Recently we discussed the extension of the agreement that Google and NASA maintain with the people at D-Wave, which will allow both names to get access to the latest version of its quantum computer. However, the main problems are two: It is still too early to obtain benefits from quantum computing, and studies have been published that declare those benefits nonexistent. This time, Google says it has concrete proof of the D-Wave's operation. Is that really so?
Quantum computing is one of the "great hopes" for the future of technology, but scientists do not like to work with hopes, to say the least. What they expect from quantum computing are facts, and so far they have not obtained them. With this I refer to the debate about the D-Wave quantum computer, first of its kind with commercial profile, and currently available to two titans like Google and NASA. Most of the doubts focus on the D-Wave system's chip. In the past, researchers from Google and other companies were unable to prove that performance improvements exist, since several conventional systems running simulations yielded very similar results. Today, the story seems to have changed.
Google has focused its efforts on the concept of "quantum annealing", extremely useful for facing optimization problems, which essentially keep machine learning and artificial intelligence experts awake at night. According to the latest information, in Mountain View they have solid proof that the "quantum annealer" in the D-Wave system is one hundred million times faster in solving a "made-to-order" problem than simulated annealing on a single core of a conventional chip (in this case, an Intel Xeon E5-1650). On the other hand, Google found that the D-Wave is far superior to another simulated algorithm on traditional hardware, the Quantum Monte Carlo.
The results are certainly encouraging, but at the end of the day they only serve to deepen the research. In Mountain View they acknowledge that there are algorithms with the capacity to take "shortcuts" that would lead them to solve their special optimization problem faster than the current quantum hardware. Even so, it is estimated that with the next generations of quantum chips, those shortcuts will lose their effectiveness completely. In other words: The D-Wave can develop a chilling speed... if the problem to solve is the ideal one.
