Epoch AI finds GLM 5.3 turned public Firefox fixes into working exploits
AIEpoch AI reports that GLM 5.3, an open-weight model from Z.ai, autonomously built exploits from Mozilla security fixes that were published but not yet shipped. The experiments reportedly succeeded against all six Firefox versions released between September 17 and October 26 for about $1,000 in compute each.
Loading source posts…
This story is featured in top picks: “Open-weight model GLM 5.3 turns unreleased Firefox fixes into working exploits”
Source details
Epoch AI · The Epoch Brief · Aiden Ament · epochai.substack.comPublished · added here
Original title: AI can turn public security fixes into working exploits before users receive an update