Scott Buchholz, global leader for Deloitte Consulting’s quantum computing efforts, is interviewed by Yuval Boger. Scott explains how Deloitte helps clients understand quantum technology and its potential applications, particularly in solving business problems. They discuss the timelines for quantum computing becoming commercially useful, referencing industry roadmaps. He emphasizes Deloitte’s role in bridging the gap between technical jargon and business needs, helping clients navigate vendor selection, shares insights into the learning curve associated with quantum computing, highlights Deloitte’s work in developing quantum-inspired machine learning solutions, and much more.
Transcript
Yuval Boger: Hello, Scott. Thank you for joining me today.
Scott Buchholtz: Yuval, it’s a pleasure to be here. Thanks for having me.
Yuval: Who are you and what do you do?
Scott: Well, I work for Deloitte Consulting and I lead our efforts to explore quantum computing. For us, what that means is we’re trying to understand what the hardware and software that others are building and providing can do to help ourselves and our clients solve problems, business problems of interest.
Yuval: At what stage do customers come to you? And you mentioned computing, but do they also say, well, I have a quantum networking or a quantum sensing or a quantum security issue?
Scott: We get a little bit of everything. It’s a little bit the virtue of working for the world’s largest professional services firm. I would say, though, that the majority of questions that we get tend to either be some version of, I’ve heard about this quantum computing thing and I would like some help understanding what I need to know, or we’ve been looking at this quantum computing thing and we’d like to know Deloitte’s perspective and what you’re seeing. We also get questions about cyber risk as it relates to sort of the future threat from quantum computers and what people should do and be thinking about. And occasionally people come to us because they’ve read our research or other things and they know that we are doing interesting work and they’d like to work with us on interesting problems.
Yuval: Is there a particular segment that you see more than others? Government or commercial pharma or finance or something else?
Scott: I would say we tend to see more interest from finance, pharma, chemicals and the government. And I would also say that interest tends to ebb and flow over time.
Yuval: One of the questions that I get a lot, but now I have you to answer it for me, is when. When will quantum computers be truly useful for a particular industry? What is your standard answer there?
Scott: Well, Yuval, my standard answer is I refer people to the roadmaps of our partners. So IBM is on record saying that they believe they will be able to do something of value sometime next year. I think they’re still working to qualify what something of value looks like. Um, and as you know, there are a whole slew of other vendors. I think most people will be able to solve a problem of commercial value. Timelines start somewhere in the next two to three years, give or take. And that’s not every commercially interesting problem. That’s the initial commercially interesting problems.
Scott: And then we’ll see where we go from there.
Yuval: Some vendors, some hardware vendors will say, well, our computer is so complex and the software environment is still immature, that you really need to work with the people that design the computer to make the most out of it. Where does Deloitte come into the picture in these situations where the customer just wants to work with the hardware vendor?
Scott: I would say a variety of ways we actually work with folks. So sometimes what the hardware vendors find is if you’re a startup in the hardware space, your board, as a general rule, would prefer that you focus on your hardware and not on delivering services. And we are happy to help hardware vendors with the services problems that arise. I would also say, you know, part of the challenge sometimes is that there’s a language that businesses in which most business executives understand value, they understand return on investment, they understand problems and other things. And sometimes it’s actually helpful to have an organization like ours who is accustomed to translating between the technical and business terms, language and concepts to actually be able to help both sides get along better.
Yuval: You find that customers come to you to help with vendor selection as well. Which computer should I use or which cloud I should use?
Scott: We do get vendor selection problems. I have no favorite children, just to be clear.
Yuval: Could you tell me about an example engagement, something that maybe you completed and you can talk about just to get a sense of how it progressed. You know, what was the question, what was delivered, what was the outcome?
Scott: Sure. One example that comes to mind is we were called by an organization whose CIO was really interested in learning about quantum computing and who wanted to know, in essence, when was their organization likely to be impacted, where were they likely to be impacted, when were they likely to be impacted, and how should they think about getting ready? And what we helped them with, for instance, was devising a roadmap that would actually help them think about the who, what, when, where, why and how around getting started, next steps, sort of the plan over the next couple of years so that they have triggers that they can wait for. In their case, that would say now is time to take the plan off the shelf, dust it off, and here are the things to try with some sequence in order. I would also say though, that we get calls sometimes to help with proofs of concept. So people would like to understand what it takes to actually run a problem on a quantum computer and will help work with the customers and typically their chosen vendor, hardware vendor or software vendor. And we’ll actually work with everybody to try to figure out what’s a problem that we can collectively scope down to something that we’re comfortable the hardware and software can handle that will demonstrate enough value to the client that they’ll actually get to the end and be happy with it. And that serves as a useful place to start when clients think about, okay, what is it that I need to do next? What do I need to think about next? What do I need? Where do I need to go from here? In essence.
Yuval: Do people start from Quantum or do they start from the problem? I mean, so starting from Quantum with me, and I’ve, I’ve heard about Quantum, what does it mean for me? How am I impacting starting from the problem, I’ve got this scheduling problem or what have you, you know, what’s the best solution that Deloitte can recommend? And should we also explore Quantum? Maybe you get a little bit of both. But what’s the most common thing at the moment?
Scott: We tend to get more. We tend to get more. We believe Quantum’s the answer. Help us figure out the question. Most, many, many commercial enterprises are focused on the next quarter. They are not necessarily asking to explore what might be the optimal solution in a couple of years. And as a result, that happens occasionally. But candidly, most of the time what people are asking is, hey, not how will I be able to solve my problem in the future, but I have a problem right now.
Scott: My foot hurts. Can you make it, can you help me make it better?
Yuval: Do you also introduce quantum inspired solutions in that discussion?
Scott: We do. One of the things that we’ve been working on is we have for instance, a classical solution, which means we run a solution that runs on CPUs and GPUs that is quantum inspired. It uses the fundamental principles of quantum computing and quantum mechanics to actually perform its work. What it’s doing is a process in machine learning called feature engineering. Basically you can think of it as a pre processing step before you feed the data into your machine learning algorithm. And it turns out that by doing this feature engineering step, we can dramatically improve the outcome of quite a few machine learning models. It actually started in fraud detection, payment fraud detection as an algorithm, but turns out that it has applicability elsewhere. The nice thing about something like that, where you have a quantum inspired algorithm, is not only can you slot it into your existing machine learning pipeline with a minimal amount of fuss, but also that along the way you don’t just get better outcomes today, but you also start being able to learn.
Scott: What does it actually take to work with a quantum computer? Because in order to understand how to tune the engine, you actually have to understand a little bit about how you would actually tune the engine if it were running on a quantum computer itself, as opposed to on today’s CPUs and GPUs.
Yuval: And as far as you lead quantum globally for Deloitte, or are there different sorts of geographical centers?
Scott: Explaining our internal organization is as much fun as explaining any large organization. But let’s just say I help coordinate things globally. And we tend to have centers in Australia, Japan, Germany, India, and a handful of other places in Europe.
Yuval: You’ve been doing this for a while. What have you learned about Quantum in the last year that perhaps you didn’t know before that?
Scott: I find that every year brings new advances and every year brings new frustrations. I sometimes there was an analogy called dolphining where you think about the dolphins who jump out of the water to breathe air and then go down for a while under the water and come out and go down and up and down. And sometimes I feel like my enthusiasm for how soon quantum computing will be useful is a little bit like watching a dolphin. I get really excited, and then sometimes I get disappointed, and then I get excited again and disappointed. I’m. I’m still on the optimistic side at the moment, so I’m hoping that nobody comes along and rains on my parade anytime soon.
Yuval: What would be disappointing events? I mean, hardware is making progress. Some algorithms are getting discovered or are improving. What could dampen your enthusiasm?
Scott: What would dampen my enthusiasm is some bright light demonstrating that things are a couple orders of magnitude harder than we all thought, and suddenly the timelines go out another couple of years all over again. If you’re asking me if I think it’s likely, I don’t think it’s likely. But I’ve been wrong before. If.
Yuval: If I were a customer in a. In a Fortune 500 organization, and your analysis showed that Quantum could be beneficial to me, say, in three years, when should I get started?
Scott: Actually, sooner rather than later. The reason is part of what’s going on is what we’re finding, I think collectively is that it can take a year or two for people to build some level of proficiency in using quantum computers. Some people tell me that feels like a lot. I just remind people that’s about the amount of time it takes to make a data scientist proficient. And this is about as different as working with classical computing as data science is. And with data science, you know, there are a bunch of APIs you can drop some data in and get an answer out. Nothing about that tells you whether or not you’ve done the right thing. You can start with quantum computing, but if you assume that it’s going to be like everything you’ve done before, you’re likely to be quite disappointed.
Scott: And as a result, if you say, well, it’s likely to take two years to identify the right people to get them the training to devote them to the process for a couple of years, the learning, the journey, the et cetera, right? Developing proficiency. And you think about our earlier part of the conversation where people’s timelines for Quantum are really two to three years. If you’re a Fortune 500 organization, it’s useful to think about that activity and getting started at staffing that team and working with the vendors and other things as a little bit like a call option with the idea that if Quantum goes to infinity, like we all hope, then you’ll actually be well on the way and well on the journey.
Yuval: If you were recruiting, maybe you are. But if you are recruiting to the Deloitte Quantum team, what’s the pitch? Why should someone come to work in Deloitte on quantum computing?
Scott: Other than that we’re the best group of people to work with and work for? Gosh, what would I, beyond that, what would I tell people? What I would tell people is we actually work at the intersection of practicality and research. And so where those two things come together wind up being really interesting problems. So how do you apply all of the research and all of the things you know to be able to solve problems that people have? We’re not the best place for people who want to do pure research, but we are a really great place for people who want to apply what they know to real world problems to understand how business interacts with technology, technology as it advances.
Yuval: You probably heard me ask that question before. But hypothetically, if you could have dinner with one of the greats, dead or alive, who would that person be?
Scott: You know, I think it would be fun to have dinner with Feynman in his prime. He’s always struck me as being a bit of a character, so that would be fun. I like characters.
Yuval: Wonderful. Scott, thank you so much for being here today.
Scott: Yuval, it’s been a pleasure. Thank you for having me.