It's basically another form of premature optimization as it pertains to discovery and research. It stems from the misbegotten idea of permeating the last few decades that cost-benefits accounting and ROI should drive the entire world (and choosing the entirely wrong accounting model to apply uniformly on top of it).