Posts

I finally wrote the material I kept telling people I’d get around to. Omnist now has a proper introduction deck and a hands-on tutorial. The deck is language-agnostic — it explains the spec itself, not any one port’s API. The tutorial uses Python to walk through every model and feature with real, working code. Why a deck and a tutorial, not just docs The specification is precise on purpose. Precise documents are a bad place to start. You shouldn’t have to read a spec before you know if something is worth your time. The deck is the fifteen-minute version: what problem Omnist solves, and why the model looks the way it does. The tutorial is for someone who’s already decided to try it. Every example is real, working code, checked and run before it went in the page. It’s built as steps, so you can do them out of order once you know your way around. ...

A library gets proven by people using it. A specification is a harder claim to back up: someone else has to build from your rules alone, without being able to ask you what you meant. I wanted Omnist to become a real document and schema standard, not just a library that happened to do well. So I planned for that early on: get the design out of one language, and check that it actually holds up in several others, before I called any of it a standard. ...

Two posts ago, R and D were two roles, two rulebooks, one person switching hats. Last time was D’s story: the loop that made shipping safe. By now the two don’t split cleanly — every decision and every build happen in the same conversation. This round adds a third: QA, an external auditor with no stake in the code already being good, brought in specifically to check R and D’s work from the outside. ...

Everyone has discovered that AI writes code fast. The harder question is when you can trust what it wrote — especially if you weren’t watching while it worked. This post is the playbook I use for that: which tasks go to which AI agents, when work can run in parallel, and why the agent that builds something never gets to declare it done. It’s battle-tested. Twice, this workflow caught AI confidently misreporting its own work — both incidents are below, receipts included. ...

Last time, I split the two weeks of building Omnist into two teams of one: a Research Assistant for conceptualization, a DevOps team for materialization. R and D ran together the whole time. Every cycle was the same: decide, build, decide again. These next two posts split that loop by role, not by time. This one is R — three decisions, and why they went the way they did, before each spec became D’s problem to build. Next time is D: the same three decisions, from the side that answered how. ...