Sometimes when we argue about whether my Quantum Volume is bigger than yours, we forget that the real opponent for quantum computing is classical computing. Although Moore’s Law may be slowing down, it hasn’t stopped altogether and continued advances are being made in semiconductor technology. And classical computing is continuing to advance in other areas. More innovation is being made in non-Von Neumann architectures and developments in GPU’s and neuromorphic computing are continuing to improve the performance of classical computing performance. What’s more, the software folks are continuing to innovate and create new algorithms that can provide solutions to problems that are more efficient and effective. And the thing to realize is that classical computing is still getting better every year. So beating it out represents a moving target.
But sometimes people can get confused about what really separates classical computing from quantum computing. The definition we use is the following. If a solution leverages the quantum mechanical principles of superposition and entanglement it can be called a quantum solution, or at least a hybrid classical/quantum solution. If the solution does not utilize these phenomena, we will call it a classical solution even though it may not look like a normal classical computing solution.
So the goal of quantum computing, in general, is to achieve quantum advantage which demonstrates that the quantum solution can provide a real world advantage over a solution attempted using a classical approach. This is when quantum computing will start providing return on the large investments they are making in this technology. Quantum supremacy, which we define as demonstrating that a quantum computer can do something that cannot be done on a classical computer is a stepping stone but it does not require the all-important factor of doing something that is commercially useful.
In the past few years, a new buzzword has evolved that is called quantum-inspired. The idea is that you take a problem which traditionally has been difficult to solve on a traditional classical computer and look at it in a new way. Perhaps some of the concepts that people studied in quantum computing can be modified so that they can be implemented on a classical computer.
Actually, this concept of looking at a problem in a new way to arrive at a solution is very common within engineering and many other fields of endeavor. Sometimes, an engineering group will bring in a new member with a different background who can suggest an approach that the others hadn’t considered. Perhaps they bring in an analog electronics engineer into a team of digital engineers who helps them develop a solution to a difficult problem. You could call a situation like this “analog-inspired”, but that term just doesn’t seem as sexy today as calling something like “quantum-inspired”.
To be clear, “quantum-inspired” is not quantum but a classical approach because these solutions do not leverage the quantum mechanical principles of superposition and entanglement. We don’t particularly care for the term, but marketing folks like it because it adds some excitement when they promote their products.
But whether you like the term or not, it is a valid approach for some companies. These companies, particularly the ones that are mature, don’t view their fundamental mission as one of advancing quantum technology. They see their role as one of helping their customers solve important computational problems to help them to become more successful. Microsoft is one such company and it makes sense for them to do this. They have a large base of customers and also a large base of employees with a variety of skills, including some who are expert algorithms designers in classical computing. So Microsoft has promoted their quantum-inspired solutions because it allows them to solve customer problems without having to wait for a more powerful quantum computer to be developed. Other companies, such as Quantum Computing Inc., have developed software approaches that can run today on a classical computer and then be converted to run in the future on quantum computers when it makes sense. These approaches make commercial sense for some companies as it allows them to start seeing customer revenue at an earlier stage.
However, whether this approach is suitable for a specific company will depend upon the background and capabilities of its staff. Some of the quantum software startup companies we know are primarily comprised of physics PhD’s who have studied quantum for a decade or more, but who may not have the classical software and algorithm experience to effectively pull off a quantum-inspired solution. For these companies, it may make more sense for them to stick to a pure quantum or hybrid quantum/classical approach and forget about providing pure classical solutions.
So every company will need to think about their strategy and what is the best approach for them to achieve success. But we would urge everyone to focus on the customers and their real world problems and realize that artificial quantum metrics are only a proxy for how close they are to achieving this goal.
November 5, 2020