surely that depends on the size of the training set? I could feed the Linux kernel one function at a time into a ML model, then coerce its output to be exactly

surely that depends on the size of the training set? I could feed the Linux kernel one function at a time into a ML model, then coerce its output to be exactly the same as the input this is obviously copyright infringement whereas in the github case where they've trained it on millions of projects maybe it isn't? does the training set size become relevant legally?