Google has solved a major problem with quantum computing. Have they effectively broken encryption going forward? Is bitcoin going to be ok? Huge implications for the future
https://blog.google/technology/research/google-willow-quantum-chip/
Google has solved a major problem with quantum computing. Have they effectively broken encryption going forward? Is bitcoin going to be ok? Huge implications for the future
https://blog.google/technology/research/google-willow-quantum-chip/
10 Comments
Submission statement:
From the article
“The first is that Willow can reduce errors exponentially as we scale up using more qubits. This cracks a key challenge in quantum error correction that the field has pursued for almost 30 years.
Second, Willow performed a standard benchmark computation in under five minutes that would take one of today’s fastest supercomputers 10 septillion (that is, 1025) years — a number that vastly exceeds the age of the Universe.”
Big implications for crypto, encryption, privacy, network security going forward.
what shall we do with the most powerful computer in the world?… benchmark it
This problem may be solved faster by building the physical arrangements of the gates, every one, and seeing what the outcome is. It would probably be faster to do that than to use a classical computer.
And yes, we’d need many universes of material to actually do that, but it’s not infinite arrangements, it has an upper limit.
The problem they’re solving lends itself to massive parallelization and quantum computers are perfect for that. The issue is, like any solution, we have to prove it’s the right one, and not a close approximation, or just jibberish.
Does this mean bitcoin could become worthless in the future?
I wanted to see what the problem was so I followed the links to https://research.google/blog/validating-random-circuit-sampling-as-a-benchmark-for-measuring-quantum-progress/
And I’ll admit I don’t understand what it means or why it would take supercomputers so long to randomly sample the circuit
For some context, the problem they are talking about here is called Random Circuit Sampling. It is not practically useful for anything, it is designed specifically to give the greatest possible advantage to quantum computers just to demonstrate that they are actually doing something that classical computers can’t.
The problem goes like this: create a completely random quantum circuit and then sample an output from running that circuit on a quantum computer. So for a quantum computer you just… do that. But for a classical computer there is no great way to simulate an arbitrary quantum circuit that doesn’t have any particular structure so it will by default be very, very slow.
Besides being practically useless, another problem with this approach is that it is essentially impossible to verify that the output of your quantum computer is correct. You just have to run it on small circuits that you *can* simulate first, check that it is working, and then assume that it keeps working when you scale up to more qubits.
Anyway, this is not to down on Google they have made a ton of progress here, but the sensationalist headline stuff oh my god we calculated this thing that takes a trillion years or whatever is not actually very helpful at explaining what they have done, because it is not a calculation that anyone really needs done in the first place. And the calculations we actually would like to do still can’t be done on this computer.
well dont hold back – how much does your mom weigh ?
“Have they effectively broken encryption going forward?” – No. They beat a very specific benchmark that is on one hand impressive, but the benchmark is kind of built to be the one specific problem where quantum computers are much more efficient than classical. Beating this problem does not necessarily mean anything for real-world problems. And to reiterate – quantum computers DO NOT solve problems by trying all solutions and finding the right one. (the closest thing to that is Grover’s algorithm, but that is not really world-shattering, it just gives you a nice speedup) There are some cryptographic primitives that are vulnerable to quantum computers, mostly public key schemes that are not built on lattice assumptions, but quantum computers do not help you that much with breaking symmetric encryption, finding preimages of hash functions, breaking anything built on lattices and so on. (okay, Grover’s algorithm gives you some speedup, but nothing you can’t protect against by doubling your security parameter)
Don’t get me wrong, getting better results at quantum error correction right is a step towards more effective quantum computers, but to frame this as “cryptography now broken” is nonsense.
Some people always throw shade on technology demonstrations like this.
Don’t forget that Michael Faraday discovered “electricity” in 1831 and it wasn’t until 1879 that Edison perfected the lightbulb. Solving these quantum problems now will yield fruits in the next few decades.
Would they kindly factor 191,287?
I would like to see verifiable progress.