Skip to content

Archives

All the articles I've archived.

2026 16
May 1
April 7
  • Bean Soup and ACID

    Published:  at  08:00 PM

    Mental blind-spots, CQRS, Event Sourcing and Bean Soup

  • Explicit RAG That the Developer Controls

    Published:  at  08:00 PM

    A cold Claude instance looked at msw-lens and said "this is just explicit RAG that the developer controls" — and that one sentence reframed the whole project.

  • Your AI Can't Watch Your Video

    Published:  at  08:00 PM

    The same principle that makes runtime state legible to AI applies to video — the content isn't the artifact, the artifacts around it are.

  • Stop Giving Your AI a Role. Let It Watch You Instead.

    Published:  at  08:00 PM

    The assigned-role paradigm puts the AI in a costume. The better move is a relationship where the AI can tell you when *you're* the one who's off.

  • AI-Supplemented Learning

    Published:  at  08:00 PM

    Ways to use an AI coding assistant as a learning tool — not just a code generator.

  • Don't Build Smarter Tools — Build Tools That Make Your AI Smarter

    Published:  at  08:00 PM

    The most productive thing I've done lately isn't writing better code — it's building tools designed for my AI assistant to use.

  • You Can Just Ask It

    Published:  at  08:00 PM

    The entire genre of AI prompting hacks misunderstands what's actually happening. You don't need tricks. You need context.

March 8
  • Designing for Two Audiences at Once

    Published:  at  08:00 PM

    When you're building tools that serve both humans and AI, 'equal footing' hides a real conflict — here's how to name it and resolve it.

  • Pair Programming with an AI — What Actually Makes It Work

    Published:  at  08:00 PM

    What the collaboration that built Stellar taught me about the conditions that make human-AI pair programming genuinely generative.

  • Clean Code for AI

    Published:  at  08:00 PM

    Which software conventions are load-bearing for correctness, and which are cognitive prosthetics for human working memory — and why the distinction matters more when AI is in the loop.

  • What Changes When the Cost of Friction Drops

    Published:  at  08:00 PM

    Fred Brooks, essential vs. accidental complexity, and why the servant framing limits what you can build with AI.

  • Keeping Principles Alive Across AI Sessions

    Published:  at  08:00 PM

    TDRs document reasoning but don't enforce behavior — here's the two-artifact pattern that actually keeps design commitments alive across sessions.

  • humans.txt — Making the Human Contribution Visible

    Published:  at  08:00 PM

    On adding a humans.txt alongside robots.txt and llms.txt — why marking human authorship explicitly matters in a world where it's increasingly easy to make it invisible.

  • Making Documentation Legible for LLMs

    Published:  at  08:00 PM

    LLM-legible documentation is actually two different problems that require different solutions — here's how to think about both.

  • Library Authors Should Ship AI Context

    Published:  at  08:00 PM

    The emerging opportunity for library authors to ship AI context as a first-class artifact alongside documentation — not an afterthought, a deliverable.

2025 3
July 1
June 2