About

I started learning to code in February 2022. I hadn’t written a line of code since college, and decided Harvard’s CS50 was the place to begin. I spent about three and a half months doing every exercise, every lecture, every lab. Then I got so absorbed in the final project that I never actually finished the course. All I had to do was upload a 1-minute YouTube video but I couldn’t be bothered with it was too busy writing code.

I’ve been building things ever since.

Three years later, my GitHub has 29 repositories—framework hopping through Next.js and React and RedwoodJS and Svelte before landing on Flutter and Serverpod. Lately I’ve been experimenting with python AI agents and Rust MCP tools.

Somewhere along the way, I noticed that pair coding with AI is its own kind of coding. Not “the AI codes for me”—genuine collaboration where something emerges that neither of us would build alone. The meta layer is recursive: systems that improve how you build systems. That recursion is what this blog explores.

What you’ll find here: the actual process of learning to build software through AI collaboration. Real commits, real pivots, real moments where nothing works. I’m documenting it as a lab notebook, not a highlight reel. If you’re curious about what AI-augmented development actually looks like for someone who started from zero, this might be useful.

When I’m not coding, I’m probably with family, eating, sleeping, swimming, bouldering, or watching my stock positions do things I didn’t expect.

You can find the code on GitHub or follow along here as I figure out what I’m doing.