骋辞辞驳濒别鈥檚 claim to quantum supremacy is far from the end of the road. The firm has just edged in front, but there are still plenty of chicanes, banks and ditches to halt the competitors in the race to build the first truly useful quantum computer.
As well as Google, research groups at IBM, Intel and Microsoft, as well as a slew of start-ups, are all developing quantum computers. And the competition is heating up. IBM has said that 骋辞辞驳濒别鈥檚 claims of quantum supremacy aren鈥檛 as good as they appear and there are still plenty of milestones left for others to claim.
The next few years will probably see each group attempting to outdo the others on tasks that are particularly hard for classical computers. They have all backed different ways of making a quantum computer, so we may eventually see one method outpacing the rest.
Advertisement
骋辞辞驳濒别鈥檚 advances have come using superconducting qubits: components made of superconducting wire laid in an insulator. There was already quite a lot known about this approach before Google started experimenting with it, which has helped the firm get this far.
But superconducting qubits are large and hard to make smaller. At 50 qubits, the size is manageable, but quantum computers may need a thousand or a million qubits to become really useful.
Intel is working both on superconducting qubits and spin qubits for its quantum computer. Spin qubits work by manipulating the orientation of electrons using microwave pulses. These can be operated at slightly higher temperatures and may be simpler to integrate with current electronics. This makes it easier to increase their number and make them smaller, says Jim Clarke at Intel.
Microsoft has decided to take yet another route. It aims to use topological qubits, which rely on quasiparticles: packets of energy that behave like a particle. 鈥淭he catch is that it requires making a new form of matter that鈥檚 never been seen before,鈥� says Scott Aaronson at the University of Texas at Austin.
There is some indirect evidence that these quasiparticles can be made, although nobody has done it yet. The benefit of topological qubits is that they can naturally correct some of the errors that creep into quantum computers. 鈥淓rror correction is an even bigger hurdle than supremacy, but it鈥檚 the obvious next goal,鈥� says Aaronson.
That goal may be years away. John Martinis at Google said in a press conference on 23 October that its team will experiment with error correction experiments over the next year or two. During that time, it also plans to build a quantum computer with 1000 qubits.
So what next? We can expect that as superconducting qubits improve, the other types of qubits will also be made into processors that can achieve quantum supremacy. A decade or two down the line, we may see quantum computers that can perform useful calculations, such as modelling chemical reactions or interactions in materials science.
But that won鈥檛 mean that most people will get their own quantum computer any time soon. 鈥淨uantum computers are going to be located either in the cloud or in large supercomputing centres,鈥� says Clarke. 鈥淢ost people won鈥檛 use them directly. But well after I retire 40 to 50 years from now, I can鈥檛 rule anything out.鈥�
