Generating code got easy. Making it last didn't.

In mathematics nothing is real until you prove it; in code the proof is tests. With an agent writing the code we need more testing and static analysis, not less — and tests good enough to be worth reading.

Model benchmark tables and one-shot demos are vendor-run, Goodharted, and blind to your harness. The only benchmark that predicts your work is your work.

Convert PDFs to clean Markdown locally on Apple Silicon with marker, so LLMs can ingest documents efficiently without rasterising every page.

Transcribe video locally on Apple Silicon with whisper.cpp at 5–8x realtime — no API key, no per-minute charges, no audio leaving the laptop.

An opinionated set of agent skills (and the Artifex CLI they lean on) that ships work from a Jira ticket to a deployed fix without you at the keyboard.

A practical guide to turning plans into durable specs that humans and agents can review, approve, and implement.

On skills as the missing layer in agentic coding — encoding tribal knowledge, composing with commands, and keeping the context window focused.

A practical guide to agentic coding — context windows, tools, modes, prompting, system prompts, skills, spec-driven development, and warnings.