My guess about why diseases are so hard to model is that all biological systems are self-organizing and thus quite stable (as systems can't really self-organize without navigating some staunch thermodynamic principles). When we want to model a disease, we think we can impose a different order externally, and the body is quite resistant to that change, as there can be redundant systems several times over. So we have to makes "super" versions of genetic mutations to make them the dominant force in the system, which is of course going to be far stronger than the "natural" mutations, because those mutations are a greater or lesser part of a multimodal failure.

I think that brain organoids' main contribution to neuroscience will not be clinical but rather in helping us to understand the underlying organizational structure by which the brain grows (and I'm not disparaging the work even a tiny bit, for the record; this would be no small achievement). The best we can hope for in regenerative medicine is to set up conditions in which the body can recapitulate its own developmental growth program.