I'm getting a PhD and it's been 90% real stuff, with no regard towards impact metrics. It probably depends a lot on the subfield. My subfield is less "hot" than most, which some would take as a sign of unhealth, but it means we're spared all these other problems.