The Stack Overflow Podcast AI cybersecurity is a cat and mouse game
12 snips
Sep 11, 2026 Sam Curry, Zscaler’s Chief Security Officer and a cybersecurity leader in zero trust and cloud security, explores AI’s fast-changing security battlefield. They discuss why secure engineering matters more than endless whack-a-mole, how runtime protection and simpler infrastructure reduce attack surfaces, and the risks hidden in SaaS, multitenant clouds, and browser-based AI. Resilience, shared standards, and smarter TLS inspection round out the conversation.
AI Snips
Chapters
Transcript
Episode notes
AI Expands Engineering Beyond Writing Code
- AI may compress coding tasks while expanding engineering’s responsibility for architecture, infrastructure, supportability, and high-quality code at scale.
- Security teams also face signals every few minutes instead of every few hours, increasing investigation and accountability demands.
Reduce Attack Surface Instead Of Chasing AI
- AI makes attacks more adaptable, turning cybersecurity into a cat-and-mouse contest involving both pattern detection and concealment.
- Sam Curry argues defenders should reduce attack surface instead of merely accelerating the same offensive and defensive cycle.
Shift Security Protections Into Production
- Put protections in production to block vulnerability probing, exploitation methods, unnecessary functions, lateral movement, and excess data exposure.
- Use zero trust, encryption, tokenization, honeynets, and realistic decoys to limit options and waste attackers’ time.