Getting Ahead of the Mythos Effect | Megan Horner

AI models like Claude Mythos now find vulnerabilities faster than any team can fix them. PHP2Shell went from a leaked exploit to millions of attack attempts in two days. Megan Horner of Seemplicity and Dan Barahona on what vulnerability management looks like when discovery is no longer the hard part. Takeaways: Finding vulnerabilities used to be the bottleneck. Now it's deciding what matters and getting it fixed. Gartner's CTEM has five stages: scope, discover, prioritize, validate, mobilize. Don't assume a finding is real. Validate that it's exploitable in your environment. Group findings by the fix, and get each one to the person who owns the asset. Start with what matters to your business, not the backlog. Speaker: Megan Horner, Director of Product Marketing, Seemplicity. Hosted by Dan Barahona, AI Security University. Chapters: 0:00 Welcome Megan Horner 3:44 What Claude Mythos changes 6:52 The PHP2Shell timeline 10:01 Discovery is no longer the bottleneck 12:37 Is AI making it worse? 16:07 Security in the software life cycle 20:20 Gartner CTEM and the poll 27:56 A better way to handle exposures 31:07 Poll results 31:48 Q&A: is it really exploitable? 37:07 How Seemplicity compares 41:15 Business risk and MTTR 43:42 Final thoughts Thanks to Seemplicity for supporting this session and helping keep our courses free: https://seemplicity.io Event page: https://aisec.university/events/getti... Free AI security courses: https://aisec.university/courses Join the community: https://community.aisec.university #AISecurity #VulnerabilityManagement #CTEM