I understand your statement here and I agree with it. Yet I guess both you and the other user here (@cyd) are missing what I am trying to describe. Maybe my explanation is not accurate and my understanding is not well developed.
Let me illustrate with an example : we can study new planes models :
a)- in a wind tunnel equipped with instrumentation (camera, smoke trail and so on), or
b)- with numerical simulations on a computer.
One method (a) is very specific to a very precise problem, it cannot be (easily) adapted to calculate various random problems. The other (b) is meant to be a versatile programmable computer and so can switch to a completely different problem in one microsecond.
For what I understand so-called quantum computers (of today) are more like option (a).