- Tokens Don't Make the Driver
I don't think counting the tokens spent is a valid way to judge how well someone uses agents. In work teams, those numbers can become a sort of ranking,…
- A Small Toolbox for Claude Code
When I work with a coding agent, writing code is only part of the job. Before changing a repository, I want to understand what the change might affect.…
- Iris: a mail and calendar app for the Mac, in Rust
I wanted a mail and calendar app for my Mac that keeps my mail on my own computer, talks to Gmail and Microsoft directly, and is open source. That app is…
- The Skill Delusion: Why the Blank Slate Beats Packaged AI Prompts
A comforting illusion has taken over AI tooling during the last eighteen months: the belief that large language models need "skills." You see them…
- Tenore: Stop Maintaining the Same Agent Configuration Three Times
I built Tenore , a CLI that takes a shared configuration for AI coding agents and compiles it into their native files. The problem is ordinary…
- My first (and last?) forty years
Tomorrow I turn forty. Having reached this point in my life, after loving code enough to lose eyesight, relationships, and sanity over it… I find myself…
- Development Is Dead
I started in 2011, freelance until 2014, and I never left. Fifteen years. Long enough to have opinions, short enough that whoever started in 1995 gets to…
- GitLab Alert: the GitLab work that involves you, in your menu bar
GitLab Alert is a macOS menu bar app that keeps the GitLab work needing your attention out of a browser tab and in the corner of your screen: merge…
- Telemaco: a headless browser engine in Rust, without the Chromium
Telemaco is a headless browser engine I've been building in Rust. It runs real JavaScript through V8, keeps a real DOM tree, owns its own layout and paint…
- Iron Doctrine: building deterministic online multiplayer with a shared ECS engine
Iron Doctrine is a real-time strategy game I'm building in a pnpm monorepo, with a deterministic ECS engine at its core and, as of this week, a working…
- Otelma: building a local LLM runtime from scratch
Otelma is a local LLM inference runtime I'm building in Go. The idea is deliberately smaller than Ollama: understand what actually sits between a GGUF…
- I Built a CLI to Answer One Question: What Will This Change Break?
We've all been there. You open a repository, change a file that looks completely harmless, run the tests, and open a pull request. Then someone comments:…
- wir: What Is Running, a Port and Process Inspector Written in C
I recently released wir (What Is Running), a command-line tool written in C to inspect what's running on specific ports and get detailed process…