OP's claim was that quantitative finance was *hard science*. Requirements regarding predictability are way more stringent than merely observing stuff and seeing how it responds to an input.
These are colloquial terms [1]. We might as well argue about whether Pop Tarts are ravioli or tacos sandwiches.
Colloquial terms whose concrete meaning does not correspond to OP's claim.
There is no ambiguity in this: if your models are not testable and fail to predict behavior then it's not hard science.