Although quantum computing is evolving rapidly, we are not yet aware of any commercial application that directly uses quantum computation in a production environment.  What we have seen are a lot of theories, small scale demonstrations, synthetic benchmarks and technical papers that have given us a lot of confidence quantum computing will eventually turn out to be a useful commercial tool that will help us provide valuable products and services in ways not previously possible.  But convincing commercial end users to use quantum computing in a production mode will not be easy. If a problem can be solved using either a classical computer or a quantum computer, they will choose the classical approach because it will invariably be cheaper and more convenient.

However, what we have seen is that quantum computing has already created value indirectly by creating a sense of competition for classical computer scientists and spurring them on and find even better approaches for solving certain problems classically.  The newest buzzword for this is “quantum-inspired”.  By seeing how a quantum computer might theoretically solve a problem, some computer scientists are figuring out ways of coming up with a similar approach that uses existing classical computers.

A recent example of this was a paper published in July by University of Texas graduate Ewin Tang on a classical approach to solving the “recommendation problem”.  This refers to how companies like Netflix and Amazon makes recommendation on a product you might like to use next based upon your previous history and data on what other customers with similar tastes have watched. Up until Tang’s paper, the best known classical algorithm was thought to be exponentially slower than a proposed quantum algorithm published by Iordanis Kerenidis and Anupam Prakash in 2016, but not anymore.

Another example of a quantum-inspired solution is the Digital Annealer introduction by Fujitsu.  The D-Wave series of quantum annealers has provided a lot of impetus for solving problems through combinatorial optimization techniques and Fujitsu is now providing a way to do similar things using custom application specific digital ASICs.

A final reference is a video talk by Helmut Katzgraber presented in 2016 titled “Quantum vs Classical Optimization: A Status Update on the Arms Race”.  One of the concepts he presents is that there is a deep synergy between statistical physics and quantum information.  Another important point made in this talk is that a hybrid approach, quantum and classical calculations working together, might be the real key for overcoming the limits of Moore’s law. By understanding the best classical optimization techniques, one can make quantum optimizations work better and vice-versa.  And you can find many more examples of “quantum-inspired” by Googling the term and viewing the resulting references.

So the point is not that the classical approaches nor the quantum approach will always work better, but by better studying quantum techniques we can improve the classical approaches and raise the bar for when the quantum approach will have the advantage.  Like many aspects of human endeavor, the existence of competition makes everyone involved sharpen up their game.