Linux kernel maintainer Greg Kroah-Hartman has said that, after an initial rough patch, high-quality AI-driven code review has “really jumped” for Linux.
Kroah-Hartman said: “Something happened a month ago, and the world switched. Now we have real reports. All open-source projects have real reports that are made with AI, but they’re good, and they’re real.”
He said security teams across major open-source projects chat informally and often, and everyone is clocking the same shift in what lands in their inboxes.
Kroah-Hartman said: “All open-source security teams are hitting this right now and the sudden flood is changing how review work gets triaged.”
For now, AI is showing up more as a reviewer and assistant than a full author of kernel code, but the boundary is starting to wobble.
Kroah-Hartman said: “I did a really stupid prompt,” and then watched the machine spray out fixes at scale.I said, ‘Give me this,’ and it spit out 60: ‘Here’s 60 problems I found, and here’s the fixes for them.’ About one-third were wrong, but they still pointed out a relatively real problem, and two-thirds of the patches were right.”
Code still needed human cleanup, better changelogs and integration work, but the tools were good and “We can’t ignore this stuff. It’s coming up, and it’s getting better,” he said.
He said AI can already crank out usable patches for “simple little error conditions, properly detecting error conditions,” today.
