Perhaps the first gate level computer that might be able to support a real world application that provides a quantum advantage is the 72 qubit machine being developed by Google using their Bristlecone (or follow-on) device. Google is currently working on a quantum supremacy experiment using this device that would indeed be a significant milestone. However this particular experiment is not intended for any real world application. We do, however, expect Google to be announcing successful completion of this experiment very soon.
Part of the problem with any quantum application is understanding how to bridge the mathematics of the real world application problem you are trying to solve with the mathematics of the quantum computer. This generally not an easy thing to do and there are many software startups that have been formed to help figure out how to do that.
But there is one application where it would be relatively easy to interface to a real world application and that is creating certified quantum random numbers. Random numbers are needed in many different situations including encryption keys, lottery picks, Monte Carlo simulation and many others. Most of the time these are generated using a software algorithm that starts with some seed number and runs it through an algorithm to create a jumble of random bits. This is called a pseudo-random generator because it has a weakness. Since classical computers are deterministic, if an adversary knows the seed and the particular algorithm used they would be able to predict the random number.
For users who are concerned about the security of pseudo-random number generation a class of products has been developed called hardware random number generators. These are little boxes that will leverage physical phenomena such as thermal noise, atmospheric radio noise, nuclear decay, or even quantum phenomena such as photons travelling through a semi-transparent mirror to create a random number that is not deterministic. These products are readily available and you can even purchase one for about $50 on Amazon with free shipping. However, there is one deficiency with these hardware random number generators. You need to trust that the designer of the device did not put in some backdoor that would allow some outsider to figure out what random numbers that you are generating and use the information for their own nefarious purposes.
So this creates an opportunity for a certified random number generator which will generate a true random number even if it was created by someone you do not trust. And Professor Scott Aaronson of the University of Texas has proposed a method to do this. The method involves sending a block of challenge data to the quantum computer, having the quantum computer compute a function based upon the data and sending the results back to the user. The function would be selected such that it would be easy for the quantum computer to compute, but very hard for a classical computer to do so in a reasonable amount of time. The algorithm would use a moderate capacity quantum NISQ computer with roughly 50-100 qubits and rely on a quantum program which is quite similar to Google’s experiment mentioned above. Such a program would run well on Google’s 72 qubit device.
But the next question is whether there a commercial market for such an application. Our expectation is that any commercial market will be very small. With so many other random number generation schemes available at a low cost, we do not think very many people will pay extra for a certified version. In addition, any application that requires that the random number remain secret, such as use as an encryption key, is problematical because the random data generated by the quantum computer could be intercepted by an adversary when it is transmitted back to the user. This could possibly be solved by using a quantum resistant QKD communications network, but that may have other logistical problems.
However, we have identified one potential application where a certified random number is quite desirable and it would not matter if the number were made public. This application would be for picking the numbers in government lotteries such as EuroMillions, Powerball or MEGA Millions. There is always concern with these public lotteries that there is no cheating, the dollar amounts are quite substantial, and it would not matter if the random numbers were made public as long as the ticket purchase window has closed. In addition, there would be some novelty in saying that the winning numbers were picked by a quantum computer and it may help sell a few more tickets! Like many other technology advancements, we do expect the initial usage to start small. Much of the early progress in some areas of classical computing, such as computer graphics, started with games and perhaps quantum computing will follow a similar path. But this will pave the way for larger and more complex applications as the technology continues to develop.
August 10, 2019