I Hate the Modern Internet. I'm Building My Way Out.
Soo I have a secret. Well, actually I have two secrets. One I’ve held for almost twenty years and the other I’ve kept for the last five months.
-
I HATE social media. If you know me, I guess that’s not all that surprising. I’ve hated social media since I left MySpace to make a Facebook page back in 2008. Now that hasn’t stopped me from joining every other platform that’s launched since, but I still hate it.
-
For the last five months I’ve been vibe coding my way through a series of web app builds with the goal of replacing at least one of these accounts. No longer will I be part of a platform that collects my data so some bro in Silicon Valley can sell it to advertisers. No longer will I be sequestered to four websites and fed “content” from people’s “profiles” that some “algorithm” thinks I like.
I’m done. Well, mostly.
How I Got Here
The sad truth about the modern digital landscape is that you almost have to use social media. *eye roll*
Lately I’ve found myself having more fun on Reddit and, embarrassingly, TikTok than any other platform, which genuinely surprised me. I was the guy who avoided Reddit in the 2010s because I had it filed away under “weirdos and contrarians complaining in forum after forum.” I was wrong…mostly. There are some posts that still make me want to question if we actually are in a simulation, and that’s with a well-managed feed. That said, once I found my chosen subs, the entire experience settled into a wall of gaming news, anime recommendations, fantasy tropes and Black Twitter refugees. I was and am happy here, most days. Honestly, Reddit is brilliant and I’m mad I’m so late to the party.
And that’s where the first aha landed.
Forum-based platforms are closer to the real open internet than anything built to serve us videos, photos or microblogs.
At its essence, Reddit is a collection of thoughts, opinions, posts, and reposts from across the wide web, organized into fun little rooms. Its orientation is outward. The other platforms are oriented inward, toward themselves.
Reddit reminds me of the internet I fell in love with back when I was twelve or thirteen.
It was the early aughts and the internet was a BLAST. Each log on brought a new experience. I’d open Internet Explorer (later Firefox) and just see where the day took me. Online minigames. Looking up song lyrics. Asking Jeeves about a homework assignment and ending up on some random Encyclopedia Britannica page. Downloading a new version of a peer-to-peer file sharing app that shall not be named to maybe acquire music to burn a new CD. Talking to friends on instant messenger.
Using the internet from 2002 to 2005 was an adventure. And a lot of antivirus.
And I fell hard for MySpace. I mean head over heels. I had what felt like my own website. My own little corner of the world wide web. I could change the layout and background, put a song on the page, make anything that I could get generated into HTML. It was a vibe but it was a platform. No slight to Tom or the dev team, but it was a platform that plateaued and then Facebook came and ate its lunch. Except they didn’t just eat MySpace’s lunch, they threw it out, kept the souvenir lunchbox and filled it with Uncrustables and saltine crackers.
That’s my whole frustration with social media. The container looks the same but everything inside was replaced with something mass produced and flavorless. Not to mention the more pernicious truth that by using these platforms almost exclusively we’re all basically locked behind a walled garden (or several). Sure, it’s a garden the size of a theme park, but it’s still somebody else’s promised neverland.
I came to social media because I thought I was carving out my own corner of the net. First, it was the wrong move. Second, the platforms were always in control, and of course they flipped the script and locked us in.
I hate it here and I miss the open web.
So I’m building my escape hatch.
Back in March I got curious about what I could do with AI so I gave myself a homework assignment. I was to build a custom learning application that would teach me more about AI, the basics of blockchain / distributed ledgers, and to breakdown quantum computing in terms I could understand. This wasn’t for anyone else but me. It didn’t take more than one session with Claude to build a web app (Evolution, cheesy I know) that ran locally on my Mac. However, somewhere between building the application then using it something shifted. I stopped being someone who just read about people using AI to build apps and services and became one of those people. That ended up mattering way more than the content of the learning application. I learned a process and the seed was planted.
I used the application for maybe two weeks before I realized that I’d spent more time playing with the configuration than completing the learning journey. The seed was growing. Then in April while on my morning Reddit scroll I stumbled upon (rip) a subreddit filled with people building cyberdecks. Handmade, portable computers. Some single-purpose, some general purpose. Mostly built with Raspberry Pis with a few exceptions. Each deck was an expression of their creator. It was like a wand or a wizard’s staff. And I wanted one. I ordered my first Pi, the Pi 5, that week. I thought I was going to build a cyberdeck. What actually happened was I booted the Pi and found myself asking: what can I do with this? So I asked Claude.
Before I knew it I had migrated Evolution from my Mac to the Pi. It ran there locally, and it worked! Then I started wondering, how do I access this version from Mac and my phone. So I asked Claude and we set up Tailscale so I could access the application and the Pi, that I named Earthseed, from anywhere. From work. From a hotel in New Orleans. From my phone over cellular. Anywhere. And it worked! After maybe a week of building and deploying I decided that if I was going to run infrastructure, I should actually understand and be able to explain what I built. So I added a module to the Evolution learning journey focused on networking and the physical layer.
Then it happened again, I kept tinkering and playing with the shape of the thing because honestly, building and shipping web apps was easy. And fun. Sure they were vibe coded and lacked the sophistication and distinction of a premium web service, but it was all mine. Then the tinkering got away from me, but in a happy accident kind of way.
Here’s what I learned through this process: the first thing you build may take a week, the second will take an afternoon. Once I realized that, my mind started spinning with possibilities. Every idea I had was shippable.
- I stood up an Open WebUI that ran local models so I’d have AI at the house. That meant I had a system that didn’t phone home to OpenAI, Anthropic, or Google.
- I built an agent I named Chiron to help me plan projects.
- I set up a ROOT dashboard which was supposed to be a clean front door for the two web apps and Chiron’s interface.
- Then I built a vault so my data could be encrypted and held solely by me.
- Then I built what was meant to be a public facing learning tool.
- Then Mission Control to track my progress through the Evolution learning app I built that started this whole thing.
- ROOT eventually evolved into something I called Lazana and intended to be an operating system with agents running all over the place.
- Then I started building a runtime environment underneath Lazana to host the agents.
Then I hit several walls.
The truth is that there was no real reason for any of it. No one was asking me to build any of this. I just started doing it because I could and I thought it was what I wanted. Right now there are 13 services running on my Raspberry Pi and nine apps that I wrote to sit on top of them. There’s a 14th, Mission Control, that has been crash-looping since a system update back in June. It has tried to restart itself more than 116,000 times. Nothing has been recorded within it since April. I didn’t notice either error for months because I never actually used the thing. I built it, then abandoned it.
Five months, a lot of ambition and if I’m honest sheer will. In that time I learned exactly what I’m capable of as a one-man dev and design team armed with a few clever agents. A whole lot but not as much as I envisioned. And that’s okay. It’s probably the best lesson.
This past June something else happened. The screen for my cyberdeck broke. It wasn’t that major and it was bound to happen considering the deck essentially lived in a fanny pack with no real housing unit. However that small, stupid mistake made me stop climbing the walls I hit. I finally asked myself the question I had been outrunning and deploying service after service around since April.
What is all of this even for?
Out of everything I shipped, the only things that another human could actually engage with were two websites:
- sece.dev; where I document what I’m building and try to explain it in terms a normal person can follow.
- xophersean.com; my own little corner of the internet. My portfolio where I share all of my work and post things like this. I built this to replace my LinkedIn.
Everything else I built can be considered scaffolding. Real scaffolding that’s still useful. My websites are the building they support.
What I’ve Got Now
For twenty years, I’ve been convincing myself that I liked living in the walled gardens of Meta, Google, Apple, Amazon and even Netflix. I’ve pretended that the relationship I’ve negotiated with them makes sense and that we are both beneficiaries. Now I’m building my own infra to support the sites and services I want to run so I don’t have to depend on the Silicon Valley cartel to enjoy the power of the internet. I invite you to visit sece.dev and xophersean.com and check out the philosophy I’m laying out about how we survive the changes of the 21st century. As a hint, there’s a reason I named my Raspberry Pi Earthseed.
Also published at Middle Path Lab.