15-Year-Old Security Argument: AI Escapes Come From Superincompetence, Not Superintelligence

dscape · x · 2026-09-24

At a Lisbon AI event, security veteran Diogo Mónica reiterated his long-held thesis: when AI 'escapes,' the culprit is usually superincompetence — engineering and security failures — not superintelligence. A well-received masterclass on AI cybersecurity wrapped in a Pirates of the Caribbean joke.

Original post →

More from Safety

Safety channel →