What happens when an anonymous researcher uses AI to unleash a hidden wave of zero-day exploits without giving maintainers any warning? In this episode of the Security Boulevard Podcast, hosts Tom Hollingsworth and Fernando Montenegro dive into the chaos caused by the "Exploitarium" code drop on GitHub, where AI tools were used to uncover zero-day vulnerabilities in critical open-source software like libssh2. The hosts debate the dangers of uncoordinated disclosure, the ethical lines around AI-driven bug hunting, and the unfair burden placed on open-source maintainers who are left scrambling to fix code. They also discuss the emerging reality of "security from AI," offering practical advice on how organizations can better prepare for a world where adversaries automate exploit discovery at scale.