Mythos and Security

The recent announcement of the capabilities of the Mythos AI model have raised many concerns about the cybersecurity implications of ever more powerful AI tools. Scott Crawford and Daniel Kennedy return to the podcast to debate the impact with host Eric Hanselman. Is this more marketing than real attack potential? It seems like the pace and scale of the discovery of new vulnerabilities could overwhelm traditional remediation and patching practices. Is this a reason to hold back the release of new AI models? If nothing else, these developments should serve as a wakeup call for enterprises to revamp the way in which they approach both their software supply chains and operational procedures. Increased speed and scale are necessary, but it's also critical to secure the source of patches and updates to counter attacker efforts to compromise them. Attackers have always been good at taking well intentioned improvements and weaponizing them. More S&P Global Content: Next in Tech | Ep. 259: The RSAC Conference – Agents on The Loose AI in action: unleashing agentic potential 2026 Trends in Information Security For S&P Global subscribers: RSAC Conference 2026: Record turnout confronts both the promise and threat of agentic AI Information security budgets rise, but must address cloud security, AI governance – Highlights from… AI's impact in security and its application are not always aligned – Highlights from VotE: Informat… Credits: Host/Author: Eric Hanselman Guests: Scott Crawford, Dan Kennedy Producer/Editor: Feranmi Adeoshun Published With Assistance From: Sophie Carr, Kyra Smith, Dylan Scheible