Have all the "easy startups" been done? Welcome to the "two PhD problem"!