Will quantum computing disrupt any industries that matter, and how soon?
If quantum computers ever work well enough to be trusted for general purposes by users outside of academia, they will need to become reliable. Making a device that depends on quantum mechanics reliable is not all that much unlike taming a herd of wildebeest. It's not, at least analytically speaking, impossible.
You'd think, though, that high probabilities of reliability, of accuracy, and of resilience would be necessary to bring about any technology that purports to offer 'digital transformation' -- a concept that implies not only a shift from state A to state B, but a considerable distance between the two. So the biggest emerging theme from the recent Inside Quantum Technology Europe 2020 conference, based in the Netherlands but held online this year for obvious reasons, could be summed up with the question every parent driver hears shouted from the back seat: Are we there yet?
Building that first quantum application
William Zeng, Head of Quantum Research, Goldman Sachs.
"We need to have a first application, before [QC] can be ubiquitous," commented William Zeng, the head of quantum research at financial giant Goldman Sachs. "As far as we can tell now, in order to get to the first application, we're going to need error correction of some kind. In order to have error correction at scale, we're going to need to scale up the imperfect qubit systems that we have today. There's enough milestones that it's not going to happen in the next year. I think the way to track it is to look at it from a user perspective -- to say, 'What is the series of milestones that need to occur in order to get there?' And however long it takes, we'll find out."
A quantum computer (QC) is a completely different kind of computing mechanism -- not an evolved semiconductor, or even a system that uses symbolic, binary logic. It's a machine that assembles atoms in a structure so that their quantum-mechanical properties may be leveraged to execute tremendously long, repetitive, analog-style algorithms. But those atoms will eventually fall apart, and even before they do, they'll occasionally get their results wrong.
So a calculation that could quite literally require centuries of continuous work from the fastest semiconductor-based supercomputers, might require a few seconds worth of work from a QC -- if you're willing to accept the small probability that the result might be wrong. That is if, and only if, QC engineers successfully tackle the current problems of system instability, and -- even in a state of relative stability --manageability.
Nevertheless, QC sounds like as transformative an idea in technology as any ever conceived -- as the vacuum tube, the transistor, or the 'off' switch. And even before it attains this mythical status, it could have a cataclysmic effect on every industry whose digital transactions are protected by cryptography.
SEE: What classic software developers need to know about quantum computing (TechRepublic)
Banking comes to mind right away. But consider also, mobile communications. All of mobile communications.
Axel Foery, Executive VP Quantum-Safe Security, ID Quantique.
"If quantum technology — in this case, quantum random number generation [QRNG] — hits the mass market, it's time to start," remarked Axel Foery, executive vice president for quantum-safe security at Swiss quantum security firm ID Quantique. In 2001, ID Quantique produced the first commercial QRNG device.
While you'd be lucky to fit a full-scale QC in the back of a pickup truck, a QRNG is a chip small enough to fit inside a smartphone. It's not a logic processor, but an optical mechanism -- like a child's rattle, except with photons. The photons rattle around inside, doing whatever photons do. And that's actually the point: Since there's no rhyme nor reason to it, the process can be leveraged in snapshots that reveal truly random numbers -- rather than algorithmically seeded values that seem random enough, at least until the conditions of the algorithm can be replicated.
Random number generation is essential for encrypted communications -- or, as it's coming to be known, communications. Encrypted transactions that use QRNG-generated keys are called 'quantum safe'. It's a real, reliable technology that 5G Wireless is putting to use today.
Andrew Shields, Head of Quantum Technology at Toshiba Europe.
"In the end, we have to seek quantum-safe communications everywhere," remarked Andrew Shields, head of quantum technology at Toshiba Europe. "It doesn't make sense if there's some part of the network which doesn't have quantum-safe communications."
Cybersecurity and cryptography
Perhaps within the decade, the capability for QC to render modern public-key cryptography pointless will transition its state, like an orbiting electron, from theory to reality. What makes cryptosystems viable today is how long it takes any algorithm to decode them, even if it were to take smart guesses rather than brute-force approaches. Once a single reliable quantum system comes online, and is usable through a cloud-based service, the time required to break any code would become trivially small. Even encrypted video chats could conceivably be revealed in near-real-time.
It's fair to say a technology transforms the way you do business, if it kicks your very foundation out from under you. A quantum-key encryption device (QKD) would go several steps beyond a QRNG, producing a channel for two and only two parties that leverages another bizarre quantum property: If interfered with in any way, rather than cracking, the channel would cease to exist.
A few years back, analysts believed the necessity of a true QKD device alone would be incentive enough to invest in QC's future. That ended up not happening. So now, QC needs at least one more raison d'être, perhaps to serve as a catalyst for all the rest -- with those catalyzed applications hopefully including the one that protects digital transactions in the post-public-key-cryptography era.
Chemistry, optimization and simulation
Robert Sutor, vice president for quantum ecosystem development at IBM Research, offered his perspectives on which applications, in his view, were the best candidates -- more specific, he said, less general:
- Chemistry, including the simulation of physical systems and the interactions between their components. For example, the production of ammonia-based fertilizers consumes as much as 2 percent of the world's energy. A quantum simulation could seek a more energy-efficient method. Also, more efficient conversion of carbon dioxide into hydrocarbons. "The more that we can do chemistry inside a computer," said Sutor, "as exactly as possible, there's going to be less time spent doing chemistry in the laboratory."
- Optimization and simulation, which is the category that a financial services institution -- historically, the earliest investor in a computing infrastructure -- may pay handsomely just to pioneer.
- Matrix algebra, whose immediate benefactor would be machine learning systems. Consider that the part of ML that 'learns' a pattern doesn't have to be the same part that memorizes it. Today's ML systems are trained repetitively; theoretically at least, a QC could run all its repetitions in parallel, simultaneously. It would not remember what it learned (that's not physically possible), but if the results could be rendered onto a conventional electronic memory, that might not matter.
SEE: What is the quantum internet? Everything you need to know about the weird future of quantum networks
Robert Sutor, IBM Research, VP, IBM Quantum Ecosystem Development.
"The way I would see it," said Sutor, "by the end of the decade, we'll start to see little, scattered examples. And then there'll be this reassessment: Scientists, software engineers, and [other] people will say, 'What do I have, and how did I do it?' And there will be improvement...Then people will be really clever, in terms of just how they're packaging the system, how they're combining it with classical systems. The hardware itself is going to be changing -- a lot of our classical components will be moving much closer to quantum computing, which will change the algorithms. Our compilers will be getting better. So there will be a second wave, then there will be a spreading."
Second waves, in any technological endeavor, take time to build potential. Think of how long mobile communications had to evolve before 3G Wireless finally came about. Now apply that logic to QC, whose evolutionary timeline would put us at a time analogous to the briefcase era of mobile transmitters. 'Tomorrow', such as it is with any quantum science, won't be tomorrow.
"We're not looking for quantum to be slightly better," said IBM's Sutor. "We're looking for quantum to be really better."