As I understand it (and I'm prepared to admit that my understanding may be a bit shakey), what quantum computation buys you is (effectively) a reduction in the order of computational complexity of certain algorithms (say, a reduction from O(e^N) to O(N), though I don't know that that's the particular reduction we're talking about). For those of us who eat CPU cycles like candy, that can only be good news. However, the last I heard, this reduction was only applicable to certain classes of algorithms and not to algorithms in general. It may be, for example, that factoring falls to QC but other problems do not. Further, the ability to compute at a higher order of complexity is potentially a sword that cuts both ways — e.g., it may be feasible to develop new encryption algorithms that are O(e^N) given a quantum CPU to encrypt with.

The so-called quantum cryptography stuff is not encryption in the mathematical sense but rather a physical technique for constructing a perfectly tamper-evident communications medium. This may be valid and important technology in its own right, but I don't know that it addresses the computational issues introduced by quantum computation. (I'm also curious how one might actually use quantum cryptography in real communications systems. Essentially what they are constructing is a channel so delicate that anything that could perceive the signal will destroy the signal. However, traditional transmission media suffer signal loss all the time anyhow — that's why we put in error correction. It strikes me that any kind of error correction techniques applied to a quantum cryptographic channel will eliminate its security, but if you don't have error correction you won't be able to communicate reliably. Maybe those guys have an answer to this; goodness knows they've thought about it more than I have, but I've never heard this issue raised before.)

They say a forty atom quantum computer can rival the power of our biggest supercomputers. And a thousand qubit quantum computer can quickly solve all the problems currently considered intractible by computer science. Cryptograpy based on factoring huge numbers appears doomed.

Or is it? Researchers in Cambridge, UK are working on quantum cryptography which may thwart QC's ability to break codes based on factoring.