But that's what I'm saying; there is lots of cryptography that is entirely unmotivated by attacks, like: work on determining the size of complexity classes lik

But that's what I'm saying; there is lots of cryptography that is entirely unmotivated by attacks, like: work on determining the size of complexity classes like CZK and SZK; Determining lower bounds of round complexity of MPC in various models; Determining the kinds of assumptions different primitives can be based upon; Determining the complexity classes supported by different delegation of computation schemes, and the assumptions these can be based on; Determining whether you can obtain particular primitives out of others, such as OWPs from OWFs in a non blackbox manner, seeing what iO can be used for, etc; Improving asymptotic efficiency of various primitives, like NIZKs (see SNARKs). I think the difference in our viewpoints comes from the communities we're; I'm in the theoretical, rooted in complexity theory side of things, whereas you appear to be on the more applied side of crypto, given your emphasis on things that are being used in practice.