Cybersecurity Headlines Department of Know: VoidLink threatens multi-cloud, flaw threatens Claude extension, China practices on infrastructure
21 snips
Feb 17, 2026 Adam Palmer, CISO at First Hawaiian Bank, brings pragmatic risk and governance perspective. Jon Collins, Field CTO at GigaOm, offers cloud architecture and threat analysis. They discuss AI-assisted VoidLink multi-cloud malware, zero-click flaws in AI extensions, shadow AI and insecure defaults, and national cloud rehearsal practices. Short, sharp takes on resilience, visibility, and treating AI as privileged.
AI Snips
Chapters
Transcript
Episode notes
Internet-Exposed Admin Tools Are Never Internal
- Internet-exposed internal admin tools are exceptionally high risk regardless of vendor brand.
- Attackers exploit trust relationships rather than brute force when tools are exposed.
Shadow AI Grows Faster Than Governance
- Shadow AI risk scales when powerful agents deploy with insecure defaults outside IT visibility.
- Governance must catch up to prevent widespread, uncontrolled deployments.
AI Raises Attacker Efficiency, Not Always Novelty
- AI doesn't yet make malware smarter but it makes it cheaper and faster to produce at scale.
- Attackers are building multi-cloud persistence and won't respect provider boundaries anymore.