3 ms·
The way this starts seems to tell a story that I feel is quite disconnected from reality: > As quantum computers have become available to the general public, t
by hannob 9y ago
The way this starts seems to tell a story that I feel is quite disconnected from reality:
> As quantum computers have become available to the general public, the need has arisen to train a cohort of quantum programmers
It seems to peddle the idea that in a few years we'll replace all normal computers with quantum computers. I don't think this is even remotely plausible.
It's still an open question whether quantum computers that do anything useful at all are feasible. But even if they are - we can safely assume that for a long time they will be expensive devices, running in complex physical experiments, that will be reserved for very special needs.
From all we know QCs aren't magically faster computers that are suitable to do everything better. There are just a few very special algorithms where QCs have an advantage. That's certainly interesting, but even if it would be possible - and that's a big if - it's unlikely anyone will add them to our smartphones anytime soon, simply because there's no need for that.
If quantum computers arrive we'll need a handful of specialized programmers that are familiar with their algorithms. But not masses of them.
- zaarn 9y agoQuantum Computers will have some applications in Cryptography, I hope. There are already some Quantum Key Exchange algorithms which can detect wiretapping by a third party reliably without requiring a PKI. If any third party is listening, the connection will simply fail as a fundamental property of nature. (Though MitM is still possible, but you have to go full, a simple wiretap is not sufficient anymore) Another possibility are quantum algorithms that allow one to perform cryptography without having to trust the underlying device, it could even be malicious, and everything would either work without leaking or just fail without leaking. (Though IIRC research in that area is still ongoing)
- hannob 9y agoQKEs have nothing to do with quantum algorithms running on quantum computers. Also they don't solve any real problem, but that's another topic. It's an extremely impractical and expensive way to implement a key exchange over short distances which, as you said, is vulnerable to mitm attacks. I don't know why anyone would want to have that.
- blauditore 9y agoI think the relevance of potential quantum computer programming can be compared to that of GPU programming: It's useful to solve particular problems with much more power than with classic approaches, but for most applications there's no gain. So most software will probably continue to run on classic computers as today.
- hannob 9y agoI think use of QCs will be orders of magnitude lower than use of GPUs.
- zitterbewegung 9y agoI remember when the only useful thing you could do with a GPU would be to play videogames. Now you can do 1. Deep Learning 2. Cryptocurrency Mining 3. CAD / CAM 4. Acceleration of certain computations like fluid dynamics
- giomasce 9y agoI suppose people thought the same of classical computers when the ordinary way to carry out computations was to deliver them to arrays of people sitting at desks. Now, history is not forced to repeat itself, but it has happened frequently enough so that it might be premature to exclude that QC will at some point be generally available and used. I agree that having a lot of programmers is not an immediate need, though.
- Uehreka 9y agoSure, but then bear in mind that with the onset of deep learning (which seems poised to be the foundation for a lot of emerging technology for at least the next decade) GPU programming has suddenly become a lot more useful. What if, just as deep learning brought life to GPUs decades after they were invented, some other algorithm or paradigm that we’re not paying attention to now becomes huge once QCs are available to test on?