Maybe you can come up with a particularly stringent conceptualization of what is or isn’t science, and maybe some fields of inquiry won’t meet that bar (like cosmology?), but that doesn’t make them “dumb”. Even without testability you can still reason about the likelihood of things.