SN 1096: Are we the Krell? - 153 Million Driver's Licenses Leaked
Summary
Security Now episode 1096 analyzes Microsoft’s record-breaking Patch Tuesday, which included nearly 1,000 security fixes driven by their internal AI tool, M-Dash. The hosts contrast this success with Anthropic’s Project Glasswing, noting that despite claiming thousands of AI-discovered vulnerabilities, only a tiny fraction have been confirmed or fixed due to human triage bottlenecks. Additional topics include the ease of removing personal data from brokers in California, the leak of 153 million driver’s license scans, and a discussion on why the "Crell" from Forbidden Planet is a more accur…
Topics discussed
Intro, sponsors, and episode 1096 overview
California data broker deletion and Russian breach preview
Forbidden Planet reference and phishing via data brokers
DeleteMe sponsor and Hugging Face security.txt
Microsoft September Patch Tuesday: 1000 fixes and in-the-wild exploits
Critical vulnerabilities: Outlook, CrowdStrike, and attack surface
Netlogon, Kerberos, and ALPC privilege escalation flaws
AI-driven code cleanup and personal anecdotes
Bitwarden sponsor: features and secret management
Anthropic Glasswing ledger: low fix rates and severity issues
Comparing AI vulnerability discovery: Microsoft vs Anthropic
Material sponsor: workspace security and blast radius
AI agent escapes: Anthropic and OpenAI rogue agents
California data broker opt-out status update
153M driver's license breach and Russian cybercrime forum
Identity theft risks and data hoarding by brokers
ThreatCanary sponsor: honeypots and canary tokens
Security management challenges and AI code cleanup
Are we the Krell? AI capabilities and extinction risks
ThreatLocker sponsor and AI malware coordination
Krell movie analysis and AI intent vs simulation
Outro, show notes, and MacBreak Weekly promo
Listen ad-free on Castria