Write that down and pin it to the corkboard in your office.
It will turn out to be one of the greatest understatements you will have ever made.
Not that quantum computing will not produce many absolutely amazing positive results, it will, but the view of them will be obscured by all the smoking craters QC causes.
Francisco
May 25, 2021
Comment to Quantum computing as a threat to Bitcoin
[I’m currently of the opinion the positive results will be on par with the smoking craters. But I’m not knowledgable enough on the subject to claim any expertise.
We live in interesting times.—Joe]
I’m not convinced. For one thing, it’s not clear QC will ever become practical. Possibly it can be scaled up far enough to do real work, but possibly that can’t be done. If it can be done, it might not be economical. Programming a QC is, at least at the moment, a mindbogglingly difficult job, far harder than conventional computer programming was in 1950.
Nor am I convinced there are any significant “smoking craters” in the future. Some crypto algorithms will no longer be useable, but others will be fine and that work is well underway. That’s just one more step in the long evolution of crypto from pencil and paper solutions to supercomputers and dedicated machines. And apart from crypto, what craters are on the horizon? The marketing hype of QC (if you get down to details) is that it will do optimization tasks faster than conventional computers, which might conceivably be true. Actually, the real question is whether the fast exact solutions it promises to offer are sufficiently better than the fast heuristic nearly-optimal solutions of conventional algorithms to justify the extreme cost. That’s unclear, to say the least. Never mind the fact that a prominent proponent of the notion (D-Wave) may not be a QC at all; a bunch of their white papers are focused on the goal of proving that what it does is quantum computation, and the data are not all that strong.
But the question is, will QC be able to run Crysis at 60fps with no loss in display? 😀
Almost certainly not.
But if it can be used practically to solve the discrete logarithm problem, say goodbye to all the current encryption methods, and the kind of people that want to be all up in your business will fund that kind of development heavily.
Only one-time pads will be safe.
Not accurate, Tirno. Yes, I think it is correct to say QC can solve discrete logs fast, but it does not in any way follow that this leaves only one time pad. The current mainstream encryption systems (like TLS, which is part of HTTPS) use it, but there has been a significant body of work on “post-quantum cryptography”. I haven’t closely tracked it but I think that there are several plausible solutions waiting to be picked when the need becomes real.
I think Bitcoin is the last worry of QC. The real worry is that it will propel AI ahead by leaps and bounds. That in turn will put a large percent of the population out of work. Millions of people with idle hands and no good way to put their competitive instinct to good works is a recipe for utter disaster.
I don’t see any connection between AI and QC, other than that they are both serious hype topics. Also, AI seems to involve, in many cases, very large data sets. QC dreams for the end of the decade are about a million qubits, which translates to maybe 10k error free qubits. That’s less than the memory capacity of ENIAC, or about the memory quantity you might find in a small research lab computer circa 1950. If 100 Gq is achievable some day, that might be a different matter.
Whether propelling AI ahead by leaps and bounds has any impact on unemployment isn’t clear either, no more than the steam engine, the assembly line, or the computer did.
The smoking holes will be due to increasing automation for lower skilled jobs or what some call the sh** jobs, but they will still be limited in size and impact. Humans will always find something to do even it is counter-productive.
QC if it becomes reality is just an enabler and may allow us to solve more complex problems. However, do keep in mind, that more efficiency or capability will not translate into increase usefulness. Randomness hidden in systems will prove to be a hard limit. For example, we will find that even with all our progress that accurate modeling of complex systems such as life or environment are limited – and will always be limited – because of randomness. Perhaps, we’ll be able to improve weather forecasting beyond two weeks to four weeks but the cost will grow exponentially and we’ll always see limits on knowability. QC cannot change that and neither can AI. That is the reason that we have Gods – they help us explain the unknown and unknowable.