Subscribe
One email when I publish — durable engineering practices for running AI coding agents in production. No hype, no news treadmill.
I write one long-form essay at a time on the engineering practices behind running AI coding agents in production: task design, context engineering, evals and verification, cost and failure modes, and how teams absorb agent-written code.
Subscribe and each new piece lands in your inbox the day it goes up. No news commentary, no promotional filler — just the durable judgment I’d want if I were building with these systems on a real team.
New here? Start with The State of AI Coding Agent Engineering — the synthesis behind everything I write, free by email. Every essay is also collected on the blog , and the primary research behind them lives in the research index .