/sci/ - Science & Math

>>10219956
>quantum computer right now has no practical application.

After seeing all the changes in technology since the 50s you can't be this retarded.

>>10219956
Factoring large products of primes has to be useful for more than just cracking RSA encryption. Also, aren't quantum computers supposed to be able to do calculations in quantum superposition so they can try several approaches simultaneously unlike regular computers which actually can't do more than one thing at once? Shouldn't that give them a huge advantage?

>>10219956
>The problem is no one uses RSA anymore. They use AES which has a different scheme that there is no quantum computer algorithm for, so a quantum computer right now has no practical application.
A quick google search indicates that AES, as a symmetric key encryption, still requires RSA or a similar public/private key encryption (??) to securely transmit the AES key. So breaking RSA could get you AES keys, which get you whatever was encrypted in the first place.

>>10219956
Even if this were true that there would be nothing to gain, why do you restrict yourself to current known applications before researching something? This is /sci/ not engineering

>>10219976
>still requires RSA or a similar public/private key encryption (??) to securely transmit the AES key

It doesn't matter. How will a quantum computer crack the AES? RSA is just used as a hashing algorithm. No one encrypts the content in RSA.

>>10219978
>why do you restrict yourself to current known applications before researching something?

1. I studied quantum computers and worked with a leading professor in the field at the University.

2. In the industry I've worked with encryption.

Now tell me im ignorant.

>>10219983
do you even know the difference between symmetric and private-public key encryption?
If you want to dtore data in encrypted format go ahead and use AES.
If you want to transmit informstion securely without having to share a key first (which someone could see) you need asymmetric encryption such as RSA

>>10219997
>If you want to transmit informstion securely without having to share a key first (which someone could see) you need asymmetric encryption such as RSA


you need asymmetric encryption such as RSA

No you don't. It's just used now as a hashing algorithm for key having in AES. Even if you cracked the RSA you still haven't cracked AES and you cannot decipher the content.

Also once you have a machine that can factor the products of large primes people will just use another hashing algorithm instead of RSA.

>>10220007
Don't they all use primes? None of the algorithms are going to be safe.

>>10219990
>In the industry I've worked with encryption.
Obviously it wasn't helpful.

>>10220021
>Don't they all use primes? None of the algorithms are going to be safe.

No AES has nothing to do with factoring large numbers. Only RSA did and no one uses RSA except as a hashing algorithm, which can be accomplished many other ways.

>>10219990
Assuming you're are true, neither of those imply you're not ignorant. Maybe you're just a shitty researcher, maybe you have bad colleagues who brainwashed you, maybe you're just in it for the paycheck, idk

>>10219990
>In the industry I've worked with encryption.
And I've worked with truecrypt for my documents, what's your point?

>>10220035

The fact is modern encryption does not solely rely on factoring large numbers. Even if you build a quantum computer today you'd have no use for it.

>>10220042
Remember when truecrypt used to be free?

>>10220042
>And I've worked with truecrypt for my documents, what's your point?

Are you gonna tell me your encryption relied solely on factoring large numbers? Maybe you don't understand me. I work with encryption down to the microprocessor instruction level. I know exactly what is happening in the registers with keys and cypher text.

>>10219956
>Why are quantum computers a thing?
They're not.
https://www.youtube.com/watch?v=mIULkvTGk2U

>>10219956
Op, how did you gain access to this pc? You shouldn't be here talking about things you don't even understand

>>10220050
What about>>10219966
I don't all this stuff in intro to proofs class (including how AES works) but I got a D in that course lmao

>>10220059
>You shouldn't be here talking about things you don't even understand

I've taken the user manual for a microprocessor and built an encryption system the ground up in assembly ans implemented it in a government facility. I also studied quantum computers at the University.

I know what I'm talking about.

Do you really think all encryption relies solely on factoring large numbers and there is no other way to encrypt things?

>>10219966
>Factoring large products of primes has to be useful for more than just cracking RSA encryption

name one application.

>>10220070
Quantum computers aren't built for encryption

>>10219956
It can be used to simulate quantum mechanics or so I've heard.

>>10220076
>Quantum computers aren't built for encryption

Tell me an application where they perform better than classical computers.

>>10220050
No, my point is based on his wording, it could be anything from actually programming encryption (which I'm not going to assume) to using software that has encryption in the name. Since he's relying on an authoritative approach, I'm going to assume the worst

>>10220074
Proving the Riemann hypothesis

>>10220077
>It can be used to simulate quantum mechanics or so I've heard

This is absolutely true. But a classical computer in the same way can excel at simulating solid state physics

>>10220081
They are still being developed. But they will be used for simulations and AI(?) for example. They aren't built for consumer application like you seem to think.

>>10220084
>Proving the Riemann hypothesis

And how is it going to do that?

>>10220086
No one intends to replace classical computers though

>>10220088
>They are still being developed. But they will be used for simulations and AI(?) for example

Can you tell me how a quantum computer will perform better at AI than a classical one?

>>10220044
There are actually quite a few practical applications for it. For example a quantum computer would be the perfect starting point to simulate large quantum systems (this one is fairly obvious).
Also quantum annealing (D-Wave is a pure quantum annealer) would have a straight forward application in material science, as well as general minimization problems (current AI learning algorithms are basically only minimization problems).

>>10220101
You know simulations don't just run an experiment on the materials in the computer, they run algorithms and equations to predict outcomes

>>10220081
Not that guy, but objectively, quantum computers can be used for annealing simulations. Soap bubble problems, travelling salesman, etc. The reason they believe they could work for cracking encryption is because there are answers which are of greater computational distance (literally, "more wrong") from the correct answer, so as you contain a probability of all possible answers, once you reduce the energy, you're quantumly reducing the quantity of answers which may be true. For a problem which could have it's answer set reduced by 100, a probably which took 100 years to solve may only take 1. Obviously when we're talking like 1024-bit encryption where it'd take longer than the life of the universe computed on a computer made of everything, 100x factor difference isn't significant.

>>10220101
>For example a quantum computer would be the perfect starting point to simulate large quantum systems

So? I said practical applications.

For instance a wave pool is terrific at simulating waves in a pool, but so what?

A car is great at stimulating an internal combustion engine.

Your body is great at simulating chemical reactions.

A transistor is great at simulating a transistor

>>10220107
A problem, not a probably. I've made that typo twice recently. I need to fix my annealing

>>10220097
Well, a quantum computer should be faster at performing certaing actions than a normal one. Like mathematic related things ( i suppose). So ai should perform better. ( If you know how AI works)

>>10220109
>he can't imagine how simulating reality could be useful
nigger. Offense intended. How about simulating protein structure from amino acids to understand diseases and genetic engineering. How about simulating the stability and reliability of new methods to construct computers.

>>10220107
>quantum computers can be used for annealing simulations. Soap bubble problems, travelling salesman,

Can't you simulate annealing better by sticking something in a furnace? Can't you simulate soap bubble in the bathtub? Aren't my alternatives a lot easier too?

>>10220109
A-are you stupid ? No, seriously how the fuck things like that can't be useful? How old are you?

>>10220118
>How about simulating protein structure

Aren't protiens observable? Why not just look at them? Why spend decades and billions of dollars to build a computer that simulates them.

>>10220106
What do you want to tell me with that? Sure some parts of the simulations will still perform better on a classical computer.
But some simulations can just not be done on a classical computer in a feasible time, but are easily done on a quantum annealer.
Simulating Hamiltonians of large systems for example.

>>10220125

I think you are stupid. My point is building a quantum computer to just study quantim systems of the same complexity does not represent a break through in computing. It's a novelty that might have it's niche but will otherwise be useless.

>>10220128
You can't always observe them in vivo. Electron microscopes for example only work on dead things. But if you know the amino acid sequence and the charges etc you can simulate how the protein will forld and what it'll do.

>>10220135
>But if you know the amino acid sequence and the charges etc you can simulate how the protein will forld and what it'll do.

I grant quantum computer can simulate quantum system of simar complexity. This doesn't represent a break through in computing in general. If quantum computers existed today they'd be a novelty in a lab somewhere and wouldn't impact you life in a substantial way.

>>10220133
Yes, that's because quantum computer aren't built for normal people that aren't in the scientific field so you're opinion here doesn't really matter. If you can't understand why simulating things or why the calculation power of a quantum computer can't be useful than I think you don't know how things work in real life.

>>10220155
>If you can't understand why simulating things or why the calculation power of a quantum computer can't be useful than I think you don't know how things work in real life.

I do understand. And I do accept your claim that there is a niche where a quantum can be useful. In general tho it's not worth the time and effort.

>Why are computers a thing? They're not faster than mechanical calculators

>>10220174
Dude, just do it in your head, it's really not much faster using machines and you don't need to carry extra stuff around with you

>>10220174
>They're not faster than mechanical calculators

Solid state computers are blindly fast and are extremely versatile in their application. They are also very simple and cheap to make and maintain. The same can't be said of a quantum computer.

>>10220182
Yes, keep feeding him, keep him alive

>>10219956
Because silicone is reaching its limits right at a time that humans need more and more computing power.
It's a step in the right direction, stop being a faggot.

>>10220193
>Because silicone is reaching its limits right at a time that humans need more and more computing power.

Let's say it is. Who is even claiming quantim computers can be more powerful than classical computers over a diverse set of applications? Like previous posters have said. A quantum computer can solve a very specific set of problems very fast, but in general will not even meant to replace a classical computer

>>10220007
>RSA is a hashing algorithm

>>10220210
Why use a car if it can't replace your feet?

>>10220182
>Perform every calculation simultaneously
>The "competition" is still working it's way through the first computation
>"Just being fast doesn't make it usefull1!"
>Tfw quantum computation is to classic computation as it, in turn, is to an abacus
>Tfw people suggest we just use the abacus

>>10220193
silicone is what goes into breast implants
silicon is the element

>>10220244
learn to read

>>10220250
>>Perform every calculation simultaneously

What type of calculation? *,+_/? Or some quantum simulation? Won't you need to iterate or perform successive steps on the quantum simulations to get back to something useful?

>>10219956
They're a thing only on pop sci like you can witness in this thread, they'll have some cool algorithms to run but it's unsure as of now how that will become a thing or even if it will become a thing

>>10220250
I doubt you've ever taken a class on basic QM, quantum computers don't do multiple calculations simultaneously, they just have a different formal structure and therefore the algorithms that are fast in it are somewhat different, it's not a parallel vs sequential comparison

>>10220260
For a quantum computer, no. The probably is more states are required for more complicated algorithms, to represent all possibilities. A 4-input binary logic requires 16 possible discrete states. I think the point of the advantage of quantum computing is that technically, as long as the system is coupled properly, all possibility exponential states exist, because each uncertain element introduces multiple possible states simultaneously. An 8-qubit register therefore contains 256 possible values, all of which are probabilities. The problem is since each step interferes with the last one, and each step is a probability of certainly transferring, even if you could control each bit with 99% correlation, self interference and measuring failure would make your answers annoyingly bad very quickly.

>>10220260
>What type of calculation?
Geometry, simulations, optimization, ect.

>Won't you need to iterate or perform successive steps on the quantum simulations to get back to something useful?
Depends on how accurate you need your answer, but additional iterations take no time at all so you might as well run it through a few dozen times to get a super accurate figure

>>10220271
I know they don't literally perform every computation, but they effect from the entanglement of the qubits is close enough to this description for a Korean J-Pop forum

>>10220256
learn to write:
>>10220007
>you need asymmetric encryption such as RSA
>No you don't. It's just used now as a hashing algorithm for key having in AES.
WHAT is used as a hashing algorithm? Do you even know what a hashing algorithm is?
(wait while you look it up on wikipedia and desperately try to not look like a total mong)

>>10220165
It is worth because of the repercussion that the results of scientifically discoveries can have using quantum computers. When the first computer was built I don't think that normal people knew how it could be useful...now look what happened with computers. That's what probably can happen with quantum computers but on a different level. Quantum computers are useful in the science fields because of their calculation power that can be used for research. If you think that scientific reasearches are useless than there is nothing that can be done.>>10220174

>>10220356
Not to mention, quantum computers are a realization of a lot of theory we’ve built up in quantum information theory. Building such a machine is the natural next step in further study. There’s a lot of money being thrown because it’s an exciting field with promises of knowledge in both possible successes and failures alike

>>10219956
A real working quantum computer with sufficient qbits would be able to break all asymmetric encryption except elliptic curve.

>>10220379
Typical elliptic curve cryptography is broken by a modified Shor's algorithm.

>>10219956
With quantum computer you can hack any classic computer, because quantum computer is stronger.

>>10220423
Not exactly. Quantum computers are very good at specific classes of problem though. There is research to see if those collapse into the existing classical complexity classes, or if they need unique classes themselves

>>10219956
I was recently surprised to discover that "quantum machine learning" is an actual thing that people are researching, and not entirely made up popsci bullshit. Apparently there is some kind of quantum speedup that is applicable to neural networks. Can anyone redpill me on this?

>>10219990
does your father work at Nintendo?

>>10220133
So you worked with quantum computers in uni and still dont understand that they were never meant to replace regular computers but to act as complement in certain areas. Either you are lying or youre a complete brainlet.

>>10220473
>So you worked with quantum computers in uni and still dont understand that they were never meant to replace regular computers but to act as complement in certain areas. Either you are lying or youre a complete brainlet.

I've already said I agree that a classical computer can't be replaced by a wuantum one. A quantum computer will only have a niche use case. This was pretty much my statement in the op.

>>10220410
>Typical elliptic curve cryptography is broken by a modified Shor's algorithm

Do you have a citation? It's not difficult to change the encryption scheme to something a quantum computer cannot Do trivially.

>who cares about asymmetric crypto
>what is diffie-hellman key exchange

>>10220517
>>10220517
https://arxiv.org/abs/quant-ph/0301141

The problem with of post-quantum cryptography systems, as I understand it, is they have rather large key sizes. Things have probably improved a lot in the last few years, though. I haven't really kept up with the state of the art.

Current quantum computers (that we know about) are shit. When people say they won't be faster than classical, they're talking theoretically (there's not a magic theory trick that make them theoretically dominate for general programming), but engineering-wise in the future they can offer better information density, less entropy loss, and will therefore be quicker, more energy efficient, but we probably need other engineering and materials improvements to get there. I think they just poured money and hype into them because they wanted an easy way to break crypto and and not share.

>>10220872
They poured money into it because it sounds awesome.

how to start learning Quantum Computing?

>>10220435
All linear problems get a quadratic speed up, so it's possible with AI where the problems are not optimized. It turns out a lot of problems can be optimized to have this speed up anyway on classical computers.