Hacker News Reader

> 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.

> 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.

I'm sorry but you're completely missing the point of my question. The question was not "what can simulations do that experiments can't". My background in simulation is not zero, and I've never had that question. The question as something else entirely.

> You can’t measure a number of internal stress-strain conditions during the moment of failure

But what if that's not what I'm interested in simulating or measuring? What if all I care about is something easily measurable, like the length of each remaining piece? Isn't that precisely my point here? There are so many things quantum computers can't compute either - yet we seem to be judging them by what they're really good at. Just like with the wooden stick. So how do I tell if they're the same sort of scenario or not? What's the distinguishing criterion?

Are you just asking if these are unrealistic, contrived performance benchmarks? Those have always existed and they're fine: https://news.ycombinator.com/item?id=20231084

Some are more useful than others. There's no strict criteria just as there is no perfect way fully characterize CPU performance.

"Unrealistic" or "contrived" don't really get to the heart of the matter, at least not as I understand them. The stick example can be extremely useful and realistic - it doesn't really make any difference to the question.

Perhaps another way to phrase it (though I'm not 100% sure this is equivalent) is: how do they know whether whatever they're accomplishing is quantum computation, as opposed to something else (like analog computation) that happens to be concerned with quantum physics.

> What if all I care about is something easily measurable, like the length of each remaining piece?

If you're not just talking about the two large pieces, but also the lengths of the small splintered pieces, I suspect the computer is going to be a lot faster at figuring that out than you will be manually measuring them.

Your responses in this thread seem to be a bit disingenuous: first you ask why a computer simulation could be better than doing it manually, but then when people tell you the things the computer can do faster/better than you, you say "but what if I don't care about that?" Well, duh, if you don't care about anything but the most trivial things, the computer probably isn't going to be of any benefit to you.

> first you ask why a computer simulation could be better than doing it manually

No, I never asked this at all. Some people changed the question to this for reasons I still don't understand. The actual question I asked was: "what are the criteria for distinguishing those scenarios from the interesting cases? And how do we know which case this one is like?"

The whole point here is to figure out if the problems they're declaring computational superiority on are analogous to the wooden stick problem here. Not to ask "when are simulations better than doing it manually".

You're going to measure every point like this? https://www.researchgate.net/profile/Andrzej-Baier/publicati...

Likewise, you can crash a car in a lab and do a simulation of one. Both tell you "fascinating" things and are still done by engineers.

This has nothing to do with the question though.

> You're going to measure every point like this?

Why would I need to? Is "tells you a bunch of answers to questions you never asked along the way" really the distinguishing factor for what constitutes computational supremacy? If all I wanted was just the lengths of the broken pieces, my simulation has to tell me the stress and strain at every point in order to be considered superior to a numerical simulator? Merely telling me the answer to the question I asked doesn't count? So if my prime factorization tool factored 4096-bit integers instantly, that's not enough? It has to factor a bunch of unrelated numbers and maybe solve TSP in the middle before we can consider it superior to classical computation?

You really just answered this yourself:

> 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?

You assert that you have output faster than a computer with more precision. However, you do not have any empirical data, just observable data; as stated by zdragnar:

> 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.

Then you further state that you can measure those with a ruler and a scale; however, this inherently takes time with significant uncertainty in your measurements and calculations. Whereas a computer will provide all of those numbers.

The other thing to consider is the method of simulation such as finite-element analysis (FEA) and the resolution you need. You can get segmented data all the way to down a specific volume of that stick, good luck with the hand calculations on that.

I don’t think you guys are tackling this right. You can observe the stick breaking to absurd precision, even using an electron tunneling microscope, and produce immense amounts of data about the sticks state at every point in the breaking and afterwards faster than a computer can simulate breaking the stick. The key is the stick breaking and it’s observation at any level of detail happens in parallel with all aspects of the system coherently acting in real time simultaneously in a way with perfect effects and transmission of information between all quantum parts of the system at the speed of light. The computer on the other hand is operating in a binary encoding Turing machine forced to rely on imperfect relatively serial (compared to the parallelism of reality) with encodings that and algorithmic mathematical models that produce encodings that are similar to what reality produces naturally. The issue is the layers of abstraction, nor the measurement and detail of the system. All the measurable information about the stick and it’s breaking is available and measurable with no impact on the time it takes to break the stick.

Here’s perhaps a more clear example. The universe influences itself at a distance with all bodies acting on all other bodies at once. This is a classical nbody problem where n->inf. We can measure quite easily these effects without perturbing the system, which would not be able to progress a single iteration in the entire span of the universe if simulated by a computer.

But the comparison is not fair.

Yes the electron microscope can image a lot of details "in parallel" but not all details from all angles, all internal microfractures. You can't easily measure all temperature gradients in all cubic nanometers of the material etc etc.

The simulation is slow because it works at that level and thus as a side effect will also give you that output.

Obviously if you don't need all that information you may find another route to arrive at the results you want

You could though, it’s entirely fair. All information about the system is available for measurement down to the quantum level across all dimensions, because it’s a perfect fidelity to reality. Regardless of whatever specific tool you choose, in theory I could build tools that measure in near infinite detail the stick. I could not however build a computer that could calculate that detail of the stick with perfect fidelity to reality. It’s a gross approximation no matter what you do. I would however point out that my nbody example is entirely analogous and is perhaps more tractable as to why it’s impossible to compare measuring reality to calculating a simulation. Simulations have the advantage that you can reproduce and tweak it repeatedly without having to configure the universe in a specific way every time. You can hold variables constant, you can change the laws of nature, you can simplify complexities to narrow in on a specific behavior. But simulations are never a replacement for measurement of reality, and the reason is the opposite of what you’re holding out. Reality can be measured to an arbitrary level of complexity literally in real time, even systems that are so complex we can’t consider possibly simulating even approximations of the system.