The expected spread of frontier AI capabilities for vulnerability discovery such as Anthropic’s Claude Mythos means that organizations will need to prioritize patching based on the exploitability of software flaws, according to Optiv CISO Rob Gregory.