I think GP may be referring to CROs who generally do most of the ops of a clinical trial. Their incentive isn’t to have positive results though, it’s to milk the trial contract (cost plus contracting w/ ~no incentive to operate them quickly and efficiently) and then to have the results never get disputed by FDA. A CRO that fakes data is a CRO that will not survive very long.

Agreed on all your other points though, GP has no clue what they’re talking about.