4 ms·
What should a dev do to stay current, given this is happening? Say you have an interest in algorithms in general, what do you need to read / what can you read t
by lordnacho 11y ago
What should a dev do to stay current, given this is happening? Say you have an interest in algorithms in general, what do you need to read / what can you read that connects to existing knowledge?
- semi-extrinsic 11y agoTL;DR: if you're not retired before QC becomes useful (which is likely), the "API" will have changed so much from now till then that you shouldn't worry about it. Think of this as an algorithm on specific hardware for solving a specific problem. You can encode other problems into the specific problem, at a cost. Currently this algorithm on specific hardware is not as fast as state-of-the art algorithms on ordinary hardware (namely Selby's). But the people making this special hardware claim (no proof) that when the special hardware (and the problem size) is scaled up, the special hardware will be much faster than ordinary hardware. So far no-one has disproved this assertion, but no-one has proven or demonstrated it either. Meanwhile, many other scientists and some companies, who refer to this type of special hardware as "dirty QC", are working on another class of special hardware they call "clean QC". This is mathematically proven to be much faster than ordinary hardware when it is scaled up, but it's very very very difficult to scale it up. But they're working on it and showing progress. Right now, no-one knows if the "dirty" or the "clean" special hardware will be the first to be actually useful in the real world. It's likely the first useful real world application is more than ten years away. We don't know exactly what the application will be, or how the special hardware then will be programmed. It's also quite likely that most the people commenting here will be retired before access to the special hardware will be given to anyone but scientists.
- gaze 11y agoWhat do you mean dirty vs clean QC? Do you mean adiabatic vs gate based?
- jessriedel 11y agoYea, that must be what he's referring to. I know a lot of folks who work on this stuff and they never use "clean"/"dirty" terminology (although it's reasonably apt).
- semi-extrinsic 11y agoYep. Picked up the clean/dirty terminology from Scott Aaronson: http://www.scottaaronson.com/blog/?p=2555#comment-963747 http://www.scottaaronson.com/blog/?p=2555#comment-963747
- jessriedel 11y agoFYI, Scott was using the quotes to try and indicate that he was making those terms up on the spot (so it's not accurate to say that "many other scientists and some companies" use those terms). But I can see how it might not be clear from reading his comment.
- semi-extrinsic 11y agoThanks for pointing this out. (As you might have inferred, my original post tried to keep the language simple.)
- vinceguidry 11y agoIf that winds up being a problem, then hold on to your butts because the Singularity is upon us.
- swalsh 11y agoI would suppose that not much will change. If you're already familiar with working with the cloud (like AWS) this is just going to be some new node, with some new API that you use when you want to run some optimization algorithm. In fact, I wouldn't be surprised if AWS/Azure buy a bunch of these, and just offer them as one more service.
- typon 11y agoIf you want to access faster computing regimes that already exist, learn GPU programming or FPGA programming. Both provide extremely large proven speedups over CPU programming and you can start learning them now. Worrying about Quantum Computing is, imo, rather premature.
- gaze 11y agoStart learning it once there's a standard programming language for quantum computers. It's so far off. I say this as an experimentalist in the field.
- hannob 11y agoI think you have a wrong idea here about quantum computers. It is not likely that they will end up on your desk any time soon. What is likely going to happen at first is that for very specialized problems there will be large scale quantum computers. Think of them as complex physical experiments, occupying whole rooms with delicate equipment. That will be the reality of quantum computing for quite some time. If they ever end up on a desk is hard to tell, but that's really science fiction, it's like asking whether we'll visit foreign planets in the future. That means unless you are working in a very specialized environment quantum computers are irrelevant for you as a developer except that you'll likely use different crypto algorithms in the future. But you will use different crypto that still runs on classical computers.
- personlurking 11y agoPardon my ignorance on the matter but, in regards to your last sentence, you mean it'll be possible for classical computer crypto to withstand QC brute force?
- hannob 11y agoQC don't "brute-force", that's an inaccurate description of how they work. But to answer your question, yes, it is possible to have classical computers with crypto that withstand QC attacks. There's a whole research field on postquantum crypto, e.g. have a look at http://pqcrypto.eu.org/ http://pqcrypto.eu.org/
- mweser 11y agoAs long as you don't write low level algorithms, you will most likely not be affected as an engeneer. Imagine you call list.sort() now and in the future.