On: The incentives of academia naturally select for less reliable results (2016)

[Fomite]

He should be worried that the "skimming" he's talking about is, in many cases, legally mandated. Direct and indirect costs not being mixed isn't something that happened because universities wanted it that way.