
One Must Draw: Studying a Game I Love So I Could Build My Own
Seven weeks, about 3,600 commits, and a spec-driven AI harness later, my latest attempt at a co-op deckbuilder finally had a baseline I trusted, heroes of its own, and a working title.
Side projects, game dev experiments, and lessons from 20 years of building things.
Explore the project timeline
Seven weeks, about 3,600 commits, and a spec-driven AI harness later, my latest attempt at a co-op deckbuilder finally had a baseline I trusted, heroes of its own, and a working title.
A spotlight that hunts for hiding Pokémon, a coloring app for the Steam Deck, and a phone-based monster drawing party game. Small games for my kids taught me more about controllers and playtesting than I expected.
A two-day Bloons-style tower defense turned into a three-week co-op roguelite deckbuilder, and the biggest lesson was that the combat I actually wanted came from Slay the Spire, not the games I grew up with.
In January I built a chat-driven second brain that filed my notes into Obsidian. Three days later I found an assistant with a better front door, turned my work into a skill for it, and then built the assistant a kanban board.
Seven years after going door-to-door with property tax appeals, I pulled the old data out of a backup, pointed it at a different customer, and planned and built an MVP for tax attorneys in a single day.

2,189 commits in Q3 2025 alone. Pokémon Vampire Survivors with my kids. A web agency hitting $1000/month in recurring revenue. When AI meets a tinkerer's mindset, everything accelerates.