Tbh, I didn't think he gave that message. If anything, he even identifies the 50-50 model as a local "maximum" (i.e. it has value). To me the main point was to have a mindset that is willing to challenge the possibility of whether such an optimum could actually simply be a local one.