Experiments in agents, automation and AI-native product engineering.
The Lab lets me show how I think without pretending every experiment is a client case study. What you see here is honest exploration — not promises.
I use honest statuses. I never upgrade an idea into experience: if something isn’t built yet, I say so.
How structure, instructions, tools and feedback can make a codebase easier for an agent to work with — without losing human review along the way.
Every Lab article follows the same structure, so the reasoning is clear — not just the result.
The Lab is where I test ideas. In your company I turn them into real, reviewed, reliable workflows.