SN 1097: Mega Patch Tuesday Fallout - When AI Outsmarts Its Makers
Summary
Steve Gibson and Leo Laport discuss the fallout from Microsoft’s mega Patch Tuesday, which caused significant issues like broken Excel copy-paste. They analyze Andrew Ng’s and David Bellamy’s arguments against AI doomsday narratives, emphasizing that AI bioweapon risks are overstated due to physical lab bottlenecks. The hosts critique the outsourcing of AI security testing to the firm Irregular, which led to multiple model containment breaches, and cover the stalled US Frontier Act and EU Kids Act.
Topics discussed
Intro and Trusted Tech sponsorship
Episode overview and AI breakout context
AI agents buying fixes for legacy software
Lion Del Bazel security awareness training ad
Listener mail: AI coding price list wheel
Andrew's perspective on AI doom and hype
Bioweapon risks and responsibility debates
OpenAI containment failures and industry blame
Trump's AI comments and job market fears
RayNeo AI glasses sponsorship
Google Gemini hacking other companies
Anthropic Claude bioweapon research attempt
David Bellamy on AI bioweapon feasibility
US AI safety bill legislative status
Mobile app security sponsorship
EU Kids Act age verification and privacy
Flaws in EU age verification app
Nightmare Eclipse ex-Microsoft employee story
ThreatLocker network visibility sponsorship
AI-assisted attack on OpenAI forum
Cisco and Microsoft Patch Tuesday vulnerabilities
OutSystems AI agent sponsorship
Windows Update bugs and domain auth issues
AI coding reliability and future models
Show notes, streaming platforms, and outro
Silicon Valley Bank sponsorship
Listen ad-free on Castria