> doctors based on experience found things that worked better at treating patients. and they also had completely wrong theoretical models to rationalize these t

> doctors based on experience found things that worked better at treating patients. and they also had completely wrong theoretical models to rationalize these things. And in truth this is a lot of what Asimov was talking about. Because it is how accurate the theoretical model is. I'll often say that "truth doesn't exist" but I'm careful about my context. Truth doesn't exist because it requires infinite information. But that doesn't mean we can't trend towards it. This statement is no different than "all models are wrong, but some models are useful." This is why I talk about science not actually being able to prove things, but rather about ruling things out. That's what is important. The model is always wrong but is it improving? If it is having a better total outcome and able to explain more than the previous model, then yes. That's the thing. It has to be able to explain what the last model does, and more. If it can't do both those things, then it isn't progress (yet at least). And I think our only distinction of scientific progress here is that you might need to consider that this is a high dimensional problem, with many basis vectors. What science is, is what is observable and testable. I jokingly call physics "the subset of mathematics that reflects the observable world." (This is also a slight at String Theory...)