_Last updated: April 2026._
Work
- Shipping the True Platform launch. Katalon's re-positioning from an automation tool into an AI-native quality platform. Public window opened April 7 and we're sustaining the release rhythm week over week.
- Leading the Kai agent team. Kai is our orchestrator agent — the piece that holds product intent and routes work to the right capability (test generator, runner, insights, bug reporter). Architect, lead, engineers, QE.
- Building toward agentic QA trust. Not "self-healing scripts" — the actual shift: let the agent decide what to verify, make its uncertainty legible, keep the human as the safety net.
Writing
- A long essay on harness engineering for QA teams — borrowing OpenAI's February framing and Peter Pang's April essay to argue for what the QE job becomes when agents do the execution.
- A second piece on running product from the terminal — what six months of the personal-agent experiment actually taught me.
Reading
- The Mom Test — rereading. The questions still hit.
- Working in Public — Nadia Asparouhova. For the open-source governance subplot.
- Forrester's AI-Augmented QA 2026 report. Agreeing with the problem, pushing back on the solution.
Experiments
- A personal briefing system that reads across Slack, Jira, GitHub, and meeting transcripts each morning and surfaces the two or three things I actually need to decide. Rough, but shockingly useful.
- Publishing this site (you're on it). Writing in public so I stop repeating myself.
Not doing
- Not taking on new advisory gigs right now.
- Not chasing conference circuits — one or two deliberate ones per year, at most.
- Not reopening side projects until the Katalon launch is stable.