A pet peeve of mine is the statement that pops up every once in a while that number theory is not completely abstract and done for its own sake, and the example given is always cryptography. It is such a cliché phrase that it raises the question of other possible uses.
To specify, I refer to number theory as the study of integers and to applications as uses other than examples in other fields (for example, I understand modular arithmetic is an instance of a group, but it alone does not involve a number-theoretical finding being relevant in a non-number-theoretical context).