Interview with Robert Sutor, VP for Quantum Ecosystem Development, IBM Research
By James Kobielus
Robert Sutor is a leading authority on quantum computing in the enterprise. He is currently IBM Research’s Vice President for Quantum Ecosystem Development.
Sutor will deliver the plenary keynote at Inside Quantum Technology New York, June 1-5, 2020. Recently, Inside Quantum Technology interviewed him for his insights on the potential, adoption, and status of quantum computing in the enterprise. Here is an abridged, edited transcript of what he told us.
Q: What is quantum computing?
You do not have to be a physicist to understand quantum computing….[T]here is behavior and there are these wonderful terms that people insist on mentioning, such as superposition and entanglement and interference. But these are very technical terms. And to be honest, from a business perspective, you don’t need to go there.
But what you do need to go to is understanding that quantum computing is a completely different programming model from what we do classically….When you start adding more qubits, there’s something mathematically logical, but nevertheless seemingly miraculous that happens, which is the amount of information you can store and work with doubles….
Sounds miraculous, but also crazy, like, how can you possibly do that? Well, the reason is….classical computing is just a system of computing that we happen to devise, but quantum computing is the type of computing that nature, that the universe actually uses. So really, in a nutshell, we are trying to harness this power of emulating how nature itself is a supercomputer.
Q: What types of applications are quantum computers good for?
Quantum computing is not good for all types of problems….Quantum computers are not a replacement for classical computers. They’re meant to work together. In fact, when I say quantum computing, I mean quantum computers plus classical systems….It’s going to take a little while until the quantum computers are big enough or powerful enough to really do anything better than we can do classically already. So, if you will, the return on investment is several years away…..
Many people are starting to talk about….artificial intelligence [applications of quantum computing]. [We must] be careful [when we discuss AI in this context] because…quantum computers are not big data machines. What [quantum computers can do] with AI is make some calculations–such as machine learning or deep learning–much faster…Try to use quantum computing to speed up some of the math….such as linear algebra, matrix calculation,…..matrix inversions, and things like that…
Q: How mature is the quantum technology market at this time?
Quantum computers cannot today solve any practical problem faster than classical. So why would anybody pay anything? Just do it classically, and, for many things, that’s what people are doing.
So the [quantum computing] market today is for people who want access to quantum computers to start working to get eventual first-to-market advantage….
What we’re probably going to see, roughly speaking, mid-decade, some examples of…quantum systems [that] do better than classical systems alone. People will then try to take what are very specific solutions and generalize….
By the end of the decade, we’ll see somewhat wider use of it. ….The underlying technology will be changing and improving in some very technical aspects, such as fault tolerance, error correction, and scaling.
Q: How can quantum and traditional computers complement each other?
There is no pure use of a quantum computer.
You must somehow also have classical computers involved, too, to translate what your application is right into terms that the quantum computer into a form of the quantum computer can use. And then to translate the answer back. Because ultimately we start with lots of zeros and ones and those zeros and ones could look on the screen like code….
So your core quantum computer is surrounded by several different classical computers and all of the quantum computers today we have or access via the cloud….So and that’s why I say quantum computing is quantum computers plus classical systems….
The limitations of quantum technology today [pertain to whether they’re] powerful enough to do problems of interest. Can your quantum computer do that work?…
There are two quantum ways of thinking about this. What is the quantum volume of your system and what is the quantum volume required to solve your problem? We need to get much higher quantum volumes in our systems to solve problems of interest. And so that means creating systems with more and more qubits and that have less and less noise associated with them, so that is that they compute accurately enough.
Q: What’s the quantum learning curve for IT professionals?
We used to hand out these stickers that say you’re thinking too classically the because there are some very unexpected behavior. Perfectly normal if you understand quantum but unexpected….
We provide many resources….[IBM’s Qiskit] open source programming platform for quantum [has] been downloaded over 300,000 times…It would take a little time [to get up to speed on it], but it’s certainly very possible…..
if you’re using Qiskit, you’re using Python, and the latest estimate I saw was that 8 million people classify themselves as Python programmers. And so if you’re a coding Python, you are used to importing libraries and then using those functions or methods that the libraries provide to do whatever you want to do. That speeds up a whole lot regarding what people need to know….
Quantum has a very strong aspect of probability at the core of it. So you must bear that in mind. It’s not always deterministic. You do not always get the same answer every time you run your application. However, we can do an analysis and say if you do it so many times, you have a 99.9 or whatever percent chance of getting the correct answer.
To hear Bob for yourself and learn from other quantum computing experts, visit this site to register for the streaming Inside Quantum Technology New York conference taking place June 1-5.