Newsletter

Occasional notes on intelligent systems and human responsibility.

Short notes when something is worth saying—usually before it becomes a longer essay. AI governance, judgment under automation, and the gap between what systems can do and what humans authorize.

Latest

Capability Is Not Authority

15 August 2026 · Issue

AI systems can draft, send, and complete tasks. That capability is not permission. On the Marshall Principle, four states of action, and why “the AI said it’s done” is not evidence.

Read the note →

From the archive

Subscribe

Leave an email if you want these notes when they appear. Low volume. No drama. Unsubscribe anytime by writing me.

Or email richard@marshall.net with subject “Newsletter”.

Constitution

The free public policy that these notes sit under:

AI Governance Policy → · Marshall AI Governance Standard → · magrs.org →