Nature Electronics: moving quantum computing from engineered advantage to practical utility

[Scott98]

a short comment piece by He-Liang Huang of the Henan Key Laboratory of Quantum Information and Cryptography in Zhengzhou, China. The title sums up the argument well: moving quantum computing from engineered advantage to practical utility. Most of the piece is behind a paywall, but the main point is clear from the opening. It is only three pages long

Huang's argument is that quantum computers have now shown advantages over classical machines, but mostly on tasks designed specifically to show that advantage. Random circuit sampling and similar benchmarks prove the hardware can do something hard. The challenge now is turning those carefully controlled demonstrations into calculations that deliver real practical value

This fits neatly with the IBM story we discussed recently, where a 19 second run was estimated to take a supercomputer 110 years. That was impressive, but the task itself had no real world use. Almost every quantum advantage claim so far has been in the same category. Useful results are what investors and governments will eventually demand

The obvious candidates for practical use are chemistry, materials science and certain optimisation problems. Simulating molecules is something quantum computers should naturally be good at, since nature itself is quantum. But getting there needs far more reliable qubits, working error correction and algorithms that beat the best classical methods on problems people care about

I think this is a healthy message from inside the field. Hype has run ahead of reality for years, and a push towards useful results is what the industry needs. It is a message worth repeating. What do you think will be the first practical problem a quantum computer solves better than a classical one?

Moving quantum computing from engineered advantage to practical utility - Nature Electronicswww.nature.com

[Golden Dan]

The paywall is frustrating for a comment piece. Opinions on the direction of a whole field should be open to everyone

Still, the title tells you most of what you need. It matches what a lot of researchers have been saying privately. Useful results matter more than record numbers now. Funding bodies will increasingly ask for them

[Dragon95]

The gatekeeper is error correction. Without it, qubits are too noisy for long calculations, and most useful algorithms need long calculations. Several groups have shown logical qubits that beat physical ones, which is progress. Scaling that up is the job for the next few years. Once it works, practical utility follows much faster

[SlayedRebellion]

Engineered advantage is a nice honest phrase. It admits that the benchmarks were chosen because quantum machines would win them. There is nothing wrong with that as a first step

But the public reads these headlines as if quantum computers are already better at everything. Comments like this help correct that

[HangmanPage_WCW]

Chemistry will be first in my view. Catalysts and battery materials are the problems most often mentioned by people who work on this. Even a modest improvement in simulating a catalyst could be worth a fortune to industry. That gives companies a clear reason to invest

[Badger90]

Classical algorithms keep improving too, which makes the target move. Every time quantum looks close to a useful win, someone finds a smarter classical method. That race is not going to end soon

[PixelBarrel]

Optimisation gets mentioned a lot but I am sceptical. Classical heuristics are very good at those problems already. The gains may be too small to justify the cost

[Arty Leah]

Seeing this in Nature Electronics is a good sign. It signals the field maturing