I'm like 99% sure Google said "what's the most contrived workload that technically counts as quantum supremacy" and went from there...
That being said, we certainly live in interesting times.
Quantum Supremacy Using a Programmable Superconducting Processor
21–30 of 71 posts
Re: Quantum Supremacy Using a Programmable Superconducting Processor
#22Reading as somebody who is not in the field, and writing this down to read the responses of those who are more qualified: How I understand it: the "computation" is actually "sampling" the cubits? And then sampling results in the sets of random numbers, which don't have uniform but some specific distribution (specific for quantum effects). Then they claim that such a distribution could not be achieved using classical…
There's additional piece to this that you are missing: while the classical simulation is _hard_ (i.e. increasingly hard for larger and larger number of qubits or circuit depths), it _can_ be computed with enough computational power for small number of qubits <=50 and depths<=20. (To get the classical simulation to 50 qubits was an achievement in itself and required a huge amount of compute power provided by google, w…
Therefore the win here is that they have verified that a 50 qubit quantum computer "works" (success rate > 10^-3 > 0.) No classical computer could verify a 100 qubit quantum computer in the same way (but you could verify various 50 qubit subsections of the 100 qubit quantum computer). Alternatively, if they had two such 50 qubit devices (and presumably they will soon) they can verify the second one with the first much quicker than they can verify the first with a classical computer, so in that sense the quantum computer beats a classical computer at some task.
Re: Quantum Supremacy Using a Programmable Superconducting Processor
#23How does performance compare with an ASIC specialised at generating random numbers? Also what is the error rate?
The ASIC is really only relevant for quantum annealers like D-Wave, not for a QC like this. Even if a classical ASIC could close the gap, it would only be delaying the inevitable.
Error is very high, their probability of success drops very fast, see the paper.
Re: Quantum Supremacy Using a Programmable Superconducting Processor
#24What use will quantum computers have? Can they do things other than factoring large numbers quickly?
They're pretty good at simulating molecules and molecular interactions. I think chemical engineering is probably the killer app for quantum computers.
How would that work? Is there an implementation on a simulator to see how it would work?
Re: Quantum Supremacy Using a Programmable Superconducting Processor
#25Reading as somebody who is not in the field, and writing this down to read the responses of those who are more qualified: How I understand it: the "computation" is actually "sampling" the cubits? And then sampling results in the sets of random numbers, which don't have uniform but some specific distribution (specific for quantum effects). Then they claim that such a distribution could not be achieved using classical…
A noise distribution is easy to simulate.
If you had some chip that took input and gave reproducible output that was impossible to simulate classically, I think you'd actually have something of interest.
Re: Quantum Supremacy Using a Programmable Superconducting Processor
#26To me this reads a lot like “quantum computer faster than a classical computer at ‘simulating’ itself”. Constructing a physical experiment where it would take a classical computer 10000+ years to simulate the outcome is trivial. I don’t see how this is very different.
Re: Quantum Supremacy Using a Programmable Superconducting Processor
#27I wonder if we will be able to replicate this work on IBM’s 53-qubit machine that’s about to hop onto the cloud.
Re: Quantum Supremacy Using a Programmable Superconducting Processor
#28"However, realizing the full promise of quantum computing (e.g. Shor’s algorithm for factoring) still requires technical leaps to engineer fault-tolerant logical qubits"
This is NOT how the media has chosen to report on this. It's always quite frustrating to figure out how a paper diverges from the press it gets.
Re: Quantum Supremacy Using a Programmable Superconducting Processor
#29Earlier quoted context omitted.
There's additional piece to this that you are missing: while the classical simulation is _hard_ (i.e. increasingly hard for larger and larger number of qubits or circuit depths), it _can_ be computed with enough computational power for small number of qubits <=50 and depths<=20. (To get the classical simulation to 50 qubits was an achievement in itself and required a huge amount of compute power provided by google, w…
Therefore the win here is that they have verified that a 50 qubit quantum computer "works" (success rate > 10^-3 > 0.) No classical computer could verify a 100 qubit quantum computer in the same way (but you could verify various 50 qubit subsections of the 100 qubit quantum computer). Alternatively, if they had two such 50 qubit devices (and presumably they will soon) they can verify the second one with the first muc…
you lost me at "presumably they will" ...
Re: Quantum Supremacy Using a Programmable Superconducting Processor
#30I'm like 99% sure Google said "what's the most contrived workload that technically counts as quantum supremacy" and went from there... That being said, we certainly live in interesting times.
Even if it's highly contrived, it's still something that they found a task for which the quantum computer is superior. It points the way towards a more general superiority where some at least have doubted the possibility.