" it seems like any system leaks detected from a high pressure hydro would be detected at operating pressure"
In the case where the fluid to be handled is more dangerous than water, why would one want to wait and see if there is a leak after the system is placed in service? Would you actually want to work near a system that is handling ammonia or naptha or other toxic fluid if the system integrity was not proven?
That being said, there are often cases where an "in service" test of systems that carry only air or water are accepted by the client for some low pressure systems ( such as circulating water systems ), provided the testing and subsequent repair of the system will not hold up commissioning.
As meteng stated, the extraordinary stresses imposed by the hydro test may work-harden some stress raisers, as well as place into a highly stressed state pipe hangars, supports, gasket surfaces, threaded or welded connections, etc. These "shakedown " stresses not only condition the pipig system and vessels for the lower continuous service stresses, but the hydrotest period serves as a logical transfer of responsibility from fabricator to operator. Without such a proof test, any subsequent component failure cannot easily be proven to be the responsibility of teh designer or fabricator, and the authority of the local inspector would have been mooted.