Friday Squid Blogging: New Jurassic Squid Predecessors

Schneier's October–November 2026 posts discuss emerging security and technology risks, including AI-assisted vulnerability discovery, hidden capabilities and backdoors in AI models, security bias in code-generating models, spyware and bloatware in aftermarket automotive head units, fraudulent North Korean IT workers, and AI exploitation of legal systems. The strongest security concern is that adversarially trained or cryptographically concealed model capabilities may evade conventional safety evaluations. No specific software vulnerability or confirmed exploit affecting a named product is rel​