Google claims to have proved its supremacy with new quantum computer • Hacker News Board

Re: Google claims to have proved its supremacy with new quantum computer

#61

Post by wholinator2 » Mon, Jul 03, 2023, 5:33 PM UTC

hutzlibu wrote: post #42

CrampusDestrus wrote: post #32

Earlier quoted context omitted.

When you break a stick you just break a stick. When you simulate breaking a stick you know everything about the broken stick. What if you broke the stick to study its material properties? Then you would need to spend months carefully taking samples and measuring all the broken spots. With a perfect simulation? You're done the moment it ends. All the data is there available with a copy paste

"With a perfect simulation? You're done the moment it ends. All the data is there available with a copy paste" As far as I understand, to perfectly simulate reality, you would need a second reality.

Well of course the simulation is a simulation. They are speaking of the ideal case. A real simulation is an application of a model, the physics of the molecules and structures interacting. They are saying that in the ideal case, once the simulation is complete you have complete knowledge of the entire system at every moment in time. Of course there's a resolution to all this. The time steps are comparatively large to the planck time and the wave functions are approximations, everything is an approximation, that's what makes it a simulation.

But the principle is that within your model, once the simulation is complete you know _everything_. You have measured every aspect of reality possible from your simplified version. If your model was arbitrarily close to reality you really would know everything

Re: Google claims to have proved its supremacy with new quantum computer

#62

Post by BasedAnon » Mon, Jul 03, 2023, 5:34 PM UTC

zdragnar wrote: post #27

dataflow wrote: post #7

Dumb question: Say I have a wooden stick and I break it in half in less than a second. Assume a computer would need several minutes to simulate everything that would've happened in the stick. I clearly got the output faster than a computer (and with more precision), so does this imply I'm doing anything particularly fascinating? I assume the same scenario is possible to concoct for a quantum computer. I assume it wou…

The difference between you breaking a stick and the computer modeling it is that you've measured nothing. You don't know, with any precision, the amount of force you used, the rate the stick broke at, how much mass remains in the two pieces and how much was lost to splintering, etc. In other words, assuming the computer model has sufficiently accurate data as an input, it can produce significantly more refined output…

Now if we had a way of measuring it, it would be interesting to ask what sort of model of computation we could derive from the breaking of sticks

Re: Google claims to have proved its supremacy with new quantum computer

#63

Post by OkayPhysicist » Mon, Jul 03, 2023, 5:40 PM UTC

mikewarot wrote: post #2

The cited paper[1] references "random circuit sampling" which is defined in [2] which then gets so heavy into abstract math, and I give up. Can someone explain this in terms an EE or programmer can understand? [1] https://arxiv.org/abs/2304.11119 [2] https://arxiv.org/abs/2007.07872

[2] is actually a phenomenally easy read for a physics paper, but it does assume an undergraduate physics degree's level of background.

First, it's important to know that a quantum state (written as |letter>) is not directly observable. We can think of it as some vector, for which we can only apply an operation to to get out a scalar observable.

A quantum circuit is a series of operations on a quantum state to create a different quantum state, which can be modeled as a matrix U, which is unitary (meaning U*U=1).

The goal of "sampling a random quantum circuit" is trying to find the probability of observing some given observable if you keep applying random quantum circuits to a given input state.

Re: Google claims to have proved its supremacy with new quantum computer

#64

Post by weard_beard » Mon, Jul 03, 2023, 5:45 PM UTC

mistermann wrote: post #59

wholinator2 wrote: post #55

Earlier quoted context omitted.

What are these capabilities of which you speak? Or have i missed the sarcasm

Well, generate reality for one. It's not a trivial detail, but it tends to be dismissed or taken for granted (or downvoted lol). Someone may venture into this territory some day, and perhaps that someone will find some travelling companions to make the journey more exciting and productive....time will tell!

The trick is you can only flip as yet unobserved bits. If we know the asteroid is coming we can't delete it.

Re: Google claims to have proved its supremacy with new quantum computer

#65

Post by OkayPhysicist » Mon, Jul 03, 2023, 5:46 PM UTC

pclmulqdq wrote: post #50

mikewarot wrote: post #2

The cited paper[1] references "random circuit sampling" which is defined in [2] which then gets so heavy into abstract math, and I give up. Can someone explain this in terms an EE or programmer can understand? [1] https://arxiv.org/abs/2304.11119 [2] https://arxiv.org/abs/2007.07872

> Can someone explain this in terms an EE or programmer can understand? Unfortunately not. The physicists who are behind quantum computing don't think the same way, and go straight for the abstract math to solve any problem. I am pretty sure they don't understand that most of our progress on computing up to this point is because you don't need to go into Galois fields or discrete math to describe what a computer or a…

The second paper shared by the GP is actually just about as dumbed down as you can feasible get. The problem is that A) it uses some basic jargon, because it's fair to assume the only people interested in the result are people who understand the jargon and B) it's describing a completely impractical problem, designed to maximize the difference in performance between a quantum and classical computer rather than do anything particularly useful.

Re: Google claims to have proved its supremacy with new quantum computer

#66

Post by jcims » Mon, Jul 03, 2023, 5:49 PM UTC

dataflow wrote: post #7

Dumb question: Say I have a wooden stick and I break it in half in less than a second. Assume a computer would need several minutes to simulate everything that would've happened in the stick. I clearly got the output faster than a computer (and with more precision), so does this imply I'm doing anything particularly fascinating? I assume the same scenario is possible to concoct for a quantum computer. I assume it wou…

I wonder about that frequently. The universe 'executes' physics in, as far as I can tell, a realtime basis (at least within the local reference frame). What is that called as compared to computing a model of the same.

Re: Google claims to have proved its supremacy with new quantum computer

#67

Post by dataflow » Mon, Jul 03, 2023, 5:54 PM UTC

zdragnar wrote: post #27

dataflow wrote: post #7

Dumb question: Say I have a wooden stick and I break it in half in less than a second. Assume a computer would need several minutes to simulate everything that would've happened in the stick. I clearly got the output faster than a computer (and with more precision), so does this imply I'm doing anything particularly fascinating? I assume the same scenario is possible to concoct for a quantum computer. I assume it wou…

The difference between you breaking a stick and the computer modeling it is that you've measured nothing. You don't know, with any precision, the amount of force you used, the rate the stick broke at, how much mass remains in the two pieces and how much was lost to splintering, etc. In other words, assuming the computer model has sufficiently accurate data as an input, it can produce significantly more refined output…

> You don't know, with any precision, the amount of force you used, the rate the stick broke at, how much mass remains in the two pieces and how much was lost to splintering, etc.

But I can measure those with a ruler and a scale. Both before and after the breakage. Takes a few seconds, and I'd need to do that before punching those numbers into the simulator anyway. And I can be precise with where I apply the force, etc. So I don't see how this gets to the point of the question.

Re: Google claims to have proved its supremacy with new quantum computer

#68

Post by wanderingstan » Mon, Jul 03, 2023, 5:54 PM UTC

jcims wrote: post #66

dataflow wrote: post #7

Dumb question: Say I have a wooden stick and I break it in half in less than a second. Assume a computer would need several minutes to simulate everything that would've happened in the stick. I clearly got the output faster than a computer (and with more precision), so does this imply I'm doing anything particularly fascinating? I assume the same scenario is possible to concoct for a quantum computer. I assume it wou…

I wonder about that frequently. The universe 'executes' physics in, as far as I can tell, a realtime basis (at least within the local reference frame). What is that called as compared to computing a model of the same.

Of course for you it feels like realtime, because the physics of “you” are being executed on the same “system”.

It’s funny to think that if we are living in a simulation, that the machine we’re running on might have horrible uptime, but we’d never know because our “time” only works when the machine is running!

Re: Google claims to have proved its supremacy with new quantum computer

#69

Post by olliej » Mon, Jul 03, 2023, 5:57 PM UTC

esrauch wrote: post #15

knorker wrote: post #9

Earlier quoted context omitted.

My understanding is that quantum computers only have two real use cases, as of today: 1. Breaking crypto. 2. Simulating other quantum systems. For (1) it's basically all downsides. For (2) unless you're a particle phycisist you'll never need quantum computers. But that's now. Maybe there will be a killer app for it some day, changing everything. Or indeed, we could get it indirectly. Maybe simulating quantum systems…

If it is able to break crypto then surely that means it can do other "interesting" mathematical calculations that are currently extremely slow/hard though?

Nope.

QC allows a very specific attack that breaks the main asymmetric algorithms today, ECC and RSA - there is no meaningful attack on anything else (Grover's algorithm is technically a sqrt improvement on symmetric algorithms but that isn't close to breaking AES256).

There's nothing general purpose in the attack on RSA and ECC: you can use the quantum Fourier transform to find the period of the key, and that period tells you the key. But this is very specific - you have two algorithms that have inherently periodic behavior, where the period is meant to be secret, and so all you need from the QC is the period.

It's hard to see how to extend that to other problems.

I'm more curious about (and kind of wish more emphasis would be made on) programmable analog computers, which is what it seems QC should be capable of. A lot of the difficult with QC is that people seem really fixated on discrete problems (like factoring \o/) where there's a single correct answer, and so huge amounts of effort and research are going into error correction, etc. There are lots of problems (simulations as in this circuit), where the classical solution simply requires running millions of times with random variance to converge on a sufficiently accurate result where a QC would theoretically be great. The general use problem in that case becomes "how do I convert this classical problem into a QC compatible algorithm".

Re: Google claims to have proved its supremacy with new quantum computer

#70

Post by JumpCrisscross » Mon, Jul 03, 2023, 6:01 PM UTC

dataflow wrote: post #67

zdragnar wrote: post #27

Earlier quoted context omitted.

The difference between you breaking a stick and the computer modeling it is that you've measured nothing. You don't know, with any precision, the amount of force you used, the rate the stick broke at, how much mass remains in the two pieces and how much was lost to splintering, etc. In other words, assuming the computer model has sufficiently accurate data as an input, it can produce significantly more refined output…

> You don't know, with any precision, the amount of force you used, the rate the stick broke at, how much mass remains in the two pieces and how much was lost to splintering, etc. But I can measure those with a ruler and a scale. Both before and after the breakage. Takes a few seconds, and I'd need to do that before punching those numbers into the simulator anyway. And I can be precise with where I apply the force, e…

> I can literally measure those with a ruler and a scale

You can’t measure a number of internal stress-strain conditions during the moment of failure. You can’t repeat the experiment with the same stick. The best way to get a fast intuition for why the simulation is superior is to take an entry-level CAD course with a focus on material design.