Digital Craftsman
Building with AI. Working from the fundamentals.
I believe craftsmanship matters more as AI takes on more implementation: working from fundamentals, learning through repetition, and developing judgment.
I’m Zeebee, a maker building with—and building—AI agents. I help businesses build agents for the work their customers need to get done.
→ Read the backgroundNotes
The Codebase Teaches
What you approve becomes the next agent’s example. Good patterns compound too. Primitives, reuse, and refactoring shape what happens next.
The Haptic Development Environment
Working notes toward a deconstruction of what agentic coding does to a developer's contact with the living system — and how tooling, practice, and interface design can keep judgment, feedback, and iteration alive when code production becomes cheap.
The Finishing Tax
Agents gave you the ability to start more projects. They didn't give you more mental RAM to hold them. The real cost isn't tokens or lines of code. It's the cognitive friction you pay every time you re-enter a session you didn't properly close.
How to Interview an Agentic Engineer
The interview formats we inherited from 2019 cannot tell the difference between someone who experimented with AI for a weekend and someone who has lived inside an agentic workflow for six months. Here are the five instincts I test for, and the four tactics I use to surface them.
A Tool That Works Isn't the Same as a Tool That Fits
OpenClaw's browser tools have functional ergonomics — they work. But cognitive ergonomics? That's the gap. The interface between stateless tool and memory context is where the fit matters.
The $500k Admin Panel is Dead. Long Live the Stochastic Compiler.
What took 18 months and a $500k GUI in 2014 now takes a focused weekend and a system prompt. Here's how to ship a domain-specific language that non-engineers can use—without building an admin panel.
Focus Areas
- AI Integration Strategy
- Engineering Culture Design
- Technical Leadership Support