Defending at AI Speed as Quantum Threats and AI Policies Won't Save You - Nolan Karpinski - BSW #467
Summary
Nolan Karpinski from Google Cloud discusses defending at AI speed, focusing on agentic detection engineering and the future of the SOC. The episode covers how organizations must build unique "harnesses" with specific context to secure AI agents, as one-size-fits-all models fail across different industries. Karpinski emphasizes the need for internal AI centers of excellence to evaluate tools and manage the cost of digital workers versus human labor. The hosts also address the rising nation-state threat landscape, the importance of quantum readiness, and the critical role of executive relatio…
Topics discussed
Intro, sports weekend, and pirate costume
Sponsor and guest introduction: Nolan Karpinski
Defending at AI speed and agent-based threats
Competitive advantage and scaling AI SOC capabilities
The need for organizational context and 'harnesses'
Building golden evaluation sets for AI agents
AI self-regulation and business leader responsibility
Advice for smaller companies and AI CoE concepts
Evaluating AI tools and task alignment
Budgeting for digital workers vs. human staff
Cost of AI triage and risk-based prioritization
Agentic data retrieval and changing security philosophy
Closing interview and Isara Advanced ad
CSS 26 event announcement
AI reshaping the nation-state threat landscape
Security as a core business function
Post-quantum cryptography and asset inventory
AI policies vs. visibility and asset management
Monitoring AI agents and 'judge' architectures
Invisible work, burnout, and leadership empathy
Personal travel plans and final thoughts on AI
Listen ad-free on Castria