
NewsReportCorrectedLens
OpenAI says its AI agents built 948 Linux root exploits in one run. What defenders should take from it
An automated pipeline found, exploited and drafted fixes for hundreds of Linux kernel privilege-escalation flaws for about $1.2 million in compute. The bottleneck now is getting patches onto real machines.
Key claims: 1 corroborated · 3 claimedSources: Help Net Security, CSO Online · 3 min read
Read the full story →






