“It may possibly even be a somewhat painful tool, each for maintainer workloads and just from a ‘it keeps finding embarrassing bugs’ standpoint,” he said of the usage of AI in security scanning. “The answer is to make sure that those LLM tools help maintainers as a substitute of just causing them pain.”
Developers ought to be free to decide on whether or not they use AI, he said. “We’re not forcing anybody to make use of it, but I’ll very loudly ignore individuals who attempt to argue against other people from using it.”
Torvalds’ measured support for AI doesn’t come completely out of the blue: Around the identical time that he was complaining about AI distorting security maintenance, he also spoke about its usefulness, claiming that it could improve programmer productivity by an element of 10.