Engineering intelligence feed — articles, AI-native notes, framework observations, and the occasional update.
A field survey — what I'm reaching for by default this year and why the defaults changed.
A practical guide to integrating AI into your workflow without overcomplicating things.
5 min read ↗Why I treat LLM evaluation infrastructure as a first-class system, not a notebook — and the four signals every AI pipeline should gate on.
8 min read ↗A pragmatic take on when double-entry / event-sourced models earn their complexity — and when they don't.
11 min read ↗State, command UX, and optimistic mutations for operator tools that real people use eight hours a day.
7 min read ↗TTFB numbers from moving a retail storefront to the edge, and the cache-mesh design that made it hold.
9 min read ↗The review loop I actually run now — agent drafts, I arbitrate. What got faster, what got worse.
5 min read ↗Fifteen years in, the highest-leverage thing I do is make change cheap. A taxonomy of safe refactors.
10 min read ↗When threads beat processes, and a benchmark that surprised me on a CPU-bound parsing job.
6 min read ↗