Digital Platforms and Services

Google reignites the ‘quantum supremacy’ debate – again

By Martyn Warwick

Jul 21, 2023

  • The problem solved by its Sycamore device had no real purpose or real-world application, but does confer bragging rights
  • Quantum computers can solve some problems much faster than a classical binary supercomputer but...
  • Classical computers will always co-exist and complement quantum devices
  • ‘Utility’ quantum computing will not be practicable until error-correction is possible

It was late 2019 when Google first claimed to have achieved “supremacy” in quantum computing. The controversial assertion was politely questioned by some scientists, publicly pooh-poohed by others and downright derided by its arch rival, IBM. According to Big Blue, the problem set for the experiment by Google's quantum computer, dubbed Sycamore, was comparatively simple and could have been solved by a classical digital computer, although admittedly over a considerably much longer timeframe. Indeed, as things stand, any computational problem can be solved by a classical computer given sufficient time to complete the task – but that could be thousands of years and thus would be intrinsically worthless.  

And that’s the issue: In any computational problem in quantum computing, “quantum supremacy” – a term first coined in 2012 and bandied about boastfully by various companies at every opportunity ever since – refers to the repeatable demonstration by a programmable quantum device of the solution to a problem no classical computer, of any size, could ever solve in any practically feasible amount of time. In other words, coming up with a verifiably correct answer in less than a period of ’n’ where ’n’ can be anything between a human generation or less, or tens of millennia and longer.

Critics say the very term ‘quantum supremacy’ infers a state of complete ascendancy of the technology and its ability to solve a specific, but not necessarily relevant or meaningful, problem faster than a classical computer, and thus is too gung-ho and triumphalist. Another, and less pejorative, term is ‘quantum advantage’, which is rather more subtle than the notion of ‘supremacy’, because it implies that a quantum computer can solve a real-world problem better than any binary-based classical computer. 

TelecomTV has been talking about quantum supremacy and other quantum computing matters with the author and scientist, Lawrence Gasman, erstwhile senior fellow in telecommunications at the Washington DC-headquartered thinktank, the Cato Institute, and the founder and president of research and consultancy house Inside Quantum Technology.

He’s a man who knows his quantum onions (as it were) and who argues that quantum supremacy “is a very unpleasant term… quantum advantage is less aggressive.” And so it is, but that hasn’t stopped Google coming back four years later to claim, for a second time, that it has achieved it via the latest iteration of its Sycamore quantum processor, which runs at 70 qubits rather than the 53 qubits of the machine used in the 2019 test.

As the addition of new qubits, the basic building blocks of quantum computers, exponentially increase the powers of quantum devices, the latest version of Sycamore is 241 times more powerful than the Sycamore of 2019. This, Google says, puts it “beyond capabilities of existing classical supercomputers”.

The Google paper “Phase Transition in Random Circuit Sampling”, which is published on the open access science website ArXiv, claims that Hewlett Packard’s  Frontier supercomputer would take 47.2 years to perform the same calculation and produce the same result that the new Sycamore did practically instantaneously.

Despite the fact that the latest Google test had no function other than serve as a demonstration to support a claim to quantum supremacy, and did nothing of practical use or apply the sort of high-level error correction that will be vital to permit quantum computers to provide answers to meaningful real-world problems, as an exercise it is flashily impressive in much the same way as a display of fireworks – entertaining but expensive and achieving little of lasting value or importance.

Or, as Gasman has it, “Quantum supremacy could mean we have constructed a problem that nobody gives a ‘bleep’ about, but we can show that the problem can’t be solved in any reasonable period of time with a classical computer, but can be solved with a quantum computer, and that shows something.”

He added, “Researchers view quantum supremacy as primarily a scientific goal, with relatively little immediate bearing on the future commercial viability of quantum computing. Due to unpredictable possible improvements in classical computers and algorithms, quantum ‘supremacy’ may be temporary or unstable, placing possible achievements under significant scrutiny.”

Horses for courses

Today’s quantum computers are ‘noisy’, error-prone and fault intolerant and it is evident that the future of computing won’t be a one-horse race but a matter of ‘horses for courses’. Quantum machines will be of immense importance and value, but they can’t and won’t be able to do everything. The power and speed of classical computers continue to increase and each time a new record for one of them is claimed it is quickly superseded by another.

Thus, for eight months last year, between March and November 2022, Hewlett Packard Enterprise’s (HPE’s) “Frontier” exascale supercomputer was the fastest on earth. Then, in November, it was overtaken by the “Flatiron” supercomputer, which is owned by the internal research division of the Simons Foundation and services five centres for computational science. You can bet that won’t be the world’s fastest classical computer for very long either.

What is for certain is that for many computer applications, the classic computer will not only continue to co-exist with quantum computing but also dominate it in many application areas. Quantum computing, because of its dependency on concepts of quantum physics and other technological limitations, will, in the foreseeable future, be limited mostly to certain areas of application such as artificial intelligence (AI), chemical engineering, cryptography, finance, logistics, materials science, modelling weather systems, pharmaceutical research and drug discovery, as well as subatomic physics and simulation of quantum and other systems and problems.

This is because quantum computers are excellent at solving complex problems via statistical approximation and optimisation. The structure and properties of qubits allows them to manage such processes much faster and more efficiently than classical computers – and that’s because they process data in a fundamentally different way than binary computers.

In classical computing, a ‘bit' comprises the set values of a ‘zero’ or a ‘one’ but in quantum computing the quantum property of superposition is exploited simultaneously to store data in qubits at an undetermined value that can be a zero or one any combination of the two. The ability to hold two bits of information at the same time means a quantum computer can work on multiple problems simultaneously. It’s also a property that allows the storage of vast amounts of data in a much smaller space than is possible for a classical binary computer, and store them in a much shorter time.

What is needed is the move to what is being called “utility quantum computing”, where devices with thousands of qubits will provide solutions to some pressing real-world problems at a much faster speed than any classical computer could ever achieve. However, that will be dependent on quantum machines being built with sufficient redundancy to allow larger and larger numbers of qubits to work together instantaneously to correct the propagation of errors that, currently and inherently, inhibit the devices from working for long enough to solve real-world problems before breaking down and halting the running of a program.

Quantum computers are extremely sensitive to the noise and errors caused by interactions with their environment, and it might be something as simple as a stray ray of light. As a program advances, errors accumulate and degrade the computation. Quantum systems do not lend themselves well to being in a definite state, even for a very short period of time, and thus a quantum system will very quickly de-phase, de-cohere, entangle and collapse.

Most quantum scientists regard error correction as being the central problem that must be solved before practical utility quantum computers can be built and operated.

In the second part of this report, TelecomTV will cover the main challenges now facing quantum computing and some of proposed solutions to them, and look at what Gasman considers to be the looming threat to public key encryption (PKE) – the asymmetric cryptography that has, hitherto, kept systems and networks (including the world's telecoms networks) secure for a generation past but maybe not for much longer.

And finally, a quick quantum joke: Schrodinger goes out for a Sunday drive down a country lane. The car raises a cloud of dust and provokes the suspicions of a traffic cop. He pulls over Schrodinger, looks at his driving licence and insurance details and then asks for the trunk to be popped. Schrodinger pulls the release, the cop looks in, then walks back round the car and says, “Sir, are you aware you have a dead cat in your trunk?” Schrodinger replies, “Well, I wasn’t, but I am now.”

- Martyn Warwick, Editor in Chief, TelecomTV

Cookies

TelecomTV uses cookies and third-party tools to provide functionality, personalise your visit, monitor and improve our content, and show relevant adverts.