Archive
Notes worth keeping.
What Survived After I Tore My System Apart
How dismantling ARCH led to TAKT — and why the most useful architecture was what survived the collapse.
AI Is a Tool, Not a Competitor
The most useful way I have found to work with AI is not to compete with it, but to use it to make better engineering loops cheaper.
When the Best Plan Is Not to Build
Engineering judgement starts before implementation: some of the best outcomes come from changing the decision, not executing the plan.
Why I Care About Friction
The best engineering work is often not about doing more. It is about removing the things that should never have been necessary.
How I Think About Software
I am less interested in writing more code than in making the system around the code easier to understand, change, and live with.
A Small Place on the Internet
A personal notebook for ideas that are worth keeping.