AI agents find exploits within minutes of bug rumors

Simon Willison · rss · 2026-08-29

Cambridge professor Anil Madhavapeddy reports that modern coding agents can find and exploit vulnerabilities within minutes of a bug being rumored. rclone maintainer confirms receiving 40+ disclosures in the last month vs. 20 in 10 years. This speed breaks existing open source security embargo practices.

Original post →

More from coding & agent

coding & agent channel →