Hacker News
Mark Nottingham: What's Missing in the 'Agentic' Story Is a Well-Defined User Agent Role Backed by Public Standards
IETF/W3C veteran Mark Nottingham argues AI agents lack the accountability framework web browsers provide through standards. Without standardized permission models, sandboxing, and behavioral expectations, users must place extraordinary trust in opaque systems with no verifiability. Proposes developing AI user agent roles through public consensus processes (like W3C) rather than proprietary platform decisions — constraining agent capabilities in 'normal' ways so people can depend on how they behave. 63-comment HN discussion engaged seriously with the standards governance angle.
↳ Follow the thread