Quantum computing is just getting going. But the hype could bring everything crashing down
For many scientists working in the field, the keen interest that investors and CIOs are taking in quantum computing is a double-edged sword.
The idea that quantum computers will transform business and usher in a new era of unprecedented computing power is increasingly making its way into executive pitches as a marker of forward-thinking and innovation, with the technology often touted as the new must-have that could deliver a competitive edge.
But for many scientists working in the field, the keen interest that investors and CIOs are taking in quantum computing is a double-edged sword. While quantum computers eventually need to move out of labs and into businesses, the technology's commercialisation might be happening too soon, they warn, running the risk of relegating quantum computing to the much-dreaded 'over-hyped' category, along with virtual reality, blockchain or NFTs.
It's not that quantum computing isn't interesting. From a scientific perspective, it's hugely exciting -- which is why research has been ongoing in the field for decades.
In the early 90s, scientists were already getting excited about the idea of using quantum mechanics to build next-generation computers. This is because it had been observed that when particles in their smallest, quantum, state behave very differently to the way the laws of classical physics dictate.
For example, quantum particles can exist in various different states at the same time, in a sort of dual reality. That property, imagined scientists at the time, could be leveraged in the context of computing, with quantum particles able to carry different data in parallel, instead of being restricted, like the classical computer bit, to either a one or a zero. The idea of the quantum bit, or qubit, was born.
Armed with qubits, a computer could theoretically tackle hugely complex problems in no time, since different calculations could be carried out simultaneously in multiple parallel 'realities'.
"In principle, we've known as a community since the early 90s that quantum computing can solve problems that are hard for classical computers," said Bill Fefferman, assistant professor in the computer science department at the University of Chicago. "Those were theoretical results -- no experiments came with that. We were just saying that in principle, if a perfect quantum computer was ever built, it could do these things."
Fast-forward to the present day, and we are now seeing early prototypes of small-scale quantum computers -- systems that can control a small number of qubits, although usually not any more than 100 or so. The most powerful quantum machines built by IBM, for example, which is one of the most prominent investors in the field, currently boast 65 qubits.
With such few qubits, there is very little that quantum computers can actually do: researchers estimate that up to one million qubits, and potentially even more, would be necessary to build the perfect quantum system that engineers were dreaming up in the 90s. But scientists can still experiment with today's small-scale machines to hypothesise how things might turn out once the technology is more advanced, and the results they are seeing so far seem promising.
Chemical engineers, for example, are anticipating that quantum computers will be able to simulate large and complex molecules to predict the combinations that will best fight off disease in order to create life-saving drugs much faster; banking giants are counting on quantum systems to determine the best stocks to buy and sell for maximum return, based on calculations that can account for many more fast-changing factors; and car manufacturers are testing how the technology could revolutionise the design of batteries, the optimisation of supply chains or the management of traffic in dense, urban settings.
These early experiments are generating a lot of enthusiasm across fields that range from oil and gas to logistics, through cybersecurity, agriculture and even weather forecasting. Every single industry, some experts claim, is set to be transformed by the technology once it reaches maturity.
"Early experiments hint that this technology will hold the promise to solve very interesting problems that cannot be solved classically," says Fefferman.
It hasn't taken long for investors to take note. The quantum computing industry is flourishing, largely driven by deep-pocketed tech giants IBM and Google, who were among the first sizeable companies to take interest in the technology. They have now been joined by Amazon and Microsoft, which have both launched their own quantum programs, as well as scores of smaller companies that combined saw a total investment of $1.02 billion just this year.
There are now nearly 200 quantum computing start-ups on the market, offering services in quantum software and hardware, and promising huge business improvements once the quantum revolution kicks in. The first publicly-traded firm dedicated to quantum computers, IonQ, was announced earlier this year in a $2 billion deal. Quantum roadmaps are multiplying, from IBM's 1,121-qubit system set to be released by 2023 to PsiQuantum's promise of a million qubits by 2025.
But some experts are now expressing doubts about the viability of this industry. For Sabine Hossenfelder, a researcher in theoretical physics at the Frankfurt Institute for Advanced Studies, the quantum computing industry is experiencing a bubble in the making -- and the consequences could be greatly detrimental to research.
"I work in basic research, and from my perspective all of the early applications of quantum are super exciting," Hossenfelder tells ZDNet. "But a lot of the stuff I read is unreasonably optimistic."
"The risk that I see is that you have all these investors who like the idea that soon enough we'll have a great quantum computer and we will make money with it because we'll solve all these problems -- but in five years or so they will realise these promises didn't pan out. Then they will pull out and it will be really hard to continue, even on the research side, because the bubble will dramatically deflate."
For Hossenfelder, the problem is mostly to do with the timeline. Hitting the one million qubit mark is a huge technical challenge, given the current less-than-100-qubit state-of-the-art. It's not only about successfully creating and controlling more qubits: engineers also have to think about ways to reduce the space needed to fit all of the equipment that's necessary to run the system. Current quantum computers already fill rooms-worth of machinery and tools; making the devices orders of magnitude larger with current technologies is simply unrealistic.
The next few years won't bring all of the technical solutions to these problems, argues Hossenfelder, comparing the challenge to building a modern-day PC a century ago and equipped only with wood.
And even if they do, even if there is a team working in secret on a new approach that could solve all of the existing bottlenecks in the next five to 10 years, it is far from certain that quantum computers will beat classical computers on all fronts. Quite the contrary: quantum systems are expected to be transformative for specific use cases, particularly simulation, but it's unlikely that they will be replacing our current laptops anytime soon.
"We have to be careful that we talk about them in an accurate way," says Fefferman. "Quantum computers are not a panacea; they won't be able to speed up all problems. There will be certain problems which even 30 or 100 years from now, when a perfect quantum computer exists, it won't be good at solving."
There is every indication that classical computers are here to stay, and that they will still be used for many tasks -- if not most of them. Quantum computers, in contrast, will be more of a special-purpose device that will generate extreme speed-ups on a set of very specific problems. What scientists are doing today, is trying to discover what exactly these problems might be.
So, where does the quantum computing industry hype start -- and end? "The high-level answer is there's definitely a fine
line," says Fefferman. There are many research teams with legitimate goals, as well as many businesses developing products that could be game-changing. But there are also a lot of companies riding the system and selling what has become known as "quantum snake oil".
Fefferman is not alone in warning against the unrealistic expectations that are being set for quantum computing. Computer scientist Scott Aaronson, for example, is a prominent critic of this bubble in the making, who writes in his blog that the call is now coming from inside the house, meaning that quantum scientists themselves are worrying about the proportions that the quantum field is prematurely reaching.
Hype isn't fundamentally bad for quantum computing: the industry needs commercial interest if it's to leave the realm of academia and achieve the dream born in the 90s. The danger lies in creating impossible-to-reach expectations too quickly -- in fact, in creating them at all. For scientists working in the field, this is only making the prospect of a 'quantum winter' all-too-imminent. The promise of quantum computing is certainly real; but realising it will require patience, and a strong degree of accountability.