This isn't for lack of trying on the part of cryptographers - unconditional proofs of security for most modern cryptosystems would imply that P and NP are separ

This isn't for lack of trying on the part of cryptographers - unconditional proofs of security for most modern cryptosystems would imply that P and NP are separate. For example, a direct proof that SHA-256 is collision-resistant would imply that one-way functions exist unconditionally.