Who fixes the zero-days AI finds in abandoned software?
Anthropic's red team discovered over 500 critical vulnerabilities using Claude AI, focusing on maintained software. The greater concern lies in the long tail of vulnerabilities in abandoned software that will likely never be patched.