Notes & Thoughts
Hi, I'm Keith
Technical systems, mental models, incident response, and agentic coding. Notes on how we think about, build, and ship software.
Recent Notes
View all →
Cynefin for Technical Support Managers: Match the Response to the Situation
Cynefin helps support managers match their response to the situation: runbooks for clear problems, experts for complicated ones, experiments for complex ones, action for chaos.
A Runbook for When Your Brain Goes Offline
A structured outage runbook removes thinking under stress — five questions, a timecode rhythm, and clear roles to cut response time and keep customers informed.
Bad Agent or Bad System
Hanlon's Razor applied to ticket quality review — how to diagnose whether a low-scoring ticket is a skill gap, a process gap, or a tooling gap before you blame anyone, including in multi-agent tickets.
-
The New Technical Manager Manual Five essential frameworks every new technical manager should internalise — from giving feedback without ego to running incidents without chaos.5 min read
-
The Incident Manager's OS: OODA × RACI × Incident Commander Speed alone isn't enough. OODA gives you rhythm, RACI gives you role clarity, and the Incident Commander gives you decision ownership. Three layers that turn crisis response into something repeatable.7 min read
-
Your Incident Response is Stuck on Observe Most incident response teams are stuck in an endless Observe loop — gathering data without building shared understanding. Here's how John Boyd's OODA framework fixes it.6 min read