Some 40 years ago now, the brilliant Nobel Prize-winning physicist Richard Feynman implored humanity to build the first quantum computer.
That famous conference of 50 top thinkers hosted jointly by Massachusetts Institute of Technology and IBM, gave birth to Feynman’s most famous quote: “Nature isn’t classical, dammit, and if you want to make a simulation of nature, you’d better make it quantum mechanical, and by golly it’s a wonderful problem, because it doesn’t look so easy.”
It certainly hasn’t been easy but everyday quantum computing is gradually moving towards becoming a reality, and now two of the top voices in the space are insisting the time has come for the dawning of the quantum revolution to become more widely embraced by enterprise and by society at large.
A couple of years ago, quantum computing properly emerged into the mainstream consciousness as Google’s Sycamore quantum computer performed a mathematically-designed calculation that it would take the world’s most powerful supercomputer 10,000 years to do in less than five minutes. Effectively, this made the Sycamore machine, which completed the task in 200 seconds, 158 million-times faster than the world’s fastest supercomputer.
The potential is staggering and quantum computing is in the hugely exciting stage of innovation where so many contrasting approaches and methodologies are being formulated, invested in and designed and tried out, all with the aim of accelerating the arrival of the time when quantum computers become the norm. But in order to get to that point more quickly, Zapata Computing CTO and Co-Founder Yudong Cao insists that there needs to be a change in mindset from enterprises.
“People are still thinking that ‘no, quantum computing is early, we need to do the research first, and then we go to production’,” he says. “But I think, only by going to production, that you can really identify the bottlenecks, and really identify what truly are the things that matter.
“Can you use that as a forcing function? It’s something I believe is unique about this paradigm, which is that we not only care about the science, but also we want to go to production and then use the customer challenges as a forcing function that drives our priorities, and research and engineering and everything else.
“At Zapata, we believe that there are fundamental software engineering issues, in many cases, even IT issues that need to be resolved before we can begin thinking about quantum computing. This, I would say, is really the rubber meets the road moment.”
Zapata is very much focused on the software side of the space. The company’s quantum platform Orquestra combines a powerful software platform and quantum algorithm libraries to deliver advances in computational power for which can be applied in a real-world context.
On the hardware side, there are multiple modalities at play – some that are very much real today, some that are more speculative and with an eye on the future.
One of these approaches is cold atom quantum computing. This is where atoms are cooled to a few millionths of a degree above absolute zero, and they then take on quantum properties. Lasers are used to arrange the atoms, hold them in place, run computations on them, and read out the results. Quantum calculations, communications, and sensing are the result.
That is the approach adopted by ColdQuanta, whose President Paul Lipman shares the view of Cao, whom he calls “one of the smartest technical minds in the quantum software are” that a change in mindset is needed from the business world when thinking about the quantum space.
“There’s a kind of mystique and romance around this concept of quantum, it’s almost like in the same category as time travel,” says Lipman. “There have even been TV shows where the quantum computer is kind of this, this all-seeing, all-knowing machine that can travel back and forward in time. We need to challenge that general kind of almost science fiction nature of it.”
Lipman points to the example of AI, and how it moved from being thought of in somewhat of a similar way a few decades ago as an example of how the thinking can change around something after a fallow period.
“If you go back to the 80s or the 90s, there was this AI winter where there was a great promise of AI but it never panned out and the overall level of interest waned,” Lipman continues.“But for the AI revolution to then actually come to pass, it was a combination of three factors. Firstly, the fact that the hardware evolved, and we had these enormously powerful GPUs that could do matrix multiplication on a great scale or linear algebra on a great scale.
“Secondly, the algorithms really evolved with the emergence of deep learning and, then, number three, there was this great abundance of data with the explosion of the internet. And now you could actually train these models.
“The net result of those three factors was kind of the AI revolution we see today. I think there’s going to be a similar revolution that happens in quantum. You’ve got the evolution of the hardware, a massive evolution on the algorithm side, and then, I think it’s about the use cases. It’s the early adopters who are exploring, ‘Okay, how do we use a quantum computer to solve a problem in chemistry? How do we use a quantum computer to optimise a machine learning process? How do we use a quantum computer to do financial simulation?
“I think it’s those three factors in parallel, which is why you’re seeing so much partnering going on between the hardware companies and software companies and businesses. And the net result of those three will herald the quantum revolution.”
There’s definitely a healthy mix of competition and coopetition in the space but how do you move the needle in terms of changing the thinking of enterprises to bring forth that revolution more quickly?
Lipman believes it will probably become part of the workflow, rather than replacing the workflow in the first instance before improvements will see it take a more prominent role in ways we can’t even probably conceive yet.
Cao, whose eyes light up at the mention of quantum computing’s potential, believes that one key issue needs to be tackled before anything else.
“What we have noticed about our enterprise customers is that it doesn’t matter what industry they’re in, they ultimately face the problem of complexity and I think that’s true for most of the problems humanity is facing,” he says. “That sort of scale and complexity, not only manifests in the sort of sheer amount of resources needed – how many people or computers are needed – but also the sort of inherent mathematical intractability of the problem.
“A common flavour of these complex problems that arose in enterprises is when, for example, you want to optimise a logistic system, you typically want to sort of maximise a particular target, but then at the same time, you have all these conflicting constraints that are in your way. This is what makes a problem hard.
“Those are incredibly complex problems, especially when you get to a very large scale, and no computer, no algorithms are known to solve these general problems efficiently.
“The complexity that’s manifested in the enterprises and also probably in humanity at large, often can be translated into these inherently intractable mathematical problems. That is the place where enterprises, or the world, will look for any computational solutions that offer even an incremental improvement.”
The key question that arises then is: how do you build the link between quantum technology and correlation and addressing that complexity?
“One of top contenders, I think, is AI, as it provides absolutely the shortest path for bridging that gap,” Cao suggests. “Because there are many machine learning techniques which essentially use some source, some sort of statistical power to infer correlations in data and also to generate new data points.
“If you can then imagine there’s a back-end quantum computer that supplies statistics that are driven by the superposition and entanglement. and then at the front end, you can produce points in a data space, you can reach points in a data space that are otherwise would have been very hard to reach with a classical solution.
“Then you can take that ability to navigate the data space, further into optimisation. Let’s say, for example, a quantum-driven solution can discover a logistic supply chain design that was previously not uncovered by any classical algorithms; or quantum computers may be able to discover a drug molecule structure that wasn’t previously attempted.
“The good thing is that we already have the legacy of the past 20 years of all these AI systems and algorithms that we can build on top of so I think AI is absolutely the shortest path to make an impact in the real world with these sort of large-scale problems.”
Lipman, on the other hand, believes that it will probably be an amalgam of different approaches which helps to lead to the point where businesses are regularly exploiting quantum computing technology.
“I was talking to somebody the other day, and they were asking should enterprises be sitting on the sidelines waiting for things to settle out in terms of which approach to quantum computing will ‘win’?” Lipman says. “And I told them that’s like imagining your business in the 70s, waiting to see whether DEC or VAX, in terms of which computer standard, will win out.
“The reality is, we probably can’t even foresee what things will look like 10 years down the line and there may well be different kinds of quantum computers that prove to be most useful for certain kinds of applications.
“Then the next step is bringing that to the question of which enterprise applications will be useful and will help define how this will play out. The way the industry seems to be guiding businesses is: now is the time to get involved. It’s not that you’re going to flip a switch at some point and all of a sudden, we’re going to be at this concept of quantum advantage (Quantum Advantage is the concept whereby a quantum algorithm will have exponential speed advantage over a classical algorithm).
“One of the big misconceptions about quantum computing is there’s going to be this point in time and the switch will get flipped and, all of a sudden, we’re in the world of quantum advantage. That’s just not how it’s going to play out. It’s going to happen in different ways, with different use cases across different industries. It’s going to be a gradual process and it’s going to happen incrementally.”
Another important aspect in bringing quantum computing to the forefront is ensuring that people know what it is, and how the machines work. This is an area which has seen a lot of growth in the past couple of years in particular, and Lipman is heartened by that.
“One of the things that’s going to be necessary for this industry to continue to thrive is a talented and educated workforce,” he says. “And we’re seeing quantum education and the emergence of quantum computing programmes in universities, and even sometimes at high school level, and it’s very gratifying.
“There’s the promise of a really kind of transformational new way of computing. It’s not that we’re going to have faster video games or better calculators or, but it’s the potential for solving a class of problems that are out of our reach today, and the impact that could have on the world – when we think about things like climate change, material science, drug discovery and probably a whole range of things we can’t even imagine.
“I think that’s critically important for us to educate people to have a healthy and thriving industry. And it’s not just on a technical side, but also business people and even the general public needs to understand what this technology is. They don’t need to understand the science necessarily, but they need to understand what these technologies are about, what their applications may be, what the challenges are and what the risks are, because it’s going to affect all of us eventually.”
