Exploits Now Beat Patches: Open Source Security in the AI Era - Patrick Smyth, Chainguard
AI tools enable low-skill attackers to operate at engineer-level, with exploits now beating patches
“everyone is moving up a level in terms of threat”
Chainguard's Patrick Smyth argues frontier AI has fundamentally shifted the open source security risk landscape by lowering the barrier to sophisticated attacks — citing a teen using ChatGPT to steal 4.4M records and a single actor using Claude Code to hit 17 organizations. The exploit-to-patch window has collapsed, with exploits now arriving before patches in a way that AI-assisted attackers can exploit faster than defenders can respond. The core claim is that every tier of attacker has effectively leveled up, meaning security teams must assume a permanently elevated threat complexity.