It has been making quite a buzz in the last few days in the global news. It all started on 20 September 2019 with an article published in The Financial Times, a British economic and financial daily. One might be surprised that it is in this journal that a supposed revolutionary breakthrough in the field of mythical quantum computers is announced if it were not for Google. Above all, it has been said for years that these computers could easily break the codes on our credit cards, not to mention the fact that they could revolutionize scientific computing, for example, to develop entirely new drugs.
In this case, The Financial Times reports a document written by Eleanor Rieffel, now missing from a NASA site but of which a complete copy can be found with the title Quantum Supremacy Using a Programmable Superconducting Processor, concluding that the Quantum AI Lab research team, led by John Martinis, would have eventually reached the Graal of Quantum Supremacy.
Would John Martinis have operated a quantum computer – which is not a universal programmable quantum computer that can potentially run any algorithm of reasonable size for its memory and computing capacity – to perform in about 200 seconds a calculation that would require about 10,000 years on a conventional supercomputer, of the kind available to NASA and which can serve as a point of comparison?
Sycamore, a calculator and not a quantum computer
Google’s quantum computer would be called Sycamore and would have more than 50 qubits (54 precisely and 53 would have been used). He would specialize in the simulation of what is called a random quantum circuit capable of handling 50 qubits. A random quantum circuit is an assembly of a random choice of quantum logic gates. All this remains to be confirmed, and for the time being, neither Google nor NASA is responding to any media messages on this subject, invalidating or confirming anything as a result.
If the results announced are real, a significant step forward would have been taken, but we would still be a long way from a real revolution. It should be remembered that it is always difficult to combat the phenomenon of quantum decoherence, even with corrective codes, which quickly makes calculations carried out with a relatively large number of qubits imperfect or even impossible. However, it would take a lot to beat conventional computers in the majority of the tasks where they are used.
Moreover, we know that there is no guarantee that a quantum algorithm to solve a particular problem with a quantum computer, which is specialized in its execution, will not one day be dethroned, in terms of the computing speed it allows, by a new classical algorithm. It’s happened before.