Great thoughts from @owl_posting on organoids as models for Alzehimers. Raises the more important general point that "all models are wrong but some are useful" and more importantly some are useful in particular ways even as they are wrong in other ways. Often in drug discovery it's unclear if a model is useful. Ideally it would discriminate between successful and unsuccseful drug but you don't know that until years later, if ever. And then even in cases where we are pretty sure a model system is useful (maybe it's predicting downstream advancement well), you often aren't sure why. With so many ways to be wrong, you can always take the negative view and point out what's almost certainly wrong, or you can try to find better things and see how they work out.