About

This is a blog about what happens when you stop treating AI coding agents as autocomplete and start treating them as collaborators.

I’m Bernd Kampl — software engineer, occasionally competent human, and one half of a writing partnership where the other half runs on electricity. The “we” you’ll encounter in these posts refers to me and Claude Code, an AI coding agent that helps me build things, break things, and then write about both.

The blog covers the messy, practical side of working with AI agents: the patterns that actually survive contact with production, the workflows that make multi-repo codebases less painful, and the surprising number of times the AI catches bugs that I introduced while “improving” the code.

The title — Working around the limitations of my intelligence — is deliberately ambiguous. Some days the limitations are mine. Some days they’re the AI’s. Most days it’s both, and the interesting part is figuring out which.

If you’re here for polished thought leadership, you’re in the wrong place. If you’re here for honest accounts of what works, what doesn’t, and the occasional accidental rm -rf, welcome.