4 ms·
Besides scientists, I don't really see quantum computers as being useful computationally. There's breaking cryptography, but then you're spending billions to in
by MichaelBurge 9y ago
Besides scientists, I don't really see quantum computers as being useful computationally. There's breaking cryptography, but then you're spending billions to inconvenience people into moving to a different quantum-resistant algorithm.
What about creating cryptography? Wikipedia has a few interesting applications:
https://en.wikipedia.org/wiki/Quantum_cryptography https://en.wikipedia.org/wiki/Quantum_cryptography
I'm not a cryptographer - are security people excited about building all sorts of weird quantum protocols?
> “Then you let the computer check all possible solutions essentially—or a very large combination of them—and come back with an answer,” he says. In a quantum computer, there’s no mathematician cracking the problem, he says. “The laws of physics crack the problem for you.”
That sounds more like an NTM than a quantum computer. An NTM would turn its owner into a God.
> Superposition is the mind-bending observation that a particle can be in two states at the same time. Bring out your ruler to get a measurement, however, and the particle will collapse into one state or the other. And you won’t know which until you try, except in terms of probabilities.
It's not probabilities because people think of those as real numbers. The coefficients are complex numbers. There's a probability monad, for example, but no quantum monad:
http://twistedoakstudios.com/blog/Post5485_what-isnt-a-monad http://twistedoakstudios.com/blog/Post5485_what-isnt-a-monad
The article sort of mentions it later on when they quote Scott Aaronson, but I suspect the author didn't understand and mentally reverted back to a probability model.
- hannob 9y ago> I'm not a cryptographer - are security people excited about building all sorts of weird quantum protocols? No. In reality so-called quantum cryptography is completely impractical. Imagine an Internet where you can't send messages to people more than a few hundred kilometers away, with no wifi and no mobile Internet and which only works if you have pre-shared keys. That's the Quantum Internet (which apparently the EU just decided to fund with billions).
- doktrin 9y ago>> > I'm not a cryptographer - are security people excited about building all sorts of weird quantum protocols? > No. In reality so-called quantum cryptography is completely impractical What about building quantum safe protocols? Is that as a field any more interesting and/or viable?
- hannob 9y ago> What about building quantum safe protocols? Is that as a field any more interesting and/or viable? Yes. Plenty of stuff happening in that field. The challenge is: Right now we can choose between probably safe, but impractical (mcbits, sphincs) and practical, but with big questionmarks about how safe they are (newhope, sidh). The problem is: We really only trust crypto algos if they have been tested and researched over a long time, but lots of these things are pretty new.
- CiPHPerCoder 9y agoThat's called post-quantum cryptography, not quantum cryptography.
- doktrin 9y ago>> What about building quantum safe protocols? > That's called post-quantum cryptography, not quantum cryptography. As I understand it, "quantum safe" and "post-quantum" describe the same thing. I never used the term 'quantum cryptography'
- CiPHPerCoder 9y agoI think I meant to respond to the parent comment, not yours. But that was hours ago.
- BatFastard 9y agoIsn't this IDEAL for secure computing? Really its not that much different than how things work now. When we send a message around the world, its going through dozens of computers to get there. And wifi can always be an end point extension.
- russdill 9y agohttps://en.wikipedia.org/wiki/Quantum_neural_network https://en.wikipedia.org/wiki/Quantum_neural_network
- moyix 9y agoI think one of the killer apps is the purpose Feynman originally envisioned – simulating quantum systems: https://people.eecs.berkeley.edu/~christos/classics/Feynman.pdf https://people.eecs.berkeley.edu/~christos/classics/Feynman.... I wasn't able to dig up too much, but the introduction to a paper [1] by some Google folks provides a bit of insight into what might be possible: > In particular, quantum simulation of molecular energies, which enables numerically exact prediction of chemical reaction rates, promises significant advances in our understanding of chemistry and could enable in silico design of new catalysts, pharmaceuticals, and materials. [1] https://journals.aps.org/prx/pdf/10.1103/PhysRevX.6.031007 https://journals.aps.org/prx/pdf/10.1103/PhysRevX.6.031007
- germinalphrase 9y agoSeth Lloyd did a talk at the Long Now Foundation in 2016 about his research into quantum computing. He made the point that his research is (partially?) funded by the NSA - which was great because the NSA's position on quantum computing/cryptography was basically "we would prefer that quantum computers are impossible, but if they are possible then we want the first one". So - every time he'd go to them and say "well...we haven't cracked the nut quite yet" they would turn around and say "great!" which is quite the unusual funding situation. edit: link to talk http://longnow.org/seminars/02016/aug/09/quantum-computer-reality/ http://longnow.org/seminars/02016/aug/09/quantum-computer-re...
- 21 9y ago> well...we haven't cracked the nut quite yet" they would turn around and say "great!" which is quite the unusual funding situation. That's a bit like insurance: Your house hasn't burned down yet. Great, here is another insurance premium for you.
- reikonomusha 9y agoQuantum computers have a lot more utility than factoring integers. How about simulating molecules, or solving classically hard optimization problems?
- moyix 9y agoWhich optimization problems are you thinking of?
- reikonomusha 9y agoGraph coloring, SAT problems, eigenvalue minimization. Almost any binary problem where you can write down a cost function.
- MichaelBurge 9y agoAgain, you're thinking of an NTM not a quantum computer. There are some number theoretic problems, only: https://complexityzoo.uwaterloo.ca/Complexity_Zoo:B#bqp https://complexityzoo.uwaterloo.ca/Complexity_Zoo:B#bqp
- moyix 9y agoAre you thinking of the adiabatic QC approach (i.e., the dwave machine)? I think they still haven't been able to show that it's faster than a classical approach for any problem.
- cfmcdonald 9y agoWhat is an NTM? Google search turns up Nontuberculous mycobacteria and a French rap group.
- MichaelBurge 9y agoNondeterministic Turing Machine
- AnimalMuppet 9y agoThat might actually be a good name for a rap group. Or maybe punk... But: Nondeterministic Turing Machine? What? How can it be nondeterministic, and still be a Turing Machine? Aren't they pretty much deterministic by definition?
- Rusky 9y agoIt's nondeterministic in the same sense as the "nondeterministic polynomial" of P = NP. Basically, a normal turing machine would also be a nondeterminisic one if P = NP.