publish or perish, lifelong grudges from irreplaceable never-wrong superiors looking to further their own Erdős numbers.
unpaid peer reviews that then become paywalled. rent-seeking from public libraries / community colleges via bundling.
data "science" aimed at producing a predefined conclusion. e.g. corporate research grants. p-hacking, dismissal/omission of inconvenient results that dont fit a narrative.
the adage "people don't leave companies, they leave managers" is simply not a thing that can happen without risking your entire carreer.