I don't believe that assumption has any testability or untestability attached to it. I gave Asimov's definition as an example of assumptions in science. Not with particular reference to testability or no testability. I think assumptions are either useful or useless, I don't believe they are necessarily consistent or have the property of truth or falsehood. They are simply accepted as they are, as a proviso to the next step. To clarify, IMO, I can attach the prefix "I believe...:" to an assumption but not a hypothesis.