S4E20 - AI Executive Order, Project Glasswing Expanding, Cybersecurity Workforce
Summary
Hosts discuss a new executive order mandating voluntary 30-day government access to frontier AI models and the creation of an AI cybersecurity clearinghouse. They analyze the Glasswing project’s expansion, noting that advanced models like Mythos are uncovering severe vulnerabilities at an alarming rate. The conversation also covers recursive self-improvement in AI, the impracticality of development pauses, and concerns about regulatory capture by major tech firms.
Topics discussed
Introduction and startup culture chat
Analysis of the June 2nd AI Executive Order
Podcast anniversary and open-source AI strategy
AI Cybersecurity Clearhouse and government coordination
Critical infrastructure entities and vulnerability scanning
Challenges in patching and operational stability
Recursive self-improvement and regulatory arbitrage
Secure AI evaluation environments and data privacy
AI's impact on hiring and the future of engineering
Listen ad-free on Castria