How Will Quantum Computing Impact Spaceflight?

[sanman]

International Business Machines, having long ago lost its premium brand in computing in spite of its long tradition in fundamental research, is now seeking to reclaim its pole position - by proclaiming a new era of Quantum Utility:



As new advances in error-correction and noise-reduction continue to be made, can Quantum Computing now become practical?
Should Nvidia be looking hard in its rearview mirror?

Sometimes we'll hear about major corporations, some of them even aerospace giants, coming forth to drink from the quantum trough. Do they truly see something useful in this nascent technology, or is it just their shot in the dark?

Will this technology find its way to data-centers only to remain locked up there?
Or could it one day go mobile and even take flight on spacecraft?
Could cold-traps in permanently shadowed lunar craters one day become sites for future quantum data centers, giving us yet more reasons to go to space?

[JohnLloydJones]

For all the constant buzz from certain quarters, quantum computing is still more promise than substance. There is a certain class of problems where quantum computing may become valuable, but I haven't seen much to indicate that quantum computing will take over general computing. Of course, some one may announce a breakthrough that changes all that, but I'm not holding my breath.
 

[Twark_Main]

Sabine still has the best video on quantum hype vs. quantum reality.

Could cold-traps in permanently shadowed lunar craters one day become sites for future quantum data centers, giving us yet more reasons to go to space?

Datacenters produce lots of heat. You would quickly "use up the cold" and saturate your ability to radiate away heat.


Also, running a cyocooler on Earth simply isn't expensive enough to justify it.

Also, the Moon's cold traps are important scientific sites, containing a record of billions of years. Hopefully we're not short-sighted enough to bulldoze them for server farms or gas stations....   

[sanman]

Datacenters produce lots of heat. You would quickly "use up the cold" and saturate your ability to radiate away heat.

Traditional data centers produce lots of heat -- but do Qubits have to produce lots of heat? Qubits don't like heat - that's why they have to be cooled in Dilution Refrigerators.

Also, running a cyocooler on Earth simply isn't expensive enough to justify it.

But there's also all kinds of other noise contamination on Earth, from all the electromagnetic waves we produce.
Just as we might like to build a sensitive radio telescope on the Far Side of the Moon shielded from the electromagnetic emissions of Earth, perhaps our quantum computing devices would also similarly function better with less disturbances in remote regions of outer space, at colder temperatures, surrounded by shielding. Project Kuiper in the actual Kuiper Belt?

Also, the Moon's cold traps are important scientific sites, containing a record of billions of years. Hopefully we're not short-sighted enough to bulldoze them for server farms or gas stations....   

The nature preserve around Starbase Boca Chica may also have importance in ecology and wildlife study -- but sometimes we make sacrifices to further our progress in another area.

[sanman]

For all the constant buzz from certain quarters, quantum computing is still more promise than substance. There is a certain class of problems where quantum computing may become valuable, but I haven't seen much to indicate that quantum computing will take over general computing. Of course, some one may announce a breakthrough that changes all that, but I'm not holding my breath.

Computing is something we all depend upon, and pursuit of it in the quantum realm may then open up that realm and ultimately improve our technological mastery over quantum physics.

The IBM presenters were talking about achieving "Quantum Error Correction" in a few years -- what would John Stewart Bell (of Bell's Inequality fame) say about that? If you can correct errors between entangled Qubits in close proximity to each other, then could you potentially correct them between entangled Qubits that are farther apart? Like, a lot farther apart?

https://www.nature.com/articles/s41586-023-05885-0

Loophole-Free Bell Inequality Violation with Superconducting Circuits
...
Our work demonstrates that non-locality is a viable new resource in quantum information technology realized with superconducting circuits with potential applications in quantum communication, quantum computing and fundamental physics

Ansible when?!

[Twark_Main]

Datacenters produce lots of heat. You would quickly "use up the cold" and saturate your ability to radiate away heat.

Traditional data centers produce lots of heat -- but do Qubits have to produce lots of heat?

Yes, of course. Why would I have brought that up if they didn't? 

Also, as with any compute system, the ancilliary systems (not just the q-bits themselves) consume a large fraction of the power/thermal budget.

Also, running a cyocooler on Earth simply isn't expensive enough to justify it.

But there's also all kinds of other noise contamination on Earth, from all the electromagnetic waves we produce.
Just as we might like to build a sensitive radio telescope on the Far Side of the Moon shielded from the electromagnetic emissions of Earth, perhaps our quantum computing devices would also similarly function better with less disturbances in remote regions of outer space, at colder temperatures, surrounded by shielding. Project Kuiper in the actual Kuiper Belt?

It's still cheaper to go deep underground, if that does happen to be your main constraint. Yes even with the heat.

Also, the Moon's cold traps are important scientific sites, containing a record of billions of years. Hopefully we're not short-sighted enough to bulldoze them for server farms or gas stations....

The nature preserve around Starbase Boca Chica may also have importance in ecology and wildlife study -- but sometimes we make sacrifices to further our progress in another area.

Those two examples are nowhere near comparable. They smell kinda similar though (ie there are vaguely similar "tribes" within the overall "battle" cliche), which is the only narrative modern social media is capable of disseminating -- sports team metaphors. So goes the logical brain hemisphere of the masses, and so goes this thread... 

"Well sure we destroyed Boca Chica, but it was all worth it because destroyed the cis-Earth system's only billions-year ice record too, obliterating irreplaceable scientific data we went there to get!" 

The actual logic makes no sense. From the start the Boca/Moon reason-by-analogy is strained beyond the breaking point.


Anyway, the point stands that making your cooling budget smaller (which this won't do btw) is a lousy reason to build your Amazon Datacenter on top of a World Heritage Site. This idea is as bad as using the Great Pyramid of Cheops to prop up your solar farm so you can save some metal...

[sanman]

Anyway, the point stands that making your cooling budget smaller (which this won't do btw) is a lousy reason to build your Amazon Datacenter on top of a World Heritage Site. This idea is as bad as using the Great Pyramid of Cheops to prop up your solar farm so you can save some metal...

Well, I know certain special places like the Apollo landing sites are considered special heritage places - but does every spot on the Moon have to be a special heritage place? Otherwise, then pretty soon all of Mars could be a special heritage place, and all the asteroids too, etc -- and then there's no room left to exploit any space within reach.

[Twark_Main]

but does every spot on the Moon have to be a special heritage place?

We've both been talking about the polar "eternal night" cold traps, not just any spot on the Moon.

[JohnFornaro]

Ansible when?!

Points.

[edzieba]

Before things are further derailed by imagined lunar nature thermal-reserves or pointless hand-wringing over how cryocoolers work...


Quantum Computing is useful for more rapidly computing algorithms that tank an unfeasible amount of time for classical computers. Currently, there aren't many of these that are in actual use - because their uncomputability means nobody is going to implement them (with the exception of scamgraphy, where that is a feature not a bug). When it comes to hypothesised applications to NP-hard problems, some of the low hanging fruit are probably in accelerating simulations like CFD.

[sanman]

Ansible when?!

Points.

It seems like an Ansible -- ie. a means of instantaneous communication that's faster-than-light:

https://phys.org/news/2023-05-entangled-quantum-circuits-einstein-concept.html

"There are 1.3 tons of copper and 14,000 screws in our machine, as well as a great deal of physics knowledge and engineering know-how," Wallraff says. He believes that it would in principle be possible to build facilities that overcome even greater distances in the same way. This technology could, for instance, be used to connect superconducting quantum computers over great distances.

[edzieba]

Ansible when?!

Points.

It seems like an Ansible -- ie. a means of instantaneous communication that's faster-than-light:

https://phys.org/news/2023-05-entangled-quantum-circuits-einstein-concept.html

"There are 1.3 tons of copper and 14,000 screws in our machine, as well as a great deal of physics knowledge and engineering know-how," Wallraff says. He believes that it would in principle be possible to build facilities that overcome even greater distances in the same way. This technology could, for instance, be used to connect superconducting quantum computers over great distances.

No.
Like all distant entanglement experiments, data cannot be transmitted. You can observe the state at one end and have knowledge of the state at the other end, but you cannot influence either state without breaking entanglement. i.e. you can observe your bit and see that it is 1 and know that the other bit must be 0, but you cannot use this to communicate because you did not get to choose whether the bit was 1 or 0, it is random which one you observe to start with.

[JohnFornaro]

Ansible

Interesting article.  Thanks for posting.

But.

They don't inform the reader in the least about what the experiment did with those 13,000 screws.  And it turns out that Bell's theorem is false?

https://www.scirp.org/journal/paperinformation?paperid=94678

It is concluded that, Bell’s theorem is false because Bell’s inequalities are trivial mathematical relations that, due to an unsuitable assumption of probability, lack essential connection with the real measuring process of the pertinent experiments.

Nobody's explained to me why two entangled particles, separated by an arbitrary distance do not "communicate".  Whatever you "do" to the one particle is "noticed" by the other particle.  If the Morse code for SOS is sent from the first particle, cannot the "message" be read by the distant particle?

What is it that can't be done? change the first particle?  Or read the second particle?

Here's a link for the lazy:

https://en.wikipedia.org/wiki/Bell_test

This article doesn't include a reference to the 2019 paper above.

And the CSHS Inequality?  More math than should be allowed by law:

https://en.wikipedia.org/wiki/CHSH_inequality

Unfortunately, the experimental setup of Alice, Bob, and Charlie doesn't readily map to the experimental machinery I'm broadly familiar with.

I keep hoping/thinking that an ansible is possible in priciple.

[sanman]

Nobody's explained to me why two entangled particles, separated by an arbitrary distance do not "communicate".  Whatever you "do" to the one particle is "noticed" by the other particle.  If the Morse code for SOS is sent from the first particle, cannot the "message" be read by the distant particle?

What is it that can't be done? change the first particle?  Or read the second particle?

Here's a link for the lazy:

https://en.wikipedia.org/wiki/Bell_test

This article doesn't include a reference to the 2019 paper above.

And the CSHS Inequality?  More math than should be allowed by law:

https://en.wikipedia.org/wiki/CHSH_inequality

Unfortunately, the experimental setup of Alice, Bob, and Charlie doesn't readily map to the experimental machinery I'm broadly familiar with.

I keep hoping/thinking that an ansible is possible in priciple.

The thing is that Quantum Entanglement is theorized to result in instantaneous state change across the 2 particles, independent of separation distance -- meaning it's "non-local" and could conceivably be FTL. So of course someone will say that's not allowed because of a "No Information Theorem", and the overwhelming effect of random noise. But as IBM and other research papers now being published are saying, advances in error correction against noise are showing that meaningful extraction of information from this "non-local" system is possible.

So if that paper in Nature is right, then perhaps we can have access to this "non-local" phenomenon, which we could one day make use of.

No.
Like all distant entanglement experiments, data cannot be transmitted. You can observe the state at one end and have knowledge of the state at the other end, but you cannot influence either state without breaking entanglement. i.e. you can observe your bit and see that it is 1 and know that the other bit must be 0, but you cannot use this to communicate because you did not get to choose whether the bit was 1 or 0, it is random which one you observe to start with.

But Ed, in quantum computing you do get to choose whether the bit is 1 or 0, because you're loading your information into the entangled qubits. Quantum computing wouldn't be very useful for computing if we couldn't take information about some problem and load it into the entangled qubits. Nobody wants to watch some "calculator" you can't load your calculation into. The whole purpose of Quantum computing is to compute, and not just sit there serenely. We load information into the entangled qubits, they do a computation, and then we read the results out.

What do you have to say about this paper published in Nature?

https://www.nature.com/articles/s41586-023-05885-0

Loophole-Free Bell Inequality Violation with Superconducting Circuits
...
Our work demonstrates that non-locality is a viable new resource in quantum information technology realized with superconducting circuits with potential applications in quantum communication, quantum computing and fundamental physics

Anyway, people are talking about the idea more and more:

https://quantumtantra.blogspot.com/2019/11/quantum-ansible.html

[Twark_Main]

But Ed, in quantum computing you do get to choose whether the bit is 1 or 0, because you're loading your information into the entangled qubits. Quantum computing wouldn't be very useful for computing if we couldn't take information about some problem and load it into the entangled qubits. Nobody wants to watch some "calculator" you can't load your calculation into. The whole purpose of Quantum computing is to compute, and not just sit there serenely. We load information into the entangled qubits, they do a computation, and then we read the results out.

You can't just "set" qubits and have them be entangled. You have to go through a special procedure to entangle two qubits on a quantum computer, and afterwards you don't know if they're zero or one.

Entanglement (by definition) requires a qubit to be in a superposition. If you know for sure it's a 1 or a 0, then (by definition) it's not in a quantum superposition. 


Furthermore, even if you can solve this problem, you still can't build an ansible. All you'd be doing is transmitting regular bits to Bob, which happens at the speed of light. The only things that "goes FTL" is decoherence, but there's no decoherence if you already know the bit values.

What do you have to say about this paper published in Nature?

https://www.nature.com/articles/s41586-023-05885-0

Loophole-Free Bell Inequality Violation with Superconducting Circuits
...
Our work demonstrates that non-locality is a viable new resource in quantum information technology realized with superconducting circuits with potential applications in quantum communication, quantum computing and fundamental physics

This paper ruled out local hidden variable theories. It does not (and doesn't claim to) build an ansible, or show that one is possible.

"As of 2015 [i.e. this paper], all Bell tests have found that the hypothesis of local hidden variables is inconsistent with the way that physical systems behave."

https://en.wikipedia.org/wiki/Bell_test#Hensen_et_al.,_Giustina_et_al.,_Shalm_et_al._(2015):_"loophole-free"_Bell_tests

Anyway, people are talking about the idea more and more:

https://quantumtantra.blogspot.com/2019/11/quantum-ansible.html

This is connecting two unrelated things. The blog post is debunking a different (hoax) paper.



Anyway FTL communication belongs in New Physics, so it's OT for this thread. Let's not get another thread locked....  https://forum.nasaspaceflight.com/index.php?topic=34854.20

[sanman]

You can't just "set" qubits and have them be entangled. You have to go through a special procedure to entangle two qubits on a quantum computer, and afterwards you don't know if they're zero or one.

Entanglement (by definition) requires a qubit to be in a superposition. If you know for sure it's a 1 or a 0, then (by definition) it's not in a quantum superposition. 

Okay, fair enough - but even if you don't know whether they're 0 or 1, you do know that they correspond to your input. Otherwise, how can you have any calculation/computation without input? Likewise, how can you have any calculation/computation without reading an output that is based on input? As long as I'm reading an output derived from input, I don't care whether it was represented as 0 & 1, or as Red-Green-Blue.

Furthermore, even if you can solve this problem, you still can't build an ansible. All you'd be doing is transmitting regular bits to Bob, which happens at the speed of light. The only things that "goes FTL" is decoherence, but there's no decoherence if you already know the bit values.

In which case, why is the computation output instantly readable within the confines of a short distance, but not at a longer discernably "FTL" distance (ie. a non-local distance)? In what crucial way does computation differ from communication?

This paper ruled out local hidden variable theories. It does not (and doesn't claim to) build an ansible, or show that one is possible.

"As of 2015 [i.e. this paper], all Bell tests have found that the hypothesis of local hidden variables is inconsistent with the way that physical systems behave."

But from what I remember, local hidden variable theories were postulated to justify being able to determine the outcome of quantum measurements (eg. reading the output of your computation) without violating the "no FTL"/causality stuff.
So by ruling out local hidden variables, that experiment is saying causality/"no FTL" can be violated -- ie. there can be FTL.

[sanman]

[Twark_Main]

Okay, fair enough - but...

Yep, I thought you might go down exactly this road, which is why I wrote the next paragraph.

I just wanted you to get an answer to your question.

Furthermore, even if you can solve this problem, you still can't build an ansible. All you'd be doing is transmitting regular bits to Bob, which happens at the speed of light. The only things that "goes FTL" is decoherence, but there's no decoherence if you already know the bit values.

In which case, why is the computation output instantly readable within the confines of a short distance, but not at a longer discernably "FTL" distance (ie. a non-local distance)

That's not what they mean by "non-local."

To a quantum physicist, anything further than the reduced Compton radius is "non-local." It's not about short vs long distances on a human scale.

But from what I remember, local hidden variable theories were postulated to justify being able to determine the outcome of quantum measurements (eg. reading the output of your computation) without violating the "no FTL"/causality stuff.
So by ruling out local hidden variables, that experiment is saying causality/"no FTL" can be violated -- ie. there can be FTL.

Remembered wrong. Happens to the best of us. 







Bonus video on Bell's Inequality:

[sanman]

Okay, fair enough - but...

Yep, I thought you might go down exactly this road, which is why I wrote the next paragraph.

I just wanted you to get an answer to your question.

Furthermore, even if you can solve this problem, you still can't build an ansible. All you'd be doing is transmitting regular bits to Bob, which happens at the speed of light. The only things that "goes FTL" is decoherence, but there's no decoherence if you already know the bit values.

In which case, why is the computation output instantly readable within the confines of a short distance, but not at a longer discernably "FTL" distance (ie. a non-local distance)

That's not what they mean by "non-local."

To a quantum physicist, anything further than the reduced Compton radius is "non-local." It's not about short vs long distances on a human scale.

But from what I remember, local hidden variable theories were postulated to justify being able to determine the outcome of quantum measurements (eg. reading the output of your computation) without violating the "no FTL"/causality stuff.
So by ruling out local hidden variables, that experiment is saying causality/"no FTL" can be violated -- ie. there can be FTL.

Remembered wrong. Happens to the best of us. 

Okay, let's split this off into another thread under the New Physics forum, so that I don't derail this thread.

[edzieba]

But Ed, in quantum computing you do get to choose whether the bit is 1 or 0, because you're loading your information into the entangled qubits.

You don't, because loading inputs bits into the computer and getting some answer bits out is how a classical computer works. It is not how a quantum computer works.

Greatly simplified: you can arrange your Qbits into the representation of a function you want to compute, but observing those Qbits means you get one of the possible arrangements of bits that satisfies that function, whilst also 'breaking' your Qbits. If you want to get multiple results (i.e. to build up the probability distribution you were trying to find in the first place with your function) you re-create that function multiple times and get multiple answers. If that sounds weird and awkward and a terrible way to go about computation - yes it is, for classical computation. Quantum computers are absurdly worthless for performing the functions we are used to computer performing. That are exceptional only at performing functions that can be represented in the way quantum computation works, and are valuable for the edge cases where a function can both be expressed in a way that even makes sense to attempt quantum computation on, and that function also has some property that makes it unusually difficult to compute classically.