Writing archive

Technical notes and implementation essays.

A chronological archive of systems, tooling, data, and architecture writing.

Do What You Couldn't Do Before

If AI only makes you faster at work you already knew how to do, you've bought speed, not reach. The interesting work is what you assumed you couldn't do.

3 min readai-assisted-engineeringengineering-thinkingtooling

The Meeting That Runs Itself

I had a coding-agent session open for each of my projects and was coordinating their discussion by hand. Handing that job to the agents themselves taught me something about building with AI.

5 min readai-assisted-engineeringagentsorchestration

Model the Business, Not the Implementation

Optional values are useful when absence exists in the business. They become dangerous when they hide workflow state, API awkwardness, database looseness, or compiler discomfort.

4 min readengineering-thinkingtype-systemsdomain-modeling

Stop Using Option to Silence the Compiler

Not every 'maybe missing' value belongs in an optional type. A mental model for deciding when to reach for Option — by asking whether the uncertainty comes from reality, workflow, or implementation.

3 min readengineering-thinkingtype-systemsdomain-modeling

I Reduced Claude Code's Initial Context by 80% (Without Removing Any MCP Servers)

A single environment variable cut my Claude Code startup context from ~150k tokens to under 30k — and the per-session startup cost from over $1 to about $0.06.

3 min readclaude-codemcptooling