As someone who frequents HackerNews and various other forums, I often see a lot of discussion about privacy, data security and self-hosting. The discussion almost always leads to someone shunning all centralised services that most people use, and pitching Mastodon, Matrix, XMPP or something no one’s ever heard of but that one user is sure is the best decentralised service there is.
However, no matter how good these tools are, I’ve never truly been convinced they’re at a stage where I’d want to use them, or be comfortable recommending them to friends and family. Here are some of my thoughts on why I personally think these services have a long way to go.
Privacy doesn't sell
If you’re someone who’s concerned about privacy, but your friends aren’t, claiming privacy isn’t a good enough selling point for most people. Privacy and Openness are great selling points when choosing a new service, now more than ever. However, they can’t be the only selling point. I’ve come across many a service that promises the highest level of security and privacy but when you use them, you realise why. There’s nothing in the service worth selling even if someone was interested. Many services die out because they focus so much on staying private, they don’t focus on features that would appeal to people who don’t care about privacy, i.e a good majority of the world. Many decentralised services like Social Media or Messaging rely on a network effect, and personally, I wouldn’t invite people to join a service that is noticeably worse in most areas than services that exist. If someone’s regularly using Telegram, it would be tough to convince them to use Matrix in the form that it is currently. Not just does it have fewer features, most importantly it has fewer people. And it has fewer people cause it can’t compete on features just yet.
No Free Lunch
I use Mastodon, a user on the Fosstodon instance. For a long time, I didn’t question how I was able to use it for free, and when I recently started thinking about it, I asked the maintainers what it costs them and how they pay for it. They replied saying it cost them $85 a month to run it on masto.host, and the money was crowdsourced with some users donating regularly. They always kept enough cash to keep them floating for the next 6 months, and the excess was donated to open-source projects they all agreed upon. While this was a really nice thing happening, it doesn’t seem like a scalable solution. Should the service grow to a scale where it can no longer be managed through donations, an instance of a decentralised service isn’t going to get VC funding unless they have some separate side plan to get paying users. Services like Mastodon don’t provide a business plan by default. Many of these services only make money through donations or by offering a cloud-hosted and managed service. Whatever service comes up next, I do hope they can come up with a plan where even other users hosting an instance can get the money needed for maintenance through the users.
The technical wall
One thing I’ve noticed in a lot of online forums, especially ones where talks of decentralisation happen, is how much many people seem to underestimate how technologically abled most of the world is. Services win when they make it easy for the average Joe to use it. Want a chat app? Download one, sign up and you’re ready to go. If someone who’s never heard of a decentralised chat app downloads one to check it out, and the first question it asks is which server do you want to connect to, there’s a good chance you’ve lost them there. Secondly, most people, even if they’re technically abled would find it tough to host and maintain a service. It might be easy to follow some installation steps or spin up a docker container with the service, but if you run into problems, it takes time and some knowledge to be able to debug it (even with the help of StackOverflow) and keep it running. This can’t be expected of everyone, and if people just use a popular instance, it brings us back to the previous problem of the cost of maintenance.
But is it all lost?
Obviously not. The best argument for federated services is the massive success of email. You have a plethora of mail services to choose from, some free, others paid. If you don't trust any of them, you can host your own service but have to deal with the pain of setting it up and getting through all the spam filters. Email isn't perfect even today. There's a lot of phishing and spam mails, but we've come a long way building services to detect the spam and filter it out. Today a lot of decentralised services are being used as "free speech" sites that harbour a lot of fake news, hate speech and more. However, there's scope to work on that and build the required tools to help fix that too. Decentralised services are definitely the way to go in the future, and will take a lot of work to build, maintain, and most importantly, educate and help onboard everyone irrespective of their technical skills.