Experts at the HumanX conference urged companies to ignore agentic AI hacking fantasies. They argue that basic, unpatched vulnerabilities pose the primary risk to enterprises. Fixing legacy security holes outweighs preparing for hypothetical superintelligence threats.