(Vice) Everyone agrees that Google’s results are important in terms of scientific and engineering advancement. Putting claims of supremacy aside, the key innovation is Google’s superconducting processor is programmable, so it can receive instructions for different tasks, some of which may be beyond the reach of classical computers.
A major concern about quantum computing in general is that our current security and encryption protocols will no longer be safe from ultra-fast and powerful machines. But the fact that Google’s Sycamore can perform one specific task that people believe would take too long for a classical computer, does not mean it can perform the tasks it would need to in order to crack important security protocols.
Scott Aaronson, a quantum professor at The University of Texas (Austin), explained. Current RSA encryption can only be broken by “several thousand logical qubits.” Current quantum computers have fewer than 100 qubits at most, and Sycamore had even fewer. Canadian company D-Wave’s machines use more, but using an entirely different approach without quantum gates.
Some academics are questioning the experiment’s methodology. Gil Kalai, a mathematics professor at the Hebrew University of Jerusalem, pointed out in a blog post that Google’s research fails to show that the quantum computer is actually sampling the intended probability distribution and not doing something else entirely to arrive at the output.