Might be O/T, but I'm always surprised at how much our understanding of medical science comes down to basically a set of integration tests. You don't really kno

Might be O/T, but I'm always surprised at how much our understanding of medical science comes down to basically a set of integration tests. You don't really know if your solution is going to be a global optimum (it probably isn't); you have to memorize a bunch of stuff in order to have the full context of a problem available when you approach a subset of it at work; and sometimes the "ideal" solution is so stochastic that you need multiple opinions / consensus. I guess it's kind of like hardware engineering; I think we only glossed over how the physics worked in a CPU worked before working on higher-level things like NPN/PNP transistors and such, but even so there's a digital abstraction layer in computer architecture that make effective reasoning so much easier (even though there's really no such thing as "digital", only analog and clear thresholds) Am I (hopefully) wrong on this? Can modern medicine ever come from a first-principles approach?