Flawed Diamonds May Be the Key to Quantum Computing

[Angus565]

So long story short, diamonds are magic and we can harness their power to make bad-ass super computers?

 

[thiosk]

Diamond based spintronics. Neat.

But the real sex is when you include boron instead.

 

[Twilight_guy]

Impossible to hack? No. Not any more then current computing algorithms. In fact today's "impossible" to hack encryption is based on a system which makes it "impossible" by making it take hundreds of years or more to crack. Quantum computing might just blow the lid off that by making it easier to hack quickly. Quantum can lead to new ways to avoiding being hacked but who's to say there isn't a computational model higher then quantum that makes quantum encryption look like a joke? The only impossible to hack thing would have to be based on something computable and that makes it impossible to encrypt in the first place.
It seems much easier to just unplug your computer from all networks and regulate who has access to it to me.

 

[CrazyCapnMorgan]

Flawed Diamonds May Be the Key to Quantum Computing

Is that so?

Well then! Looks like I'm off to do some more collecting in Diablo 2, I suppose...

 

[FluxCapacitor]

Folks, quantum computing is explicitly NOT the same as trinary computing - as system utilising 0, 1 and 2 as values. There was actually a fair amount of work done on trinary by the Russians way back when, but it's been completely superseded by the optimised binary everyone uses today.

Quantum comuting is something else. It involves completely rewriting the way that we calculate solutions, using completely new mathematical rules that are likely to shake down to WAAAY fewer operations per problem. How? Right now we don't really know - that'll be a job for some really sharp mathematicians for the next few decades.

But what we can do is apply known physics to the concepts presented to get some idea of end applications for this awesome stuff. For example, quantum based comms traffic will likely be faster than light, and unhackable in the conventional sense. It will literally be impossible to do things like passively listen to network traffic, because through the utter craziness of quantum physics packets will be changed by being observed in any way. When the packet reaches the destination, the receiver can tell it has been listened to, and knows the network has been compromised.

TL;DR - Anyone wh says they really know exactly what quantum computing will bring doesn't understand quantum computing.

 

[Angel Molina]

We're building legitimate technology with irregularly shaped crystals now? What are we, the Protoss?

This^^, I laughed so hard that I scared my cat off the bed.

Hmm... There's always some article claiming that some research has found a super cool way to do something or has come up with a break through. But in the end, it always comes down to its marketing or how well it could sell. If no one wants and/or has the funds to get a quantum computer built, then the information/research just sits there and is eventually forgotten...

But on the bright side... diamonds sure are shiny! =D

 

[Knife]

Folks, quantum computing is explicitly NOT the same as trinary computing - as system utilising 0, 1 and 2 as values. There was actually a fair amount of work done on trinary by the Russians way back when, but it's been completely superseded by the optimised binary everyone uses today.

Quantum comuting is something else. It involves completely rewriting the way that we calculate solutions, using completely new mathematical rules that are likely to shake down to WAAAY fewer operations per problem. How? Right now we don't really know - that'll be a job for some really sharp mathematicians for the next few decades.

But what we can do is apply known physics to the concepts presented to get some idea of end applications for this awesome stuff. For example, quantum based comms traffic will likely be faster than light, and unhackable in the conventional sense. It will literally be impossible to do things like passively listen to network traffic, because through the utter craziness of quantum physics packets will be changed by being observed in any way. When the packet reaches the destination, the receiver can tell it has been listened to, and knows the network has been compromised.

TL;DR - Anyone wh says they really know exactly what quantum computing will bring doesn't understand quantum computing.

Lets review, you don't know what quantum computing really is, but its definitely isn't trinary computing. Its somehow supposed to improve problems running time, but you dont know how, and noones knows how and probably will not know for an extended period of time. Right?
So basically we are left with a device that we can build at a very high cost that we have no use for (at the very least for the time being).

Now I'm no quantum physicist, but I know my fair share on computer architecture and algorithms. We have physical phenomenon that can exist in 1 of 3 possible states, which we can use as storage device for 3 numbers (sounds like trinary computing so far). Then we add the ability to transfer that state/information from one place to another (presumably at high speeds, atleast as high as current computers). So what we get is this - a bit better storage device but nothing stellar, if we use it as a processor then we will only gain speed if we replace the RAM and the hard drive with similar devices (which admitedly could run at much faster speeds), the downside to all of this is of rather practical nature - diamonds aren't cheap, replacing silicon chips with diamonds crystals would be very costly. The question is then whether or not the speed boost worth the trouble of manufacturing such computers (similar problems have arised with RAM and hard discs beforehand, where RAM is slower than CPU but far cheaper to produce, and hard discs are even slower than RAM but even more cheap, theoretically a computer can be made out of CPU alone that would work at much higher speeds than standard PCs but it would be worth a fortune, the cost to make one would make it unfeasible for anyone to buy so computer companies discarded that idea. I'm afraid that similar issue may arise with quantum computers as well, even if it will be faster it will most likely be far more expensive than what its worth, unless somehow within a couple decades diamonds will be dirt cheap).

 

[Valanthe]

snip

You sir, win an internet.

 

[pneuma08]

Spoiler

Folks, quantum computing is explicitly NOT the same as trinary computing - as system utilising 0, 1 and 2 as values. There was actually a fair amount of work done on trinary by the Russians way back when, but it's been completely superseded by the optimised binary everyone uses today.

Quantum comuting is something else. It involves completely rewriting the way that we calculate solutions, using completely new mathematical rules that are likely to shake down to WAAAY fewer operations per problem. How? Right now we don't really know - that'll be a job for some really sharp mathematicians for the next few decades.

But what we can do is apply known physics to the concepts presented to get some idea of end applications for this awesome stuff. For example, quantum based comms traffic will likely be faster than light, and unhackable in the conventional sense. It will literally be impossible to do things like passively listen to network traffic, because through the utter craziness of quantum physics packets will be changed by being observed in any way. When the packet reaches the destination, the receiver can tell it has been listened to, and knows the network has been compromised.

TL;DR - Anyone wh says they really know exactly what quantum computing will bring doesn't understand quantum computing.

Lets review, you don't know what quantum computing really is, but its definitely isn't trinary computing. Its somehow supposed to improve problems running time, but you dont know how, and noones knows how and probably will not know for an extended period of time. Right?
So basically we are left with a device that we can build at a very high cost that we have no use for (at the very least for the time being).

Now I'm no quantum physicist, but I know my fair share on computer architecture and algorithms. We have physical phenomenon that can exist in 1 of 3 possible states, which we can use as storage device for 3 numbers (sounds like trinary computing so far). Then we add the ability to transfer that state/information from one place to another (presumably at high speeds, atleast as high as current computers). So what we get is this - a bit better storage device but nothing stellar, if we use it as a processor then we will only gain speed if we replace the RAM and the hard drive with similar devices (which admitedly could run at much faster speeds), the downside to all of this is of rather practical nature - diamonds aren't cheap, replacing silicon chips with diamonds crystals would be very costly. The question is then whether or not the speed boost worth the trouble of manufacturing such computers (similar problems have arised with RAM and hard discs beforehand, where RAM is slower than CPU but far cheaper to produce, and hard discs are even slower than RAM but even more cheap, theoretically a computer can be made out of CPU alone that would work at much higher speeds than standard PCs but it would be worth a fortune, the cost to make one would make it unfeasible for anyone to buy so computer companies discarded that idea. I'm afraid that similar issue may arise with quantum computers as well, even if it will be faster it will most likely be far more expensive than what its worth, unless somehow within a couple decades diamonds will be dirt cheap).

Woah, I think you're mixing things up a bit here. A CPU is a processing unit that actually computes information, whereas RAM is quick, cheap, non-permanent memory and hard drives (HDD) and related solid state drives (SSD) are slower, more permanent storage (with HDD being dirt cheap and SSD being faster but less cheap). I think you may be referring to the cache on a CPU, which is also temporary storage, and "virtual memory" which is using the HDD/SSD space for temporary memory, and speed implications thereof, although the implications you draw aren't entirely correct; that is, while it is theoretically possible to build a computer entirely out of cache memory, the reason why people don't isn't monetary concerns but rather that it needs to be physically close to the CPU in order to function at cache-like speeds.

As for quantum computing, if it takes ram-access speeds to use qubits, that's pretty slow on the order of computing potential. I mean, speaking of cache memory, that alone is many times faster than RAM and still slower than just using registers (that is, storage attached to the CPU itself, which is used for its calculations). That being the case, we won't have quantum computing overtaking regular computing, as a CPU with qubit registers would be multitudes slower (by my notes about 80 times slower than a 2GHz processor). Rather, the power of quantum computing is that it propagates at the speed of RAM access across many km. That's superfast for that distance. And if it can do so without error, that would be amazing.

Oh, and yes, it would be "unhackable", or at least a single connection is no more hackable (and in fact probably much less hackable, with our current understand of physics) than plugging two computers together with a physical cable and sending data between them. No, the reason why the internet is so vulnerable is because it relies on many, many computers working together to propagate information across long distances. Of course this also means that a quantum network would be no less hackable than a regular network, assuming we use the same algorithms that we do today (which would be very likely; it would just replace the physical layer of networking with quantum thingamabobs instead of metal wiring). Meanwhile, really important data links could be forged from a single, direct connection, which is theoretically possible but not really feasible with current technologies (and such a long wire would be hell to maintain and easy to break, unlike quantum entanglement).

 

[FluxCapacitor]

Lets review, you don't know what quantum computing really is, but its definitely isn't trinary computing. Its somehow supposed to improve problems running time, but you dont know how, and noones knows how and probably will not know for an extended period of time. Right?
So basically we are left with a device that we can build at a very high cost that we have no use for (at the very least for the time being).

Now I'm no quantum physicist, but I know my fair share on computer architecture and algorithms. We have physical phenomenon that can exist in 1 of 3 possible states, which we can use as storage device for 3 numbers (sounds like trinary computing so far). Then we add the ability to transfer that state/information from one place to another (presumably at high speeds, atleast as high as current computers). So what we get is this - a bit better storage device but nothing stellar, if we use it as a processor then we will only gain speed if we replace the RAM and the hard drive with similar devices (which admitedly could run at much faster speeds), the downside to all of this is of rather practical nature - diamonds aren't cheap, replacing silicon chips with diamonds crystals would be very costly. The question is then whether or not the speed boost worth the trouble of manufacturing such computers (similar problems have arised with RAM and hard discs beforehand, where RAM is slower than CPU but far cheaper to produce, and hard discs are even slower than RAM but even more cheap, theoretically a computer can be made out of CPU alone that would work at much higher speeds than standard PCs but it would be worth a fortune, the cost to make one would make it unfeasible for anyone to buy so computer companies discarded that idea. I'm afraid that similar issue may arise with quantum computers as well, even if it will be faster it will most likely be far more expensive than what its worth, unless somehow within a couple decades diamonds will be dirt cheap).

See, the 3 possible states is where you come unstuck - I'll try and give a bit of an example. At the moment, all of our computing hardware is based on a process-based paradigm where if a + b = c, you need to know a and b before you can calculate c. We can write systems that will predict how varying values of b will affect the end result at c, but the system does that by running the calculation a bunch of times for all values of b. At bit level, with b only equal to 1 or 0, this isn't that much extra work (only 2 cycles instead of 1) but at byte level it's 256 different cycles and at 128-bit encryption level that's like 3.4 x 10^38 extra operations. (That's whay it takes so long to decrypt without the key in modern system architectures - you basically do this with a = encrypted data, b = key and check c to see if it makes any sense, over and over and over again...)

With quantum computing we can express in a single operation the superposition of all possible states for b. So in a single processor state we could represent both b= 0 and b= 1, and have a result for c that is meaningful in quantum processing. Once again, it doesn't seem like much at qubit level but the advantage scales exponentially according to how many qubits you can build your system to handle simultaneously. You'd only need to build a quantum system of 128 qubits to simultaneously represent every possible 128-bit encryption key, and then it's just a question of plugging anyone's data into the system as a and selecting the c that makes sense. Single operation unencryption of ANYONE's traditional database - that is some serious shit. Granted, at the moment they're still working on the maths of building meaningful qubytes (8 qubits), but that is the sort of capability that these things could get to in the end. So yes, I think that they will indeed get built - by governments, research institutions and the private sector. Will you and I get them in our home PCs? Probably not until the theory is established and economies of scale make them cheap to mass produce.

Also, did you know that diamond can be grown artificially? For that matter, since it's flawed diamonds that are good for this technique, they would almost have to be grown in order to miniaturize the whole thing and compress many of them into a chip. This is not the first technology to do this - NASA used artificial sapphire in its hardware for years. If the benefits are great enough (and they are), someone will justify the expense.

 

[Knife]

Woah, I think you're mixing things up a bit here. A CPU is a processing unit that actually computes information, whereas RAM is quick, cheap, non-permanent memory and hard drives (HDD) and related solid state drives (SSD) are more permanent storage. I think you may be referring to the cache on a CPU, which is also temporary storage, and "virtual memory" which is using the HDD/SSD space for temporary memory, and speed implications thereof, although the implications you draw aren't entirely correct; that is, while it is theoretically possible to build a computer entirely out of cache memory, the reason why people don't isn't monetary concerns but rather that it needs to be physically close to the CPU in order to function at cache-like speeds.

As for quantum computing, if it takes ram-access speeds to use qubits, that's pretty slow on the order of computing potential. I mean, speaking of cache memory, that alone is many times faster than RAM and still slower than just using registers. That being the case, we won't have quantum computing overtaking regular computing. Rather, the power of quantum computing is that it propagates at the speed of RAM access across many km. As far as networking is concerned, that's superfast. And if it can do so without error, that would be amazing.

Oh, and yes, it would be "unhackable", or at least a single connection is no more hackable (and in fact probably much less hackable, with our current understand of physics) than plugging two computers together with a physical cable and sending data between them. No, the reason why the internet is so vulnerable is because it relies on many, many computers working together to propagate information across long distances. Of course this also means that a quantum network would be no less hackable than a regular network, assuming we use the same algorithms that we do today (which would be very likely; it would just replace the physical layer of networking with quantum thingamabobs instead of metal wiring). Meanwhile, really important data links could be forged from a single, direct connection, which is theoretically possible but not really feasible with current technologies (and such a long wire would be hell to maintain and easy to break, unlike quantum entanglement).

I was refering to CPU registers - fast but expensive, bringing data from RAM over to the CPU registers takes a lot of time (compared to arithmetic opertaion on registers and constants), if instead of storing your data in RAM you would make an array of registers directly available for CPU use it would certainly cut down on time needed for caching - which is by far bigger than any other than the time needed for other operations. Similarly when your computer runs out of memory on RAM it copies it over to your HDD which takes even more time to access. A lot of time is saved by limiting the amount of reading/writing to RAM and HDD, and background caching/buffering, but with enough processes running on a computer simultaneously it is still unavoidable. If RAM and/or HDD were to be replaced by faster technology it would boost up computer's speed - physically RAM and HDD aren't that far away from the CPU, it is how their reading/writing is implemented that affects their speed.

On the "unhackable" business, it not really of any real life use, most hacking is done through software not hardware, by using an existing IO channel and sending nasty programs over to the computer which are disguised as nothing out of the ordinary. The computer then activates these programs and voila... hacking in progress. Basically unless the two computers are completely offline (except for each other) then this little feature hardly helps.

 

[FluxCapacitor]

Spoiler

Woah, I think you're mixing things up a bit here. A CPU is a processing unit that actually computes information, whereas RAM is quick, cheap, non-permanent memory and hard drives (HDD) and related solid state drives (SSD) are slower, more permanent storage (with HDD being dirt cheap and SSD being faster but less cheap). I think you may be referring to the cache on a CPU, which is also temporary storage, and "virtual memory" which is using the HDD/SSD space for temporary memory, and speed implications thereof, although the implications you draw aren't entirely correct; that is, while it is theoretically possible to build a computer entirely out of cache memory, the reason why people don't isn't monetary concerns but rather that it needs to be physically close to the CPU in order to function at cache-like speeds.

As for quantum computing, if it takes ram-access speeds to use qubits, that's pretty slow on the order of computing potential. I mean, speaking of cache memory, that alone is many times faster than RAM and still slower than just using registers (that is, storage attached to the CPU itself, which is used for its calculations). That being the case, we won't have quantum computing overtaking regular computing, as a CPU with qubit registers would be multitudes slower (by my notes about 80 times slower than a 2GHz processor). Rather, the power of quantum computing is that it propagates at the speed of RAM access across many km. That's superfast for that distance. And if it can do so without error, that would be amazing.

Oh, and yes, it would be "unhackable", or at least a single connection is no more hackable (and in fact probably much less hackable, with our current understand of physics) than plugging two computers together with a physical cable and sending data between them. No, the reason why the internet is so vulnerable is because it relies on many, many computers working together to propagate information across long distances. Of course this also means that a quantum network would be no less hackable than a regular network, assuming we use the same algorithms that we do today (which would be very likely; it would just replace the physical layer of networking with quantum thingamabobs instead of metal wiring). Meanwhile, really important data links could be forged from a single, direct connection, which is theoretically possible but not really feasible with current technologies (and such a long wire would be hell to maintain and easy to break, unlike quantum entanglement).

Also good points - I'd say this is the computer engineer's answer, as opposed to the computer scientist's answer... I agree that quantum computing won't replace traditional computing in everyday society, most stuff will stay binary since it's WAAAAY easier to get your head around as a programmer. But top end cryptanalysis will all move to quantum on a long enough timeline, with comms and supercomputing getting into the action as well.


EDIT: Oh, and @Knife - you won't be able to slip in packets that look like regular traffic into a quantum network. Your rogue packets would be detected instantly at hardware level (long before your packets are reassembled into a program) because they would not be correctly entangled. Your only real option is social engineering.

 

[Knife]

See, the 3 possible states is where you come unstuck - I'll try and give a bit of an example. At the moment, all of our computing hardware is based on a process-based paradigm where if a + b = c, you need to know a and b before you can calculate c. We can write systems that will predict how varying values of b will affect the end result at c, but the system does that by running the calculation a bunch of times for all values of b. At bit level, with b only equal to 1 or 0, this isn't that much extra work (only 2 cycles instead of 1) but at byte level it's 256 different cycles and at 128-bit encryption level that's like 3.4 x 10^38 extra operations. (That's whay it takes so long to decrypt without the key in modern system architectures - you basically do this with a = encrypted data, b = key and check c to see if it makes any sense, over and over and over again...)

With quantum computing we can express in a single operation the superposition of all possible states for b. So in a single processor state we could represent both b= 0 and b= 1, and have a result for c that is meaningful in quantum processing. Once again, it doesn't seem like much at qubit level but the advantage scales exponentially according to how many qubits you can build your system to handle simultaneously. You'd only need to build a quantum system of 128 qubits to simultaneously represent every possible 128-bit encryption key, and then it's just a question of plugging anyone's data into the system as a and selecting the c that makes sense. Single operation unencryption of ANYONE's traditional database - that is some serious shit. Granted, at the moment they're still working on the maths of building meaningful qubytes (8 qubits), but that is the sort of capability that these things could get to in the end. So yes, I think that they will indeed get built - by governments, research institutions and the private sector. Will you and I get them in our home PCs? Probably not until the theory is established and economies of scale make them cheap to mass produce.

Also, did you know that diamond can be grown artificially? For that matter, since it's flawed diamonds that are good for this technique, they would almost have to be grown in order to miniaturize the whole thing and compress many of them into a chip. This is not the first technology to do this - NASA used artificial sapphire in its hardware for years. If the benefits are great enough (and they are), someone will justify the expense.

I think you might be off somewhat with the math there (either that or I'm off with the physics), if by using 2 states we can decrypt x amount of keys then by using 3 states we should be able to decrypt Log2(3)*x keys - here is the thing, correct me if I'm wrong but say b is 0 or 1, wouldn't a+b be undefined without knowing whether b is 0 or b is 1? And if it is how do you compare it to c? It seems like you're trying to solve an exponential problem in polynomial time and it goes against everything I know.

Yes, I did know about artificial diamonds - they are still expensive to produce, luckily however from what I remember they are mostly flawed and used for industry, like drilling, so I guess they would probably be good for this job.

 

[Knife]

EDIT: Oh, and @Knife - you won't be able to slip in packets that look like regular traffic into a quantum network. Your rogue packets would be detected instantly at hardware level (long before your packets are reassembled into a program) because they would not be correctly entangled. Your only real option is social engineering.

That's not quite what I had in mind, as I said earlier if the 2 computers are only connected to each other and noone else then it would be "unhackable" but so would 2 computers standing in the same room connected into a LAN. The problem of hacking starts when one of the computers goes online and downloads something it shouldn't have, from that point onward the technology of it all is inconsequential.

Not that long distance, wireless LAN isn't cool in and of itself.

 

[Thorvan]

When quantum computers go mainstream, people will ask why we played computer games that were less realistic than reality.

I'll go grab a pair of hipster glasses, just in case I'm still alive by then.

 

[Tharwen]

Impossible-to-hack now?

Yeah, no more so than current ones.

Eh... quantum entangled particles are pretty much completely unhackable.

Issue is, nobody targets the encrypted data even now, seeing it'd take ages to break that encryption. But you got to get the data unencrypted at some point, and that's where you strike if you want to acquire the data.

OK, I agree with that.

Um... I don't really have much else to say...

Hurrah for quantum computers?

 

[FluxCapacitor]

Spoiler

I think you might be off somewhat with the math there (either that or I'm off with the physics), if by using 2 states we can decrypt x amount of keys then by using 3 states we should be able to decrypt Log2(3)*x keys - here is the thing, correct me if I'm wrong but say b is 0 or 1, wouldn't a+b be undefined without knowing whether b is 0 or b is 1? And if it is how do you compare it to c? It seems like you're trying to solve an exponential problem in polynomial time and it goes against everything I know.

Yes, I did know about artificial diamonds - they are still expensive to produce, luckily however from what I remember they are mostly flawed and used for industry, like drilling, so I guess they would probably be good for this job.

Yep, you're off with the physics, I'm afraid. It's not 3 states - it's 2 states, plus a bizarre quantum mechanical superposition of the 2 states, which can be functionally used as either state simultaneously. It's weird, I know, but it is NOT trinary. Thus, a + b can be defined without knowing b (or a for that matter!), because we've defined b as the superposed state. The processor is happy with this, and doesn't generate errors. It is then possible to evaluate for c without explicitly knowing b. Your output c is in a quantum state which is the superposition of all possible inputs at b added to a. It is then the system architecture's job to locate the real portion of that superposed state by... doing something nifty. Will it resolve to real simply by being observed, as in the Copenhagen model? Will we need to do some kinda funky parallel matrices thing in conventional binary? I certainly don't have the answers to these questions, and I think that no one does at the moment really - quantum computing is a very young field and it's really still in the baby steps as far as physical experimentation goes. But the possibilities are a comp sci nerd's wet dream.

I totally agree with you, it is weird and semi-incomprehensible and goes against everything you know, and that's the point - quantum physics is weird, man. Ask a physicist.

 

[Knife]

Yep, you're off with the physics, I'm afraid. It's not 3 states - it's 2 states, plus a bizarre quantum mechanical superposition of the 2 states, which can be functionally used as either state simultaneously. It's weird, I know, but it is NOT trinary. Thus, a + b can be defined without knowing b (or a for that matter!), because we've defined b as the superposed state. The processor is happy with this, and doesn't generate errors. It is then possible to evaluate for c without explicitly knowing b. Your output c is in a quantum state which is the superposition of all possible inputs at b added to a. It is then the system architecture's job to locate the real portion of that superposed state by... doing something nifty. Will it resolve to real simply by being observed, as in the Copenhagen model? Will we need to do some kinda funky parallel matrices thing in conventional binary? I certainly don't have the answers to these questions, and I think that no one does at the moment really - quantum computing is a very young field and it's really still in the baby steps as far as physical experimentation goes. But the possibilities are a comp sci nerd's wet dream.

I totally agree with you, it is weird and semi-incomprehensible and goes against everything you know, and that's the point - quantum physics is weird, man. Ask a physicist.

How do you cast magic missile? Quantum physics man! By the power of nifty!

My problem is if 'b' contains all possible outcomes then adding it to 'a' would result in all possible outcomes as well, so 'a+b' is everything, which makes 'a' obsolete, basically your output after comparing to 'c' is all possible inputs of 'a' decrypted, without any indication which one of them is the decryption of our specific 'a'. That kind of kills the idea of encryption-decryption. There is no way to pinpoint the one we need out of all the possible outcomes.

 

[pneuma08]

With quantum computing we can express in a single operation the superposition of all possible states for b. So in a single processor state we could represent both b= 0 and b= 1, and have a result for c that is meaningful in quantum processing. Once again, it doesn't seem like much at qubit level but the advantage scales exponentially according to how many qubits you can build your system to handle simultaneously. You'd only need to build a quantum system of 128 qubits to simultaneously represent every possible 128-bit encryption key, and then it's just a question of plugging anyone's data into the system as a and selecting the c that makes sense. Single operation unencryption of ANYONE's traditional database - that is some serious shit. Granted, at the moment they're still working on the maths of building meaningful qubytes (8 qubits), but that is the sort of capability that these things could get to in the end. So yes, I think that they will indeed get built - by governments, research institutions and the private sector. Will you and I get them in our home PCs? Probably not until the theory is established and economies of scale make them cheap to mass produce.

Let me see if I understand this correctly: if a is (1), and b is (0 or 1), then c, defined as a+b, is (1+0 or 1+1), and all of these states are meaningful in quantum computing. Yeah, that'd be pretty revolutionary. I don't even know how that could be useful at all, even though it probably is...? Still, the speed hit is significant enough to be a drawback in many cases, and if I was worried about strict optimization I wouldn't want to trade my 100-or-so instructions for a single qubit operation all the time; it would depend on the algorithm involved. So perhaps we'll see a "quantum module" with a quantum processor or something built into computers rather than pure quantum computers. I'm picturing a "quantum card" that fits into a PCIe slot in my head right now (although in actuality it would more likely be a component on a motherboard or perhaps even a cpu chip).

If RAM and/or HDD were to be replaced by faster technology it would boost up computer's speed - physically RAM and HDD aren't that far away from the CPU, it is how their reading/writing is implemented that affects their speed.

Firstly, while this statement is true, the article clearly states such quantum computing propagates data at RAM-speeds and only exists for a few seconds. So, it wouldn't makes sense to replace either RAM or HDD with qubits.

Secondly, distance is a factor in propagation of data on a motherboard. If you put registers on the opposite size of the motherboard they will most definitely be slower to access than those built into the chip; one of the reasons why they are so darned expensive is because they have to be physically close to the parts of the CPU, and so room has to be made for them.

And as for quantum networks, if FluxCapacitor is on the ball here, it's sounding incredibly expensive since every single link has to be forged in the fires of quantum entanglement, and you would need a pair of magic quantum boxes for every single link in the internet. It would certainly solve identity issues, though, if the connections must be all unique.

That said, I'm certain there will be some avenues of attack left open on a quantum network. I think the best thing we can say is that it would be different. Quantum links seem very solid, though, especially when set alongside comparable technologies like wifi and 3/4G.

Man, this is getting well beyond realms familiar to me, as my knowledge of quantum mechanics is pretty basic and probably outdated, and I'm not terribly intimate with networking either.

Anyway, if I had to do a bit of prediction, I'd say the network of the future would look very much like the network of today except that certain key connections will be done via quantum entanglement instead of wire, probably all on the backbone, away from consumers. That said, general quantum computing may impact the state of routers (assuming a large increase in speed of quantum algorithms, they may have a quantum component to do their routing). The end result: faster internet for everyone! Hooray!