2026-06-06
Week 0 — Why I'm building in public
AI is a thing, that’s for sure. In order to practice using it, I want to make a game. I love games and always played them with my family. At the time, playing games together just seemed normal, like part of life. Looking back, it was what brought us all together, creating experiences that bonded us together.
I know that’s sappy but I just turned 50 :-). Anyway, I thought this was the perfect time to use AI to try and build a few of my own, just to see what it can do.
I can’t promise any of them will be good, but hope to share more in the coming weeks, and would love to get your feedback on this journey with me.
The game
The idea isn’t new — it’s borrowed from a game we all played back in the late 1990s called BattleMail. Same concept: a challenge would land in your inbox, you’d set your moves, then watch the results play out. I wanted to do a similar thing — but with stick figures, plus a few things I love from Fortnite: custom emotes, and the ability to play levels you build yourself.
What’s working, and the stack
The basic game is already up and running — you can challenge a friend and watch the match play out. I wired it together on Hostinger, a basic host I already had (it runs my son’s music site). It uses a Postgres database — but instead of Hostinger’s MySQL, I’m on Neon, a free database, for now. The plan is to move everything to AWS once I scale; right now I’m deliberately in a cheap, simple proof-of-concept setup. Don’t over-engineer day one.
A dev shop of one — run with AI
Even solo, I run this like a real dev shop. Every change starts as a ticket in a free Jira instance — and my AI updates, comments on, and closes those tickets through MCP, following my SOP. The code goes through GitHub the same way: AI creates a branch, opens a pull request, and I review and merge. Hostinger auto-deploys on every commit, so shipping is effortless. All of it, with AI as my co-builder, for $20 a month.
What it took
All of this was about 10 hours in the first week. Not all of it was smooth — getting my dev environment running (WSL on Windows) came with plenty of gotchas, enough to make me consider switching to a Mac more than once. But that’s a story for its own post.