Touvaz on his "love-hate relationship" with AI: Tools are useful yet dangerous; any vulnerabilities found by AI should be treated as public information

Linus Torvalds, the creator of the Linux kernel, recently stated at the Open Source Summit North America hosted by the Linux Foundation that his relationship with AI is ‘love-hate’ — “Technically speaking, I really like it; the tools are useful and interesting, but they do bring quite a few headaches.” He mentioned using AI to generate code for his personal ‘side projects’, though he still reads through the resulting assembly-level output himself: “You need to understand both your prompts and the final results, since those are what actually run.” Torvalds expressed skepticism toward claims that codebases are ‘100% AI-generated’, insisting that “AI is a great tool, but it’s just a tool.”

During the same talk, Torvalds also proposed a new rule regarding AI and security: if someone discovers a security vulnerability using AI, it should be treated as publicly disclosed information — “Because if you found it using AI, 100 other people have too.” This implies that AI-driven vulnerability discovery will drastically shorten the window between detection and disclosure, raising demands on security response timelines. On a positive note, Torvalds pointed out that the rise of AI has motivated NVIDIA to contribute more code to the Linux kernel — a change he welcomes. He also believes one of AI’s most promising uses is sparking young developers’ enthusiasm for programming.

ZDNet | TechRadar