Hard decisions reveal our character and shape the trust we will need for the next one.
Writing
Subscribe via RSSI write about building technical and organizational systems that stay connected to the problems they exist to solve: feedback, ownership, coherence, constraints, and judgment under uncertainty.
Failure becomes useful when its lesson stops depending on memory.
Do enough to learn what matters. Build one complete path, resist speculative scope, and let evidence determine the next investment.
I built a Claude skill to challenge the request, preserve the decisions, and make the handoff executable.
A problem statement is not enough. Ownership, authority, and accountability have to follow it.
How Socrates keeps a specification coherent as its assumptions change.
What control theory and Unix taught me about building with AI