The possibility that P = NP but with a “large polynomial degree” or constant is too ridiculous to be taken seriously; the algorithmic complexity of the set of NP-complete problems does not permit a shortcut that characterizes the entire set in a way that would allow such a solution to exist.
Would you elaborate.
Under the right conditions, yes.