this is the case with almost all drugs. for example almost every antidepressants was the result of a weight loss or something like that.

drug research is worse than training an LLM. you feed it to random people and then compare effects with initial conditions and of something looks good you try with more control.