In 2019, the majority of practical problems conventional computers can easily solve would leave quantum computers sputtering in the dust of their own decoherence. Quantum computers are far from useful when compared to classical computers. It may be important to ground oneself in this reality before attempting to verify the validity of a startling recent claim by tech juggernaut Google: that they've achieved quantum supremacy.
In a 63-page paper entitled "Quantum supremacy using a programmable superconducting processor," the Mountain View-based corporation performed a complicated physics experiment using a 54 (though technically 53 because one qubit isn't working) qubit quantum computer code-named "Sycamore." In the paper, Google claims that its Sycamore quantum system correctly verified results generated by the system's output distribution generated from a random quantum circuit of specific form. The test is known as "random quantum circuit sampling."
Google claims that Sycamore performed this task in 200 seconds, while asserting that today's fastest supercomputer, IBM's Summit, would take 10,000 years to perform the same task. IBM was quick to refute Google's claims of quantum supremacy with a paper published on Monday called "Leveraging Secondary Storage to Simulate Deep 54-qubit Sycamore Circuits." IBM's main refutation is that Google did not correctly factor in that Summit could be configured in a way that allows it to keep up with Google's Sycamore. In their paper, IBM wrote that Summit could simulate the quantum system and perform the random quantum circuit sampling calculation in 2.5 days if correctly configured, which is a long way from Google's 10,000-year estimation.
A Pioneering Experiment on the Entanglement Frontier
Quantum information science is largely funded by the unproven assertion that classical computer systems lack the capability to emulate highly entangled quantum systems. Google has achieved something remarkable, but whether it’s actually quantum supremacy depends on how you define it.
There are two terms to consider: quantum advantage and quantum supremacy.
Quantum advantage means that a quantum computer system can perform computations many orders of magnitude faster than a classical computer. But quantum supremacy is accomplished only when quantum systems have computing power so great that they can perform calculations that conventional supercomputers could not.
Google's Sycamore quantum computer solved a problem with quantum algorithms that is essentially a super-polynomial speedup relative to the same operation performed on conventional computers. In their paper, Google laid claim to the achievement of quantum supremacy by indicating that IBM's Summit supercomputer would take 10,000 years to finish the problem their Sycamore accomplished in 200 seconds.
Given IBM's refutation that it would only take 2.5 days for their Summit computer to simulate the quantum computing system designed by Google, it seems more likely that Google has achieved quantum advantage in this particular niche computing case, rather than quantum supremacy.
Quantum advantage is still a remarkable achievement given the unstable nature and ultra-high degree of complex engineering needed to make quantum computers perform at all. If IBM's assertions are correct that Google did not account for an optimized configuration of IBM's Summit supercomputer, then quantum supremacy was not achieved.