experiment log
Agentic Experiments
Building Core Warden with AI agents. Tracking what the machines do — and what they break.
This week
PRs merged—velocity
Agent success—of agent-run issues
Cognitive load—complexity index
AI cost—cumulative spend
This week's learning
First dispatch coming Saturday.