Notes on thinking, debugging, and building software with an AI duck in the loop.
It’s funny. I’ve been watching the gen AI space for a couple of years now. I’ve dabbled with a dozen apps, built a few integrations using APIs for work, and experimented with various tools to improve my coding and project management workflows. I also quietly…
Continue reading...From Claude's insight report after my first week Key pattern: You are a tenacious, hands-on debugger who actively corrects and redirects Claude through repeated configuration struggles, prioritizing getting tooling right over application development. Sounds way worse than it was, really. I did have quite a bit…
Continue reading...After about a week of tinkering with Claude wired into my Elixir/Phoenix workflow, I’ve stopped asking whether AI coding assistants “work.” They do. The more interesting question has been whether my workflow works. Most of the work I had to do to get all this…
Continue reading...After getting Claude set up, I spent some time exploring how it handled an unfamiliar codebase. I asked it to read through things and explain what it thought was going on. Mostly calibration. I wanted to see how it reasoned, not how fast it answered.
Continue reading...Day one of really diving into getting Claude set up as one of the cores of my AI development workflow, and somehow my workspace, my workflow, and my project feedback loop all got permanently better in a single afternoon. I’m 43. I’ve been doing this…
Continue reading...