giving the boot to many of the people that made it such a compelling service. Twitter is just starting to be despised and they haven't even monetized the service yet. Just because a company is despised, however, does not mean it will fail. Comcast's legendarily high margins attest to that. Right now two companies have a monopoly on how we communicate with one another. Unlike Comcast the oligopoly of social media is made only of software and consumer attention. These things can be crafted a lot easier than new cables can be laid.
Ello recently received a lot of interest simply on the platform of not being Facebook. Most stayed away, realizing that it's the same familiar business model controlled by the same familiar parties. Nobody wants to help build a service that'll be more of the same, especially when it lacks the features and social graph people are used to. Regardless, the fleeting immigration to Ello shows that there is immense consumer interest in competition to established social media services. The rapid emigration from Ello shows that immense consumer interest is really only going to be sustained in the form of distributed publication and decentralized control, not corporate ownership.
In 2010 there was an attempt by a small team of CS undergrads to create a distributed social media service. It was called Diaspora. They raised a six figure sum on Kickstarter (it was the second biggest crowdfund of all time, to show again that massive consumer interest) to make the idea happen. It was an utter failure for a number of reasons. I knew as soon as I looked at their code and saw ImageMagick handling all the image processing that it was going to be a failure. Distributed systems are extremely difficult undertakings. They are not the kind of thing that you should do as a first major project while you're still in college. There are companies like 10gen that produce distributed systems as their first product, but they are undertakings entirely of their own. You don't see distributed systems companies also becoming social media companies. Diaspora was predicated upon too many things becoming successful, and was doomed from the beginning by people whose vision exceeded their ability to implement that vision.
Those making a distributed competitor to existing social media immune to centralization and censorship must avoid incursion into difficult areas of computer science. Building a compelling user experience is a hard enough hurdle for a new company to jump without adding the most difficult areas of computer science to the obstacle course. There are a host of new and exciting technologies such as blockchains that one might envision social media being built upon, but I think the key technology here is much older. Most of those who are on the early Internet probably got their start posting upon the oldest distributed publishing system known to man: Usenet.
Usenet has operated reliably longer than I have been alive. Usenet already scales to a degree that far surpasses Twitter. The amount of text UGC (excluding the JSON syntax) in the Twitter firehose is a mere ~1.08 terabytes a day. Compare that to Usenet, which currently sits at around 18 terabytes of daily feed volume. Usenet is rife with content that violates the law because it is so difficult to censor. That being said, there's an essential thing that Usenet got wrong that allowed for the success of today's social media.
Usenet was born of academic discussion. It came of an age where all who had access to the service were the intellectual elite of their time. Like an ancient forum, where the philosopher-kings whose aptitudes had ascended them above concerns of daily drudgery could travel far distances to with the intent of discussing matters of eternal importance. Usenet was the world's greatest discussion forum until September of 1993, when AOL gave its subscribers access to the service. You see, Usenet's view has traditionally been by category. You would subscribe to an interest just like following a hashtag on Twitter, and thus see all the posts relevant to that subject. This was perfect when most posters were experts in their field. This was unsustainable when the average IQ of Usenet posters dropped by thirty points.
Modern social media sorts by the identity of a poster, not by category of content. This makes content bearable. It incentivizes participation as well. Everyone's motive for using social media is self-interest, be it for the purpose of hawking something, entertainment, or narcissism. Everyone is there to build a brand for themselves or get attention.
The underlying messaging infrastructure of Usenet is still fine. The only thing that prevents it from threatening its modern competitors is how the messages are displayed to clients. A service that allows people to claim a username tied to their usenet posting ID, presents a web profile with their posts in a Twitter-like view, and provides a hook into ejabberd for direct messaging is a very surmountable task, and there could be a real market opportunity in it.