Notebook
Writing about the system: one mechanism at a time, what broke, and what changed.
studio-log ·
I spent $21.10 to find out two of my own skills make the model worseAn eval harness over eight cases, three runs per arm. Two of the five skills I wrote scored below the no-skill baseline.
Two posts is where this starts. It will read thin for a while, which is accurate.
One email when something ships, or when a rule changes.
That is the entire trigger list. No weekly cadence I can't keep with a full-time job and two kids. Some months that is zero emails.