Linus Torvalds has described the current situation of the kernel security list as nearly unmanageable. The creator of Linux denounces an avalanche of bug reports generated by artificial intelligence, where different users use the same AI tools to report the same failures, generating massive duplication which he describes as a waste of time and useless work.
The automation dilemma in kernel security 🤖
Torvalds explained that the errors detected by AI are not secrets, so keeping them in private lists only worsens the problem. Duplication occurs because several people run the same automated tools on the same code base, generating identical reports. This saturates the security team's communication channels, which must manually filter between legitimate reports and duplicates, slowing down the correction of real vulnerabilities.
When AI discovers the same bug 47 times 🔄
It seems artificial intelligence is not smart enough to realize that bug has already been reported. Now kernel maintainers must read 50 nearly identical reports before finding a useful one. Soon they will need an AI to filter reports generated by another AI, forming an infinite loop of useless work. At least the mail servers are happy with all the traffic.