Not sure why parent is being down-voted as I believe this is an important point. In my opinion this would be applying the unix philosophy of having small tools that does one thing and does it well to code libraries.
Sure you could just "finish" Linux at 5.0 and then introduce e.g. io_uring via Linux-with-io_uring 1.0 instead of adding it to Linux 5.1. Same goes for all the libraries that add support for io_uring.
Yes, you could "finish" some software on the feature level, but you would still need to maintain it if you want to add support for new platforms, etc., or it will become obsolete sooner or later. In the case of still maintaining libraries, this would solve nothing in the context of this attack vector.