Slashdot Mirror
First 'Quantum Computer Chips' Demonstrated
holy_calamity writes "The first quantum computer chips have been made by two US groups, New Scientist reports. Both NIST and Yale have demonstrated chips where information was transferred between two superconducting qubits using a 'quantum bus'. The bus is made from a cavity that traps a single microwave photon as a standing wave — the NIST group also managed to use the bus to store data from one qubit for a short time. 'After encoding information in one qubit, they transferred it into the cavity for 10 nanoseconds before transferring it to the other qubit. Yale's chip used qubits around 1-micron square built on silicon, while NIST used larger 10-square-micron qubits on top of sapphire. In both prototypes, the bus between the qubits was between five and seven millimeters long.'"
maybe I'm first, maybe I'm not.
Is it just me, or do you hate it when people say "Is it just me..."?
Howdy. I don't claim to understand all of this. However, the more I read, the more I am convinced the universe makes no sense. I am waiting for the guy who is dreaming all of this to wake up and for all of us to stop existing.
See my journal for slashdot ID's by year. Mine created in 2005. http://slashdot.org/journal/289875/slashdot-ids-by-year
I was going to tell you, but I changed the outcome by reading it!
Well, back to rejecting software patent applications.
What if you put a cat inside that Beowulf cluster you're not imagining?
... but will it run Linux? (Or will it run and not run Linux at the same time?)
You can't know how many cats wide it is or fast it is until you transfer data over it.
To clarify both sides, unless I've missed something in the last couple of years, AES was designed[1] with the possibility of quantum computing in mind and the solution is to use double the bit length you'd otherwise need (which is the same for at least some elliptic curve-based Public Key algorithms but for different algorithmic reasons). Is this still computable by standard computers? Yes. Does it make it harder to use "strong" crypto in limited hardware, a little. Could there be improved algorithms down the road that push it to the point that it takes the same order of time to decrypt on standard computers algorithms knowing the key as it does to decrypt (break) on quantum computers without knowing the key? Possibly (in the sense that I don't know of any proofs showing limits on efficiency gains etc.).
[1]Designed is probably not the right word, but basically, brute force searching of 128bit symmetric keys is believed to be secure in the sense that using all atoms as non-quantum computers would find it some point after expected heat death of universe. However, quantum computers can (being lazy, start at wikipedia's entry on cryptoanalysis, look for grover algorithm) do a brute force search in quadratic time (so 128bits would take on the order of 2^64 steps which is much more tractable... however, using 256bit AES keys (which would otherwise be overkill for most things) now take on the order of 2^128 steps which again hits that whole heat death thing, unless either a better algorithm comes out or someone comes out with some sort of hyper-quantum-computing idea)
I work in a physics lab, and a few days ago the unimaginable happened.
A quantum experiment had gone horribly wrong, going completely out of control and destroying itself in the process.
The devastation was unimaginable.
All that was left of the experiment was a crater, almost a nanometre across.
As soon as we get the electron microscope on it, I'll have more details of what went wrong.
Why is it things like this never have pictures? I wanna see pictures. Its no fun to read about things that you don't (quite) understand unless you can ooh and ahh at pictures while you pretend to understand. Then you can point at your screen and say "See? Its THAT piece. That's what makes it work. Its the, err.. umm, thing that makes it work!"
No, this is not correct. While it's true that if you put N qubits together in the correct superposition, you can make a state that is "equally spread out" over all 2^N possibilities, you cannot make the computer "favor" the correct one (at least not in the sense you are implying). Using Shor's algorithm you can factor a number in O((log N)^3), which is an exponential improvement to crack RSA. And yes, I am a physicist working on quantum computing.
The one time pad, where the key length = message length is still safe as long as you never reuse the key. (the "one time" in one time pad.
As simple proof of this is that for any encrypted text of length N, there exists a key also of length N that will decrypt the etext to any plain text of length N. Therefore there is no way for an attacker to determine if an attempted key is valid or not. There if an attacker were to try every single key of length N, which is possible on some super large future quantum computer, all he will get out is every single decryption of length N, with no way to determine which is correct.
Suppose the plain text was "attack at dawn" and the etext was "xbdhgfhwteriur". After the attacker used his q-computer he'd have "attack at dawn", "attach at noon" and "attack at fred", along with 64 quintillion other combinations.
All ideas^H^H^H^H^Hprocesses in this post are Patent Pending. (as well as the process of patenting all postings)