There's a version of this that doesn't get talked about enough -- what happens to the compounds already in study when the researcher who designed the safety protocol leaves. Institutional knowledge about why certain interactions were flagged or screened against isn't usually documented well enough to hand off. It just lives in the PI's head.

We've been building Bio-Twin (biotwin.io) partly for exactly this reason -- AI pre-screening that externalizes the safety logic so it's not dependent on which scientist is still employed. Not pitching, just -- this is a real downstream consequence of the brain drain that seems underdiscussed here.