Ben SandbergI was lying in bed at night, scrolling through Reddit because my brain wouldn’t shut up, when I saw a post about something called “chat control.” I had no idea what that meant, but the comments made my stomach drop.
The EU wants to scan all our private messages. Even the encrypted ones. They want to scan everything before it gets encrypted, right on your device. For the children, they say. But the privacy experts in the comments kept calling it what it really is: mass surveillance with a PR team.
I couldn’t stop reading. Every article made it worse. These politicians were seriously discussing whether I have a right to send a message that only my intended recipient can read. Like this was some reasonable policy debate instead of a fundamental attack on human communication.
I’ve always been the kind of person who gets completely consumed when something pisses me off this much. I couldn’t think about anything else for weeks.
The angry realization about everything
What really got to me wasn’t just the chat control thing. It was suddenly seeing how I’d been complicit in my own digital exploitation for years.
I’d been passively scrolling through Facebook and Instagram since 2012, posting maybe one photo in all that time. Every time I opened these apps, I felt vaguely gross afterward, like I’d eaten too much junk food. But I kept using them because that’s where everyone was.
Now I was reading about how Meta deliberately designs their algorithms to make people angry because angry people engage more. How they’ve known for years that Instagram makes teenage girls feel terrible about themselves. How they collect data about people who don’t even have accounts.
I started feeling genuinely stupid for feeding my personal life into these machines that see me as a product to be optimized for revenue.
The worst part was feeling completely powerless. What was I supposed to do, delete my accounts and become a digital hermit? Send angry emails to politicians who probably don’t even understand how email works?
Finding out there was another internet hiding in plain sight
Somewhere in my angry research spiral, I stumbled across this thing called the “fediverse.” It sounded like sci-fi nonsense, but I kept reading.
Apparently there are thousands of independent social media servers that can all talk to each other using an open protocol. Millions of users across all these servers, but instead of everyone being trapped on Twitter or Facebook, you can join any server you want and still interact with people on all the other servers.
It’s like email, but for social media. Your Gmail account can send messages to someone with a Yahoo account. Same idea, but for following people and sharing posts.
The biggest fediverse platform is called Mastodon, and it looks a lot like Twitter but without a single company controlling everything. Each server has its own rules, its own culture, its own person running it. If you don’t like your server, you can move to a different one and take your followers with you.
I made an account on mastodon.social just to see what would happen.
This felt different in a way I couldn’t articulate
Within maybe a week, I realized I was actually participating instead of just consuming. People seemed to be having real conversations instead of performing for an algorithm.
Someone would share a technical problem they were debugging, and three people from different servers would jump in with actual helpful suggestions. When people disagreed about something, the conversations still felt human-scale instead of turning into mob warfare.
It reminded me of what the internet felt like in 2005, before everything got optimized for engagement metrics. But now I could interact with millions of people across thousands of servers while still feeling like I was part of actual communities.
But I kept thinking: why am I trusting some random mastodon.social administrator with my data? If the whole point is decentralization, why not run my own server?
I’d been tinkering with a home server setup for months, completely obsessing over converting everything to Kubernetes. Not because I needed Kubernetes. I was basically using it to run Jellyfin and a few other services that would work fine on a regular Linux box with docker-compose. But I hate not understanding things that seem important, and Kubernetes kept coming up everywhere in discussions.
So I’d spent three months rebuilding my entire homelab with six virtual nodes running this stripped-down Linux that you can only manage through APIs. Everything automated, everything configured as code. Probably ridiculous overkill, but that’s how my brain works. The goal was if my server explodes, I can just bootstrap the cluster again and everything comes back.
Suddenly that obsessive infrastructure project looked like the foundation for my own small rebellion against corporate social media.
The moment I decided to build my own corner of the internet
I couldn’t get the idea out of my head. If I ran my own Mastodon server, I would control the rules. When the next surveillance law gets proposed, I could tell them to go to hell and deal with the consequences myself.
Maybe I’d get in legal trouble eventually, but at least it would be my choice. I could moderate discussions the way I think they should be moderated. I could decide which other servers to connect with. No corporation could arbitrarily change the algorithm or shut down my community.
This wasn’t just about learning new technology. This was about using infrastructure as political resistance.
The beautiful thing about the fediverse is that it’s designed for this kind of thinking. There are art servers, programming servers, regional servers, servers for specific languages. Some have strict moderation, others are basically lawless. The diversity is the point. But they all connect to the same network of millions of users.
I registered goingdark.social and started figuring out how to actually do this.
When nothing works like the tutorials say it will
Of course, nothing worked on the first try. I tried to deploy Mastodon using Kubernetes instead of the Docker Compose setup that every tutorial recommends, because I’m apparently incapable of doing things the easy way.
Mastodon is architecturally complex in ways the documentation doesn’t prepare you for. You need the main web app, a streaming API, multiple background job queues, Redis, PostgreSQL, and Elasticsearch all working together. Each component scales differently and fails in its own special way.
I spent a whole weekend debugging why posts from other servers weren’t showing up, only to discover that the scheduler queue had crashed and I hadn’t noticed. If that one queue goes down, most of the federation stops working entirely.
Then I tried to run everything in my homelab and realized it was destroying the performance of my other services. My Jellyfin pod kept stuttering during movies because Mastodon was chewing through CPU processing federation traffic. My home assistant was unresponsive from time to time. Another issue I found was that I now had made my homelab production critical. If my internet goes down, then my users suffer downtime.
I had to swallow my pride and move the whole thing to cloud servers. So much for complete self-hosting independence. But sometimes you have to compromise between ideological purity and actually making things work.
Building something that nobody asked for
The monthly hosting costs ended up around 50 euros, which felt significant for running a social media server for myself and maybe a few friends. But it was the first time in years I’d spent money on something that aligned with my values instead of just being convenient.
I set up automatic scaling systems, monitoring dashboards, automated backups. The system could theoretically handle anywhere from 20 users to 2000 users without me touching anything. Complete overkill for the handful of people I expected to join, but I wanted to be ready.
The weird thing was how quickly we found each other. I posted about the instance a few places, describing it as “for homelab tinkerers, privacy advocates, and developers taking control of their digital lives,” and within a couple weeks we had 24 active users.
These weren’t random internet people. They were folks running their own DNS servers, flashing custom firmware, setting up their own VPNs. People who don’t just complain about digital surveillance but actually do something about it.
What it feels like to own your own piece of the internet
We’re still tiny compared to my goal of eventually reaching 2000 users. But our 24 users can interact with millions of people across thousands of other servers in the fediverse. Something about the conversations feels different from any social media I’ve used in years.
When someone shares their homelab setup, they’re actually trying to help people solve problems, not building their personal brand. When people disagree, it stays conversational instead of turning into performance art for the algorithm.
I’m still figuring out community moderation across servers with different policies. Still debugging occasional federation issues and search problems. Still learning how to grow without losing the authentic culture we’ve stumbled into.
But for the first time since maybe 2012, I’m participating in social media that feels genuine instead of performative. I’m not optimizing my posts for engagement metrics because there aren’t any metrics to game. I’m just talking to people who share similar interests and values.
The chat control proposals keep coming back
Those EU surveillance laws are still being debated in meetings. Politicians still don’t seem to understand or care about digital rights. Meta is still harvesting personal data to feed their advertising machine.
But I’m not just angry and helpless anymore. I’m running infrastructure that matches my values. When the next privacy-violating law gets proposed, I have my own platform to organize resistance from.
Maybe this sounds naive or idealistic. Maybe 24 users will always be 24 users. Maybe I’m spending too much money and effort on something that doesn’t matter in the bigger picture.
But at night when I can’t sleep, instead of rage-scrolling through corporate feeds designed to keep me angry, I’m checking in on conversations across the fediverse where I actually encounter the kind of authentic connection and understanding I never experienced on corporate platforms.
The fediverse works because it’s not trying to be one giant platform that makes everyone happy. It’s thousands of small platforms run by people who care about their communities. That feels like what the internet should be.
And if those 24 users grow into 200, or maybe even 2000, we’ll be ready for them. I really hope we do reach that goal, because this journey has been extremely liberating so far. I hope others will get to experience what I did.
Future steps will probably include needing another moderator if we grow. But I’ll deal with that when I need to.