While we all can agree to that, reducing the PhD to just the new MSc comments is a bit worrisome.
But then take a look at so many ML PhD theses and one can truly see this is the case. So much unregulated garbage on ICML, NeurIPS as well.