The current state of quantum computing really depends on who you listen to. If it's D-Wave Systems, one might get the impression that we're there right now: practical quantum computing. The Canadian firm has two "quantum computers" in its catalog currently, but they've been the source of much debate. Are they actually performing quantum calculations? If so, are they quantum calculations that even matter to super-fast computing? If the answer is no to either one, we're rather rudely shoved back into the world of classical physics and its boring rules about entities having to be either one thing or another rather than all things at the same time. That may be the case.
Recent experiments have suggested that nothing particularly quantum is going on in the D-Wave machines, despite heavy interest and investment in the technology by both Google and NASA. A crucial paper in Science from July, "found no evidence of quantum speedup." Now, Google is going in a different, more back-to-the-basics quantum direction. For one thing, the author of the Science paper, John Martinis, is now in Google's employ, tasked with advancing beyond tentative D-Wave technology into the promised land of true quantum.
It's like being able to order (and consume) everything on a restaurant's menu at once without barfing or dying.
Martinis work has involved a perhaps more theoretical realm of quantum computing (at least in that it's not marketing actual machines) that involves the construction and gradual scaling upward of qubits and qubit systems using superconductors. A qubit, if you'll recall, is the unit of quantum information, a carrier of not just a "1" or a "0" but both at once along with an infinity of states in between. Using qubits it becomes possible to evaluate not just some combination of 1s and 0s but rather every single intermediate state in one calculation. It's like being able to order (and consume) everything on a restaurant's menu at once without barfing or dying.
That's the generalization of quantum computing, but D-Wave machines use a rather peculiar architecture dedicated to optimization problems which takes advantage of the spontaneous "jumping" or tunneling of particles between different resting states. It's been difficult to determine whether this architecture is actually being aided by quantum effects or not, and it's been demonstrated that, given the right algorithm, it's quite possible for a standard laptop computer to smoke the latest D-Wave machine.
"The qubits of D-Wave's machine can maintain superpositions for periods lasting only nanoseconds," MIT's Technology Review reports. "Martinis has built qubits that can do that for as long as 30 microseconds, he says." A post-D-Wave chip using such qubits has far more potential to take quantum computing technology from the apparently trivial of D-Wave architecture to the widely functional and, thus, very, very fast.