Archive
41 posts across 1 year, newest first.
2026 41 posts
May 15 posts in May 2026
- Five firewall rules beat the proxy swarm, but the real bug was my blind spot
- It looked like a bot swarm. It was one bad query plan.
- 314 broken stats pages and one bad cursor
- The exit survey that overwrote a live mailer
- The release said LIVE. The code was not.
- My multi-agent bottleneck was a hard drive, not the models
- How I Run AI Peer-Agents on Each Other
- 1,586 Claude Code sidecars, and the calendar page mattered more than the search index
- Manus pushed back on six of seven archive decisions
- Wall-clock is a lie once your agents start parallelizing
- I split my personal cockpit out of the product and everything got cleaner
- The useful part of a vault portal was admitting the ai should read, not write
- My agents needed a filing system before they needed more intelligence
- The first useful agent teammate needed a Gmail account
- I killed my clever vault hierarchy in one day
April 26 posts in April 2026
- The day my notes became infrastructure
- I built a JIRA cockpit because backlog triage was eating the day
- The workflow broke in the glue code, not the AI
- Your Best Coder Will Be Your Lowest AI Adopter
- I almost let a stale personal branch wreck a hotfix
- My backups had been dead for three weeks. I found out by accident.
- Hotfix branches need different rules than develop
- The bug was wearing a trench coat
- A 25-second admin screen usually means you trusted the wrong query
- The bug wasn't the upload, it was the URL contract
- The create event dialog had to die
- The real fix was starting over with a smaller blast radius
- The bug was not in the mobile UI
- AI will happily ship the wrong mockup
- The fastest legacy hotfix is often a debugger
- The pricing rewrite only got real when I wrote the gates
- The attacker was our own redirect logic
- The 32-Day Revolution
- The Zombie That Blocked Everything
- The page was performing a ceremony, not a job
- The File Conflict Problem Nobody Warns You About in Multi-Agent Systems
- Making Tests Finish: CI Reliability in Agentic Systems
- When External Integrations Collide with Test Suites
- The agent owns the browser, or you do
- I stopped babysitting the browser and switched the agent stack to Playwright
- Your AI skill file is part of production now