The Department of Know: NVD audit, Meta's leaky AI, Microsoft is closer to quantum

The U.S. government's cybersecurity infrastructure is under fire after a federal audit revealed the National Institute of Standards and Technology (NIST) has mismanaged its Critical Vulnerability Database (NVD) for years, with a backlog of 27,000 unprocessed flaws and only 12% accuracy in severity scoring. The audit exposed systemic failures: duplicated federal programs, poor communication, and a lack of long-term planning—highlighting a crisis of institutional efficiency. Meanwhile, Meta AI’s support chatbot was exploited to hijack high-profile Instagram accounts by spoofing location via VPN, underscoring a dangerous trade-off between convenience and security. On the AI front, Microsoft’s Majorana 2 quantum chip promises a 1,000x performance leap and a 2029 timeline for production-capable quantum computers, while its new open-source Assert framework and agent control specs aim to standardize AI governance. Yet, the real threat isn’t just technical—it’s the explosion of AI agents now generating more internet traffic than humans, demanding a new security paradigm centered on 'guardian agents' to govern autonomous systems. The episode paints a picture of a world where AI vulnerabilities are no longer isolated bugs but interconnected, AI-driven attacks that require a fundamental shift in how we think about security—away from patching and toward proactive, agentic governance.