There's nothing particularly controversial about it as far as I can tell. They have a mechanism for optimizing systems fairly quickly using either "true" quantum parallel tempering, or a classical (perhaps analog) simulation of the same - either way, the speed at which they can optimize is the deciding factor, and not whether there are "true quantum" effects. It certainly doesn't seem to be based on the traditional model of a quantum computer and I'd guess it isn't technically able to implement something like Shor's algorithm.
