Podcast with Doug Finke, Publisher and Managing Editor, Quantum Computing Report

My guest today is Doug Finke, publisher and managing editor of the Quantum Computing Report. Doug and I spoke about export licenses and regulations in quantum computing, avoiding the quantum hype, about the growing importance of business translators, and much more.

Listen to additional podcasts here

THE FULL TRANSCRIPT IS BELOW

Yuval Boger (Classiq): Hello, Doug. Thanks for joining me today.

Doug Finke (Quantum Computing Report): Hi, Yuval. Nice to meet with you.

Yuval: So who are you and what do you do?

Doug: So I'm Doug Finke. I am the publisher and managing editor of the Quantum Computing Report. That was actually the very first market analysis website, an organization devoted to quantum computer computing. We formed it in 2015. At that time, quantum was sort of I would call a research backwater, but I got interested in it. I had spent several decades in the classical computing industry and I wanted a break. I wanted to do something new. So I formed that. I guess my timing was pretty good because the quantum industry has certainly taken off over the past six years. I've been around to report about it. We have over 500 pages of a material on this website, as well as thousands of subscribers and tens of thousands of page views. So it's been quite popular.

Yuval: So Doug, I saw a presentation you gave at the Chicago Quantum Exchange a couple weeks ago, and I think you had some bullet points about providers and investors and end users and educators and so on. I wanted to dive into a couple of these points if I may. The first thing is about investors. I think that you mentioned that investors should perform careful due diligence because quantum computing is hard and some efforts will fail, but isn't that true of all investment proposals and all markets. What's unique about quantum here?

Doug: Well, it definitely should be true that particularly if you're a venture investor in an early stage, there's always a high amount of risk. Venture capitalists will make investments with the assumption that a large portion of them then will fail. Sometimes I hear stories that investors maybe forget about quantum. It's so magic that I've heard stories where an investor have a meeting with a prospective company, a prospective quantum company had hour’s meeting and said, we want to invest 50 million. The guy couldn't believe it. He says, "Well, you're not an investor, you're a gambler." So just people the need to be careful, because I do believe that there's going to be a lot of failures in the industry. We track over 225 private and startup companies on the quantum computing report right now.

My prediction I'll go out on a limb, my prediction is that 10 years from now, the majority of those companies will not be around in the same form they are today. Now they won't all go bankrupt. Some will be bought out. Some will merge some will reconfigured, but some of them indeed will go bankrupt. So and investors just need to be careful. They need to think about where a company will add value, how they'll fit into the ecosystem, which is actually getting to be quite large. They just need to be careful like they do with all their other investments.

Yuval: There are many companies that do hardware, so I share your view. There is probably not going to be a place for so many companies to do at least superficially the same thing in the market. So some will fail and some will merge and some will wildly succeed. So we hope, but actually let's talk about the bigger companies. When you look at the Amazon or Google or Microsoft, that all have a significant bets in quantum that are hiring a lot of people that are building quantum computers and so on. Which large company are you surprised that it is not in quantum right now, at least not publicly.

Doug: Well, you want to look at the large classical companies. I would point out one company that I'm a little bit surprised, that I haven't heard of an effort is HPE, Hewlett Packard Enterprise. They do a lot of work in enterprise computing. Another on the cloud side is Oracle they're quite big in cloud computing and they haven't jumped in yet. Dell has just started it was very, very interesting if you'd asked me a month ago, I would say that Dell would also be on the list, but they just did announce they're starting to work a little bit with IBM and I listened to one of their webinars maybe week or two ago.

They said that they'd been looking at this market for a couple of years, but so certainly any large company that has large exposure to the classical industry really should be looking at this. I won't say that quantum computer will completely take over classical computing, but I definitely feel that it will certainly take a chunk and companies really, if you want be involved in high performance computing, you really need to have a strategy for quantum.

Yuval: When you look on the customer side, on the large companies that are experimenting with quantum that are either “quantum curious” or setting up proof of concepts and so on, how much are they worried about IP in the same of how much do they want to do their work internally versus outsource, bring consultants, bring someone that says, "Oh, I'll just write the code for you and I'll give you it. I'll give you a license to use it." Do you sense that's an issue with the end customers?

Doug: I think it probably varies with a specific customer. I would say though at this stage, many of the enterprises, they realize that they are really in a learning and a training mode. So they may be just be trying to learn how to program these things. So I don't think they're quite as protective around IP as they might be for a more mature technology. It's really doing maybe a few proof of concepts as the industry matures. I think they'll be more sensitive about that

Yuval: Today. The largest quantum computer has about 50 qubits or in that vicinity. I know the qubit is obviously not the only measure of the capabilities of the quantum computer, but it's a good first approximation. I think some companies say, "Well, 50 qubits is sort of the maximum that you can simulate on a classical computer. Therefore, next year, when we have hundred qubit machines, we'll be at that inflection point. We'll be at that point where a chasm starts to open between classical capabilities and quantum capabilities." Others might say, "No, this is just hype. The industry is over hyped. It's going to take longer to develop. Right now be careful not to set your expectations too high." Where are you on that spectrum?

Doug: Well, I do definitely see the roadmaps where people are talking about a hundred qubits or more in the next few months and hundreds of qubits in 2023 and then maybe a thousand qubits or so in a couple of years. I would point out though that if you could simulate 50 qubits on a classical computer, those are perfect qubits - a hundred percent fidelity - versus the quantum machines that may hit a hundred or 200 in the next few months, they won't have perfect fidelity. They'll have fidelity of maybe 99% or 99.5% certainly depends on the particular vendor. So I do think we're certainly entering the area of what people NISQ where there is not enough qubits to do full error correction. You're going to have to live with errors in the computers, but people develop algorithms that can help work around that. Things like the QAOA and VQE. 

My personal opinion is I do believe there'll be a few applications. We’ll start seeing a trickle of applications that do quantum advantage or what I would call using quantum in a production mode for real world problem within the next year or two. But it won't be just a huge amount. There are many problems that will require thousands or even tens of thousands or hundreds of thousands of qubits. Some of them will actually require error corrected qubits. When you're talking error corrected qubits, you really need machines that are in the hundreds of thousands or millions of physical qubits.

But I do believe that you'll start seeing some production level type of applications within the next year or two, and it'll get certainly people, some hope that things are going to move on, but it just won't be a wave immediately. It's going to be what I would call a trickle, but gradually year after year, as the machines get more capable, more and more users will do that. It definitely will be  what you might call a mainstream significant technology by the end of the decade.

Yuval: What's your guess on the applications? What would be those trickling applications that could deliver some value in a year or two?

Doug: Well, I certainly think that optimization is an area that has a lot of possibilities so certainly D-Wave has been talking about doing optimizations on their quantum annealer. I know that's a different form of quantum computer than most others, but they've been working on that for a long time. They have users who have been putting together proof of concepts. They have users who are in what they call pilot production. So optimization, in some respects, I think might be easier for people to conceptualize because if you're a data analyst at a company, I think it's easier for you to think about your specific problem. 

Let's say you're trying to optimize deliveries for a grocery chain or something like that. It's relatively easy for you to identify what the constraints are and to identify what you want to optimize and put that into some type of equation that you want to find and you want to minimize. So I think optimization may one of the first areas, things like quantum chemistry probably I think will be next that requires a little bit more thinking about how you want to structure the problem and how you want to convert the problem from what I would call the problem statement to the Hadamard gates and the CNOT gates and stuff like that.

Quantum machine learning or quantum AI, I think is going to take a little bit longer just because I don't think the theorists really understand how classical AI works. A lot of it's been done empirically, they don't have the full theorems on that. So I think moving that to quantum is just going to take a little bit longer. So that would in general, that would be sort of the sequence I see.

Yuval: Going back to the slide you presented in Chicago, let's talk a little bit about governments and about cloud versus on-premise computers. Most cloud computing today is on the cloud. That's probably a good idea. It's easy to experiment. You don't have to spend a lot of money buying one, but there's some thought some countries may absolutely want to have quantum computers there because they worry about export restrictions. They worry that at some point, this becomes so strategic that some of the quantum superpowers may say, you just can't use my computers anymore, or you cannot use them now and so on, and so on. How many are you hearing about this?

Doug: Yes, I am hearing that relatively frequently, although I would not phrase it as a cloud versus on-premise. I would say that what countries are worried about is their data crossing the border, but they would be comfortable with quantum machines that are sitting inside their countries and I would point out for example, a relatively recent move by IBM, they've placed or are placing a quantum computer in Germany, actually they have already, and they have another one in Japan. So if you're a German user of the IBM system, you don't have to send your data to New York anymore. You could send your data to the IBM quantum computer that's sitting in Germany. I think you're going to see more and more of that. On-premise is going to be a little bit tricky in the near term, just because these machines are so finnicky . You have things like, how do you calibrate it? How do you have spare parts? A lot of logistical details that are not trivial. 

It's not like walking into your nearest Best Buy and picking up a quantum computer and taking it home in your car. These require pretty much onsite maintenance. That's why all the current providers really want to have the machines in their own facilities. IBM is lucky. They have facilities around the world, but in places where they can have an engineer. So if something goes wrong, they can tend to it very, very quickly. So on-premise is going to take a while, because the machines will need to get small. They'll need to get more reliable and require less maintenance.

Yuval: But will governments want them in country?

Doug: I think in country will become important over the next few years.

Yuval: Today there seems to be almost like language barrier, not so much English versus French, but a language barrier between the chemist and the supply chain experts and so on. Then the quantum engineers, some companies talk about the need for a business translator that can hear the problem and then talk to the scientists and so on. Do you see that role continuing into the future or do you see the chemist and logistics guys getting a lot of quantum education or do you see the software tools providing abstraction layers that allows the chemist just to write a program just like they write one in MATLAB today?

Doug: Yeah. Well I think all of the above will happen. I would point out I would agree though that if there's one thing that's going to limit the growth of the quantum industry, it is that problem. How do you translate a problem statement into a program that can run on a quantum computer? As someone once said, let's say you're an aerospace engineering and you're an expert in computational fluid dynamics, how you going to translate to quantum. I don't think there's anyone in the world who would be an expert in both computational fluid dynamics and quantum computing at the same time. So in order to solve that, there are a number of approaches. The first is indeed software. The software needs to be able to abstract some of the details I know Classiq is doing a lot of that.

Some is with consultants, there's a fairly significant consulting group, a lot of consulting companies that will do that type of work. A lot of software companies do that work as well as I would say, the classical business consulting firms will do that. There are other companies that are continuing to develop software. I would call application specific quantum software. I'll give an example like Multiverse, where you can take a financial problem and you can put your problem statement in a spreadsheet and then they'll take a spreadsheet and convert that to a quantum program.

There's even other companies that will take a program that's written in a the language like C and convert that to a quantum program. So in both those two latter cases, someone wouldn't even need to know quantum at all. The only issue is there's still very, very early the number of different use cases where you can do that is still only a handful. But that's going to be critical to develop more of that type of software that helps make it easier for people to make use of a quantum computer. That's going to be critical for the growth of the industry.

Yuval: As we get close to the end of our conversation. I wanted to ask you about governments again, let's say you were advising the US government and maybe you are. I don't know. What would you advise the US government to do to make sure that the US remains super competitive or leading the quantum revolution?

Doug: Well, first I want to say that they've done a lot and they've done some very, very good things. They were the early supporters of some of the research five, 10 years ago. But more recently they're doing a lot with regard to developing the quantum workforce. There's multiple programs, internships, those types of things, just to, if you read through the quantum act, the 1.3 billion that was passed a couple years ago, workforce development was a key factor. So at least in the US, they're doing a lot and I know the governments in many of the other countries are doing a lot too in terms of workforce development and encouraging the industry. But I will make what I call an off the wall suggestion if they're listening.

That is when I looked at how the semiconductor industry really took off 60 years ago, the reason it took off was because of the Apollo space program, where they wanted to put a man on the moon and they needed to take a computer, which was basically the size of a conference room and shrink it down to like a 70 pound box that could go in the lunar module. So they funded the development of integrated circuits and they bought millions and millions of dollars’ worth of integrated circuits from companies like Fairchild Summit Conductor and Texas Instruments. So I tell people that nothing motivates a company like getting a purchase order. It's an off the wall suggestion, but one thing that certainly would help is to buy lots and lots of quantum technology of various types. A lot of it is computer time, as well as quantum sensors, even quantum communication, those types of things. That I think could be a good supplement to what they're already doing.

Yuval: Related to that, I was at the University of Maryland Quantum Investment Summit a few weeks ago. A lot of speakers were worried about regulation, that all of a sudden the Senate's going to come in and regulate quantum or the sales and that will cause big problems. How much are you worried about that?

Doug: Well, I'm actually more worried about any restrictions with regard to immigration. I just said a second ago, I think the workforce development is one of the key factors in there. Again, if you look towards history, one of the things that really helped the semiconductor, even the computer industry is that we had a pretty open immigration system. So students from foreign countries could come here they could study and get their graduate degrees or even undergraduate degrees. Then they decide to stay here. They like the US so much. They could find a company. 

One story I would say is I worked many years ago, I worked for Intel. My first day at Intel, my boss took me around to meet people like they typically do. He made a comment working at Intel is like working at the United Nations. So I do believe that being able to attract people from other countries with more opportunities, of course, if you're interested in quantum is, would be very, very helpful for the development. With regards to sales, I think the US just needs to be an exporting country. We import a lot of materials, but we need to have a good balance of payments and we do that by exporting services and those types of things. So we certainly want to have a balance between national security type of things, but the US still has a lot of friends around the world and we need to be open with them and try to work with them as much as we can on quantum technology.

Yuval: That's excellent. So, Doug, how can people get in touch with you to learn more about your work?

Doug: Okay, well, we have a website, the website is quantumcomputingreport.com. You can read many of our pages, as I said, we have hundreds of pages of material there. You can learn about us. We also publish a newsletter once a week. It comes out on Sundays. Basically what it does is it summarizes all the new items that we've added to the website over the past week. So I did that because I realized that people in the industry are very, very busy and they can't take the time to try to go through all the various feeds to find out all the new announcements. So I make it simple for them.

They can just get the newsletter on Sunday. Then when have a few minutes of spare time, you're not in daily meetings, they can look at it and they can just click on the items that they're interested in. So those are the ways and you certainly can get in touch with me. My email is dfinke@quantumcomputingreport.com and I'll always willing to work with people. Part of our mission is to help people accelerate the quantum industry. Hopefully we've done some of that and can want to do more.

Yuval: I think your newsletter is very helpful to me personally, and I think you're doing a great service for the industry as well. So thank you so much for that. Thank you for joining me today.

Doug: Okay. Thank-you. Was nice talking with you.