Writing / Quiet argument atlas
Essays arranged by the questions they keep open: how AI systems fail, how agent-assisted work is governed, and how production systems make trust visible.
3 thinking regions / 4 records / latest May 9, 2026
Lane index
Essays grouped by the recurring questions they are trying to work through.
Region 01
AI failures where confidence arrives before the evidence.
1 record
Region 02
Practical ways to scope, review, and verify work done with agents.
2 records
No. 01
Essay10 min read
As agents thin the human buffer between direction and finished work, the quality of the brief stops hiding. On leadership as the threshold decision before action and the ownership decision after consequence.
Opens
When agents execute the brief literally, the quality of direction becomes visible.
Why read
What the work of leadership looks like once execution is no longer the proxy for it.
No. 02
Process Note10 min read
The engineering workflow behind this website: control lanes, execution lanes, slice-based delivery, and what breaks when you let AI agents build production software.
Opens
Using fast AI agents without losing scope, memory, or review quality.
Why read
The control/execution workflow behind the site.
Region 03
Production systems that make time, source health, and trust visible.
1 record