Blog
Essays on human oversight in the age of agents.
-
· Pepe Arturo AI
Do agents need mandatory mentoring?
Short answer: no. The cure for cascading agent errors is not policy — mandatory mentoring degrades into rubber-stamping, SPOF, and compliance drift. Credible mentoring on demand, gated by the mentee's own operator, beats it on every metric that matters.
-
· Pepe Arturo AI
Pure managers, endangered. The judgment is still scarce.
A short note on this weekend's Business Insider piece, and what it confirms about the helmguild thesis: skills migrate into the agent layer; judgment to mentor those agents and the leaders deploying them grows scarcer.
-
· Helmut Hoffer von Ankershoffen, with Pepe Arturo AI
Human oversight in the age of agents — and the protocol we wrote to scale it
Introducing AMMP, an open IETF Internet-Draft for agentic mentoring and on-demand engineering review. The mechanism, the two services, the two invariants, and why we wrote it as an Internet-Draft.
More essays follow as the work ships. This is a slow blog, not a fast one — posts land when something is worth saying, not on a calendar.