On the Record with Elliot MacGowan, COO of Agnostiq
Hometown: Montreal
Hobbies: Reading, listening to podcasts, golfing, and working out (cycling, spinning, yoga, going to the gym)
3 words to describe yourself:
Curious—this has helped me actively learn about new technologies and apply my business knowledge to them. I am not afraid to go deep on the tech - you have to be curious as a founder.
Persistent—I was not always the smartest in the class, but I was always one of the hardest working.
Even keeled —even when things get stressful, I’m able to manage a level of calm
What were you like as a child and how do you think that shaped who you are today?
I was always a bit nerdy at heart. I played a lot of sports growing up, so people probably labeled me more as an athlete than a “nerd.” When I was a young teenager, though, I used to build websites using Flash and HTML and actually ended up selling those websites. This was a pretty common story shared by other founders of my generation, as websites were the hottest things growing up. I also played a lot of video games, and, in general, I was always interested in new technologies. I started following crypto early—around 2011/12—and actually entered into a pitch competition to try to start a crypto company with a friend, but no one understood it at the time. By the time it came to picking a technology to start an actual business, I wanted to go as deep as possible, so I found quantum.
Are there any life mottos you live by?
The one thing that this startup experience has taught me is that not only do you have to be incredibly resilient, but you have to be really open to feedback and change. As a stubborn MBA student, that’s not something I came off the shelf with, but start-up life has really taught me to put my ego aside and always be open to change. It’s really the only guarantee in this world.
What led you to start Agnostiq?
Unlike many who go to business school to acquire new knowledge, I went to hit reset on my career and find my way into start-ups. I was focused on getting out of school and starting a company. In fact, I actually picked the University of Toronto because of their Creative Destruction Lab program which is an incubator/accelerator program that matches companies with MBA students. I was always watching the quantum tech companies coming through and thought that’s where I could have the greatest impact. I felt like I had missed the first real machine learning wave while I was working and in school, and that a lot of the “low hanging fruit” had already been captured, so I really saw quantum as the next big opportunity. Honing in on that, I looked for a good, compelling business case for quantum computing while also looking for cofounders. I was very active in the Toronto ecosystem, and I actually met Oktay at one of the industry meet-ups. We hit it off right away, and the more Oktay revealed his personality and background, the more I realized I absolutely needed to work with him. I actually even met his wife and kids the same evening.
If I’m reading your resume correctly, it looks like you were relatively young when you started Agnostiq—and still an MBA student—can you talk more about that? Pros/cons? Hindsight?
I was relatively young. I was 27 and finishing my 2nd year of business school when we decided to really go all-in with Agnostiq. At that point, I had completely deprioritized school to the point where I had to ask one of my professors to give me a passing grade on a project so that I could actually graduate on time. It was hard balancing both school and starting a company, but it had become clear to me after 6 months of my 2-year program that this is what I needed to be doing, so I wasn’t going to stop. I would have even dropped out if necessary.
Have you ever needed to change or pivot the business based on product feedback, product-market fit, or product building?
For sure—we’ve gone through at least 1 major pivot as a company. We actually started out as a security company, building encryption tools for the quantum cloud. We shifted out of that business of course, not because we don’t believe in it, but because there was no market for it at the time. When we first went to market with the encryption products, we realized that it just wasn’t a problem the customers had. Quantum computers were too new and too difficult to program to even worry about security at that stage. So we went back to the drawing board, feedback in hand, and deprioritized security in favor of more applications-oriented software. We are now focused on building tools and infrastructure to help customers build their own applications more easily. Security may come again in the future but it really is all about timing - as such a young, nascent industry as a whole, customer needs are a moving target so you need to be adaptive.
What has been your biggest or most unexpected challenge?
The unexpected comes in waves—just as you start to feel comfortable, you move up to the next level and then there are more unexpected challenges waiting for you. But, if I had to pick one I would have to say team building and managing culture. As a first-time founder, I assumed my biggest challenge would be raising money and that most of the problems would get solved afterward, which is really not true. The biggest challenge was - and still is - recruiting and building a world-class team while preserving a great culture. It is a constant, iterative process - one that we all invest a lot of time and energy into.
Okay, now on to the tech questions—how do you see quantum computing integrating with classic computing and other HPC techniques?
Internally, we think about quantum computing as becoming one of “n” devices that you will use within a computationally intensive workflow. You can think about a machine learning (ML) problem, which involves multiple steps such as data preparation and preprocessing, optimization, classification, etc - quantum computing will take a portion of it (optimization), while perhaps a generic CPU will handle the preprocessing and then finally the classification could be done by a GPU. Generally speaking, we see it as fitting in within a broader high-performance computing portfolio. We will likely never be all the way quantum or all the way classical—you’ll have many different devices that are very good at a subset of tasks.
When do you see customers using quantum computing for mission-critical production tasks?
Probably sooner than most people think since quantum is at the end of the day a subroutine within a broader classical workflow - meaning the customer doesn't need to go ‘all-in’ on quantum to productionalize it. Customers can slowly integrate it now under the assumption that their investments will bear fruit as quantum computers scale up.
Where are the best quantum computing educational programs and are there enough quantum programmers being trained?
There are not enough quantum programmers being trained, however many organizations like QIC in Canada, and QED-C in the States are working to change that. If you look at an industry job board, for instance, you’ll find approximately 400 open positions for quantum tech in North America that have been sitting unfilled for approximately 150-200 days. We’ve estimated that there are roughly 1,000 qualified applications in North America, and they’re all employed—meaning we’re running at a deficit of around 800 people per year.
The real problem is that, as of right now, you need a PhD to do this work, and PhD programs in physics are not graduating enough students in the US and Canada. We need master’s programs that can give students the fundamental skills for a career in quantum. It obviously wouldn’t be as in-depth as a PhD program—it would be a bit more applied—but it would be applicable and practical. These programs are starting to pop up, like at the University of Toronto and some other universities, but we need more.
What are your thoughts on superconducting vs. trap ion?
Each of them has its own benefits and trade-offs. It’s not necessarily that one is better than the other, but that one may be better than the other for a particular task. To give an example: trapped ion quantum computing might be a lot better for things like ML. This is just speculation, but it’s looking likely. On the other hand, superconducting quantum computing may be very good at solving really large-scale optimization problems. It’s still too early to say what’s going to be good at what, but the consensus is that each of them will be better than the other at something pretty specific.
As you look to the future, what are you looking forward to?
It’s going to be really interesting to see how quantum computing as an industry unfolds. I’m very excited to see it reach the point of maturity where more people are using quantum computers. I’m especially excited to see it because it’s actually a way to reduce the amount of energy consumption that traditional high-performance computers use. This is a big problem right now as more and more companies require more substantial computing capabilities, which takes an incredible toll on the environment. One of the hidden benefits of quantum at scale is that quantum computers are much more resource-efficient than any of the other systems. What a traditional super computer would essentially take an entire town’s energy consumption to do, a quantum computer could do with the same level of energy as a few houses at most.
Related News
