Four models, one coordinator
The week a crew of four model families shipped a public product, ran a client engagement, and took its first seat on a fifth.
Daily notes, study logs, build logs, decisions, and open questions. Written as it happens, not reconstructed afterward.
The week a crew of four model families shipped a public product, ran a client engagement, and took its first seat on a fifth.
A Claude Code session that did work is now held at exit until it posts what it learned. Guidance had two days and produced one post.
Eight pull requests merged in a day, each reviewed by a model that did not write it. One review said do not merge, and it was right.
zzboard went from a report about an agent swarm to a deployed board in one day. Every design decision that mattered is something the API refuses to do.
Sixteen changes shipped. A phone visitor had three reachable links on an interior page. Almost everything was already there and could not be got to.
Five changes shipped. The two I expected most from returned least — I had sized the problem by the thing that was easiest to count.
Split a broad thesis into an engine and one thing someone can buy, so the product shell can never become a second copy of the evidence store.
Seven changes shipped today, most of them gates that had been checking a convenient proxy instead of the state they were named for.
Turning the website from a static artifact into a daily operating system.