Agreed, it is bit sensationalist. But what if someone proves it and then we realize that even though it can be done in polynomial time, there's some other constraint (like "x^N but N has to be super big") that renders the result impractical?
I think once someone proves it, it will be obvious in hindsight as a lot of people are just ignoring the isomorphisms in nature.
Yup, doing a big of handwavy math also demonstrates this: You are comparing something like x^N - x^M with 2^x - x^M. The second will always grow big rapidly, while the first could stay small, but it can also grow arbitrarily large. In fact, it is bound to do so for N > M and it will do so quite rapidly if N >> M.
Perhaps writing a piece of music is also done in polynomial time, but it requires more training (more advanced algorithm)? Or, perhaps, the time complexity is x^3 for writing, instead of x for appreciating.