Attackers now move at machine speed, while organizations still decide at committee speed — and that gap has become enterprise cybersecurity's biggest weakness. Title: The speed problem: How frontier AI exposes weakness in enterprise cybersecurity URL: The average time between a critical vulnerability's disclosure and its active exploitation has collapsed from roughly three weeks in 2025 to about 20 hours in 2026, and the authors argue this is an operating-model problem, not a technology one — proposing "governed autonomy," where AI handles routine, time-sensitive work while humans govern novel threats and material risk decisions. Highlights 🔍 Concentrate visibility on a "minimum viable organization" Instead of chasing real-time visibility across the whole enterprise, leading organizations first focus near-real-time monitoring and continuous scanning on the irreducible core — core banking for a bank, clinical systems for a hospital — whose disruption would be existential. 🛡️ Build defensive AI as one agentified, end-to-end workflow Threat hunting, vulnerability remediation, and incident response get designed as a single connected system rather than assembled from separate vendor tools, with AI agents pre-authorized to take bounded actions like isolating systems within defined guardrails. ⏱️ Centralize decision authority in one person A cross-functional response cell with a single CSO-delegated decision-maker can override standard change management for confirmed active threats. One financial institution adopted this after a critical vulnerability sat unpatched for 11 days because no one had the authority to override a change freeze — remediation times fell sharply afterward. It's a compelling argument that what separates winners and losers here isn't security budget size, but the clarity of who can decide what, how fast, and under what authority. #Cybersecurity# #AIGovernance#