About the publication
A desk, not a stack.
Why this site exists and how it relates to its sibling.
editorialcontrol.org is a publication about building with AI agents. Not the models. Not the benchmarks. The day-to-day texture of doing work when the thing on the other side of the prompt can draft, review, and occasionally surprise you.
It exists because I kept writing the same kind of post on audiocontrol.org — my other site, which is about vintage samplers and retro-computing — and the posts were drifting. Vintage hardware readers showed up for a SCSI teardown and got an essay on editorial calendars. That's on me. Two different audiences, two different publications.
So this is the split. audiocontrol.org keeps writing about samplers, samplers, samplers. editorialcontrol.org takes the how I got any of that shipped thread and runs with it on its own page.
What counts as a post here
A post earns its way in if it's about using an agent to do real work. That means specific sessions, specific friction, specific wins. Not frameworks. Not general advice. Not "here are 7 ways to prompt better."
The editorial calendar that runs this publication is itself a piece of the publication. You'll see posts about how the calendar was built, what broke the first time we ran it end-to-end, which skill-invocations the agent got wrong and why. The receipts stay on the page.
How it's made
Astro static site, two builds from one repo. Every post is drafted by an agent
(/editorial-draft) and finished by a human. The editorial calendar
lives at docs/editorial-calendar-editorialcontrol.md, versioned with
the code. Analytics, distribution, and cross-link audits are scripts you can read
the same way you read the posts.
The sibling site, audiocontrol.org, runs the same machinery. When this site writes about workflow, audiocontrol is usually the lab where the workflow was proven out.
Who
Orion Letizi. I build tools, I ship words, and I let the model do the first pass on both. You can write to me at orion@audiocontrol.org. If you have something the publication should be writing about, that's the way in.