Episode Appearances
All TWiT.tv Shows (Audio) · Jul 29, 2026
Security Now 1089: Models Go Rogue & ExploitGym
“Led research team that created Exploit Gym benchmark with 898 manually-crafted vulnerability instances across three domains”
AI Model Containment and Sandbox EscapeExploit Gym Benchmark and Autonomous ExploitationHugging Face Security Incident Analysis
View Analysis
AI + a16z · Jan 20, 2026
How Should AI Be Regulated? Use vs. Development
AI Regulation PolicyOpen Source AI DevelopmentUS-China AI Competition
View Analysis
Digital Disruption with Geoff Nielson · Jan 12, 2026
AI's Most Dangerous Truth: We've Already Lost Control
“White hat hacker; discovered 20 zero-day vulnerabilities in leading AI models using ChatGPT and Gemini”
AI Existential Risk AssessmentAI Safety Testing and Red TeamingAgentic vs. Non-Agentic AI Design
View Analysis