Would I Have Typed This?
AI answers are verbose because the model needs its own words to keep working - not because the reader does.
I write to think - about software architecture, agentic engineering, and how knowledge moves through organizations. Nothing here claims to be right; everything here is an invitation to disagree.
AI answers are verbose because the model needs its own words to keep working - not because the reader does.
Humans silently compensate for bad knowledge distribution. Agents can't - and that changes what documentation has to do.
When every architecture decision becomes a harness decision
The missing link in harness engineering