A POWERFUL quantum computer could crack encryption and solve problems that classical machines find impossible. Although no one has yet succeeded in building such a device, recently we have seen a gathering pace of progress – so could 2022 be the year?
At the moment, efforts are focused on an important milestone known as quantum supremacy: the point at which a quantum computer is able to complete a calculation that a classical computer can’t, given a reasonable amount of time.
Google was first to reach this goal in 2019 using a device with 54 qubits, the quantum equivalent of regular computing bits, to perform an essentially useless computation known as a random sampling calculation. In 2021, a team at the University of Science and Technology of China solved a more complex sampling problem using 56 qubits, and later pushed it even further with 60 qubits.
But Bob Sutor at IBM says this game of leapfrog is an academic achievement that is yet to have much real impact. True supremacy will only be achieved when a quantum computer is significantly better than classical ones and is capable of solving different problems, rather than the random sampling calculations currently used as benchmarks.
He says IBM is working towards “quantum business advantage” – the point at which a quantum computer can solve genuinely useful problems for researchers or companies significantly faster than classical computers can. Sutor says this hasn’t arrived yet and won’t during 2022, but can be expected within the decade.
Nir Minerbi, co-founder of quantum software company Classiq, is more optimistic. He believes that 2022 will see a demonstration of quantum supremacy in a useful problem.
“Remember when the first electric cars came out? They were useful to drive to the grocery store, but perhaps not to drive 300 miles to drop your kid off at college. Like electric cars, quantum computers will get better and better over time, making them useful in a wider range of applications,” he says.
There are a number of hurdles to solving practical problems. The first is that devices need thousands more qubits to do so, and these must also be more stable and reliable than existing ones. It is likely that researchers will need to group them together in clumps to work as a single “logical qubit”. This helps with fidelity, but will gnaw away at improvements to scale: thousands of logical qubits may require millions of physical qubits.
“Quantum computers will get better over time, becoming useful in a range of applications”
Researchers are also working on quantum error correction to fix glitches when they occur. Google announced in July 2021 that its Sycamore processor was able to detect and fix errors in its superconducting qubits, but the additional hardware needed to do this introduced more errors than it fixed. Researchers at the Joint Quantum Institute in Maryland later managed to pass through that crucial break-even threshold with their trapped-ion qubits.
Even so, it is early days. Scott Aaronson at the University of Texas at Austin says it would be “pretty shocking” if a general-purpose quantum computer solved a useful problem in 2022. “Error correction is just now starting to work, and we don’t even seem near the point of protecting a single encoded qubit for an arbitrary amount of time, let alone doing computations on thousands or millions of encoded qubits,” he says.
More on these topics: