I would say it is both very pragmatic and very reasonable approach. Agents will probably not go anywhere, they are indeed useful tools, banning them seems pretty much unrealistic since people will use them anyway.
i would be very costly to try and build detection systems as models improve. its pretty much impossible.
They should make AI an option in Linux instead of what Microsoft did to Windows 11 and make it installed with the OS.
Torvalds is talking about the use of AI to develop the kernel; not about any kind of AI features included in Linux. (Which would be a user-space thing anyway, and therefore up to the individual distributions.)