/sci/ - Science & Math

Yes and no.

At
The
Same
Time

>>15280945
Not yet. I mean they work, you can rent time on them on AWS or whatever. But they suck at the moment. Maybe in 5-10 years they’ll be decent but a lot of the companies in the space can’t survive that long.

>>15280945
They are pretty legit as noise generators.

>>15280945
I never had anyone explain what they do and what the most basic implementation would be.
A quantum computer that adds 1 and 1, how would that be in an abstract sense and in a lab setting?

>>15281852
>I never had anyone explain what they do
They compute ..... fast

>>15280990
Knower

>>15280945
Quantum computers aren't even real computers. They aren't turing complete and can simply perform certain terminating functions significantly faster than the way we normally compute. Using a qc is like using a very special gpu that only works in certain cases.

>>15281852
>>15282047
The quantum state means each bit has 3 positions not 2. This means additional quantum bits give exponentially more states then a normal bit. More states equal more processing power. This is basic shit guys.

>>15280945
If you have a steady supply of liquid helium

>>15282113
why did you tag me faggot?

>>15280945

Quantum computers of the future will be essential to machine evolution as their non-deterministic nature permits copying errors without a reliable prevention mechanism, and thus, mutation

>>15280945
Yes, the hardware works, but I am not convinced that any Quantum Instruction Set Architecture is fully scalar, as designing a fully scaled ISA for Quantum processing is a task that is beyond the capability of humans without several generations of work, and/or the aid of AI.

Source: Spent months trying to write a SIMPLE Quantum ISA based on equivalencies to x86.
It's ridiculous to attempt to create an ISA with standard compute ability, with quantum processing at this time, as well as still have instructions that take advantage of the maximum possible limits of what it could potentially do.

The I/O is excessive, you can't just truth table it out.

>>15283255
>The I/O is excessive, you can't just truth table it out.
Isn't this where language models with math capabilities will be actually useful? Predicting and executing upon system calls? I know nothing but have been thinking of AI sufficient enough to emulate hardware under any OS within its own chips. That might be ridiculous though.

>>15282218
>non-deterministic
Lol, this spook again

>>15283275
Having an AI handle the I/O differentiation for quantum sets on the fly, would probably be the ideal situation, since pre-coding the gates seems like an impossibility.

You're not far off from what is probably necessary.

We just can't do it yet, as far as I know.

GPT4 can't score a 5 on AP calc tests, I doubt it's going to get the logic complexity of quantum output and gate control.

>>15282106
>They aren't turing complete
Proof?

>>15282113
This is literal retardation. Disregard it entirely.

>>15281852
When a normal computer has N bits, it can hold N pieces of information, with each piece of information being either a 0 or a 1. Due to quantum entanglement, when a quantum computer has N bits, it can hold 2^N pieces of information, with each piece of information being an arbitrarily precise number (in practice it could never get arbitrarily precise, but in theory it can).

>>15285163
Shut the fuck up pseud

>>15280945
Dead fad, AI has completely taken over.

>>15285600
>quantum AI
trillion dollar buzzword

>>15285149
They are. The circuit model is just a lot easier to work with than the quantum Turing tape. These results were proved way back in a Bernstein and Vazirani paper in 1993.

https://dl.acm.org/doi/10.1145/167088.167097

>>15282113
>observation crashes the bit into a binary state

so it can calculate things but it can't observe the results?

>>15286011
the trick is to develop an algorithm which produces constructive interference toward the solution and destructive interference otherwise. In practice this has been very hard to do.

>>15285629
I will do it, anon. I will build a trillion dollar quantum AI company.

>>15280945
The real quantum tech of the future is photonic computing

>>15286402
like a photonic quantum computer or an analog photonic classical computer?

>>15282113
lmfao why would we need quantum processing for trinary bits

>>15285406
Eat a dick, fucknigga

>>15286402
requires extraordinarily short wavelength light to potentially outperform presently available technology, xray or stronger, which sets the energy per bit very very high and makes working with the technology physically hazardous

>>15286653
latter probably
>>15287051
can't they be made physically larger to compensate? so tradeoff between size and energy consumption

>>15280945
No, because of decoherence and because they are useless without algorithms designed to exploit the actual features that set them apart from classical computers.

>>15282218
Quantum mechanics is deterministic. GTFO

>>15287051
that’s not true, there is a lot of work being done on telecoms wavelength (NIR) quantum photonics

Currently adding to LeBron James legacy in my toilet bowl

>>15280945
The main hold up is the large overhead for quantum error correction to compute anything useful and getting the errors down low enough for quantum error correction to work. There is a lot of false hype because scientists and entrepreneurs need funding and 15 years-olds have all the money.

>>15287164
The no-cloning theorem is a real thing.

https://en.wikipedia.org/wiki/No-cloning_theorem

>>15280945
Yes. Progress is steadily being made.
>>15282106
Wrong

>>15285990
Show me an example of a quantum computer used in the last 5 years that is turing complete

>>15288877
Show me an example of a non-quantum computer that is turing complete. You can't.

>>15285406
>spastic adversarial outburst for no reason towards random person
hi schizo anon. I know it is you

>>15280945
I don't know but there sure are a lot of people hiring for quantum computing roles

>>15289778
One of those posts in your epic reddit screenshot is mine.

>>15289779
bubble imo, hopefully i am wrong

>>15289791
you'd be too ashamed of yourself to admit that if this were not an anonymous forum. the power of the anonymous forum draws out honesty that would be missing on other platforms. nice to see shii's theory of internet forums working in action

Pretty legit at getting fundings

>>15289971
That's what the threat of the chinks breaking out cryptography does

>>15280945
Already cracking your encryption for NSA.

Wouldn't quantum computing make actual analog inputs achievable? That is huge, if I understand it right.

>>15287164
based

>>15281852
The way I understand quantum computers are able to perform multiple computations at the same time, a bit like if every iteration of a for loop was executed at the same time.

>>15291338
youre describing parallel processing which isn’t exactly how a quantum computer works. They are fundamentally quantum in nature which makes it difficult to conceptualize. An explanation I like is that they work by turning your 1s and 0s in your computer into a sea of waves. Each wave being a superposition of 1 and 0. Next you need an algorithm which creates constructive interference of all the waves towards the correct solution and destructive interference otherwise. famous examples of algorithms like this are Shor’s and Grover’s. There is a quantum fft algorithm as well. You could say that a quantum computer is doing parallel processing of all the possible pathways in the problem but thats a classical way of looking at a quantum solution, because you’re not actually processing any of these.

>>15291492
Are there any limitations associated on of said wave or the algorithm that interprets it? If no then I would be right thinking this>>15291105

>>15280945
They are legit in the sense that the theory is selfconsistent and works.

The problem is noise. Current qubits and quantum operations just accrue errors at rates that make current hardware utterly useless. Whether or not qubits will ever reach noise levels which are acceptable is unknown. Not even the professors are entirely certain about this, they are mainly just hopefull.

I guess the main take away is, there is no theoretical reason why they wouldn't work, just practical reasons they may never work.

>>15282113
Superposition doesn't mean you have 3 states you fucking retard. Please educate yourself before you start posting shit like "this is basic shit guys". It isn't, and most people will likely never understand this shit because they don't understand how quantum mechanics works. I take it you're probably one of them.

>>15291526
Quantum computers are indeed analog, but there are also plenty of non-quantum analog computers which take continuous (real-valued) inputs and produce continuous (real-valued) outputs without resorting to digital signaling or processing at all. One famous mechanical example is the Antikythera mechanism, another is the Sliderule. There have even been analog electrical computers built without threshholding or logic gates.

The advantages of a quantum computer go beyond analog computing because in addition to processing at continuous/analog precision and throughput, they exploit the uncertainty and entanglement principles to effectibly run multiple analog computtations at once -- "in different universes" -- and only have the "correct" answer render in our qubit, our quantum computer, our universe.

The first paper describing a Universal Quantum Computer (in 1985!) actually called it a "machine to test the existence of parallel universes," which couldn't yet be built with the technology of the era.

>>15291617
>the "correct" answer
perhaps a better phrasing would be "the desired result"

>>15282113
>The quantum state means each bit has 3 positions not 2

>>15291105
Sound effects allows that, literally audio processors...

>>15285163
One reasonable answer in this entire thread and it only has one reply: “shut the fuck up, pseud”

Jesus, this board is a fucking barren wasteland of retards.

>>15291331
A Brazilian-American-British Jew mutt Marxist.
That's actually why he put so much work into his fully deterministic quantum interpretation - because Marxism demands that the future is fully determined by the past.

It promises exponential speedup in computing algorithms by conveniently leaving out the exponential amount of energy and qubits you need to error correct it.

>I'M GOOOONNNAA QUANTOOOOOOMMMMM

>>15294934
This. It's an inherent feature of measuring quantum systems. It is impossible to extract the information faster than classical.
All a huge scam basically

>>15294934
Energy is cheap, and the resources needed for error correction does not necessarily scale exponentially. It could be nearly linear in the qubit count

>>15297295
You don't understand the point.
Actually getting the answer of your computation always negates the "quantum speedup". Jewgle and co have been coping about this for years but it's evidently true.

>>15281852
Gate-based QCs are linear algebra machines. That's what they do quickly, perform reversible operations on vectors.
Adiabatic QCs leverage quantum effects like tunneling to find optima in complex energy landscapes where you have to shape that landscape to reflect your problem and the ground state to represent your problem's solution.

>>15280945
They are about as legit as fusion power plants.

>>15288421
That doesn't mean it's non-deterministic, idiot.
>>15287164
Measurements are inherently non-deterministic you spastic.

>>15297687
Measurements are an illusion. The multiverse continues on deterministically. You only perceive your branch.