Quantum Computing: A view from the enemy camp

SIGFPE writes "There seems to be an unthinking acceptance by many people that quantum computers are now on their unstoppable way up and before too long we'll be cracking RSA and simulating protein folding on complex quantum computers. However there is another point of view that considers quantum computers to be as difficult to make as perpetual motion machines - and for much the same reason: entropy. As an antidote to all the successes that have been reported on /. here is a just published and highly readable preprint by a sceptical mathematical physicist."

Comments

This is all well and good.

[CyberBlood]

While everything stated in this paper is all well and good and valid, he's missing the fundamental point of innovation.

You make no progress with pessimism.

CyberBlood

Re:This is all well and good.

[Jerf]

False idealism doesn't get far either. "If we just keep working on the epicycles, maybe we can finally crack the orbit problem!"

Re:looking sideways

[famebait]

> any lifeform that does not take full advantage of its enviroment will not survive

Whoah. So any phenomenon seen or imagined in science that is not manifested in humans is bullshit? And furthermore, all lifeforms can in fact swim, fly, breathe and photosynthesize and do all other things that are possible for a life form to do; either that or the very theoretical possibility of a doing these things is bullshit?

Re:looking sideways

[Pentagram]

And because no organisms use lasers as a defensive weapon, they cannot exist?

evolution dictates that any lifeform that does not take full advantage of its enviroment will not survive to procreate

Evolution dictates nothing of the sort. All that happens is the fittest of any set of organisms is more likely to survive. There is no guarantee of optimality (otherwise I would currently be making a fortune with my perfect genetic algorithm-evolved stock market model).

Achieveable accuracy

[tony_gardner]

An interesting article, but it seems to rely on several strange arguments. I wouldn't agree that the accuracies required in manufacture and operation are not achieveable. It would seem to me that a fabrication accuracy of 1 part in 10^5 is scarcely unachieveable, especially since we have everyday devices, for example LCD monitors in which less than 1 pixel in 1 million is dead, and those are consumer products.

I don't necessarily buy his argument about the inherent fault-intolerence of quantum computing either, since it relies on the idea that a simple computer with no fault tolerence built in is fault intolerant! Scarcely a surprise. He didn't make a convincing argument that building in fault tolerance is impossible, only that it isn't being done in the simple designs he notes. Maybe I've misunderstood, but it would seem that this is an avenue for more research, rather than less.

It seems to me that in the end the whole of his argument relies on the engineering argument that we don't at present have any way to measure large numbers of single spin states (or indeed any single spin state). This would seem again to be an argument for more research, rather than less, since engineering serendipity is not a predictable mathematical process.

I understand his frustration that quantum computing is taking a lot of research funding from other areas, but I'd be a little more cautious than he is about saying that it definitely _can't_ work.

Re:Achieveable accuracy

[SIGFPE]

I'm inclined to compare building a quantum computer to building a perpetual motion machine. There is nothing in the fundamental laws of physics that says it's impossible. They're obviously not impossible, an individual atom is in effect in perpetual motion. But start building multi-atom constructions and it's going to get harder and harder. Although we don't have a macroscopic perpetual motion machine I'd be disinclined to actually put a limit on how big a perpetual motion machine can get. I just think it gets exponentially harder the larger you want it to be. The same goes for quantum computers - and for the very same reason: entropy.

Re:Achieveable accuracy

[superflex]

you're absolutely right about how atoms are in fact perpetual motion "machines". the whole idea of a "PMM" is a misnomer, because the idea behind it is not one of perpetual motion, but of perpetual work/energy. that's the part that isn't kosher, because it violates the first law of thermodynamics "energy cannot be created or destroyed".

Re:Achieveable accuracy

[norton_I]

I used to be a skeptic about the possibility of quantum computing, but I have become a believer. Here is why:

1) It once was thought that gate precisions would need to be accurate to 1 in 10^12 or so, depending on the length of the computation. The problem was, unlike classical computers, quantum computers are not allowed to "latch" to a 1 or 0. However, there has been a substantial body of work that has shown how to implement error correction codes on quantum computers, without destroying the computation.

2) Decoherence, or the problem with entropic contamination by the environment, is a serious one, and the one that was always thought to be the nail in the coffin. However, now that people have actually started trying to make systems resilient to decoherence, they have measured systems that decohere on time scales from ms to hours -- giving considerable time to do computation.

3) One particular QC proposal, germanium quantum dots on silicon, has made incredible progress recently. A couple of years ago, someone wrote down the "six impossible tasks to make Si quantum computers work", three of which have already been accomplished.

We have already accomplished so much in QC that people thought would be impossible, it hardly seems logical to give up now. I can't say for certain, obviously, that QC will ever be a reality, but I would give it better than 50-50 odds at the moment.

Re:Achieveable accuracy

[WolfWithoutAClause]

Further, people have actually demoed quantum mechanical electronics using MRI so the theory seems cosher.

Ok, a 4 bit circuit when you're aiming for ~10,000 bits is a bit restrictive, but atleast it shows proof of concept.

Re:parallel processing

[jvv62]

"If it is possible to utilize parallel processing then life would have had to evolve the ability to take advantage of that." - Perdo

Ummm....

While this sentence is taken out of context, it is still completely wrong. We use parallel processing all the time. The parallel computation part of quantum computing seems to work in both theory and in the lab as well.

Assume that there exists an efficient (classical) algorithm for calculating f(x) for a given x, however if n is large, say 1000, calculation for all x would take quite a lot of time. The quantum computer, in a certain sense, performs all these calculations much faster, during a time which is not exponential, but polynomial in n. (I note in passing, that the difficulty of building a quantum computer increases exponentially with n, which is a kind of Nature's revenge). -Dyakonov

The argument against QC is really much simpler than Perdo's convoluted logic requires. Dyakonov says that the problem isn't that a quantum computer couldn't do the calculations fast. The problem is that it would take nearly forever to build the computer in the first place.

Re:looking sideways

[SIGFPE]

For some speculative ideas along these lines you could do worse than read Teranesia by Greg Egan. Actually Egan thinks the idea is purel hokum but at least one biologist has taken the ideas in Teranesia seriously.

Re:looking sideways

[SIGFPE]

Why would it change the debate on whether humans have souls? My memory is poor but I don't recall any religion that makes any connection between quantum computing and souls.

Watch out for Clarke's First Law

[alispguru]

Found here among other places:

When a distinguished but elderly scientist states that something is possible he is almost certainly right. When he states that something is impossible, he is very probably wrong.

I don't know if the author in question qualifies here, but it is something to consider.

Re:Watch out for Clarke's First Law

[SIGFPE]

It's a good quotation with lots of relevance to many things in physics. However the word 'elderly' is crucial to the meaning of his quotation. We all know that we have to wait for the old generation of physicists to die off before a scientific revolution can be completed. The quantum computing critics aren't the sceptical old however - they're the enthusiastic young who think quantum computing is a cool subject but makes some fundamentally flawed assumptions.

Re:looking sideways

[SIGFPE]

I still don't get the connection between souls and quantum mechanics.

I don't recall any religious tradition that makes unpredictability a property of souls. In fact man religious traditions claim an omniscient God making souls entirely predictable. And even if they did claim an association betwen uncertainty and souls I see no reason for that to be evidence for the existence of souls, because, as I point out, people were unable to predict human behaviour before QM was invented.

Why would unpredictability leave room for a 'higher power'? If it turned out that these so-called random events were controlled by such a power they would no longer be random would they? They'd be amenable to study like anything else. If we expect something to happen 50% of the time but due to a higher power it happened 51% of the time instead then that wouldn't be QM 'leaving room', it'd be QM being wrong. QM in fact leaves no room for such 'higher powers' because the 'random' quantum events are just as subject to the laws of statistics as any other random event.

Anyway, why might there not be laws describing these 'higher powers'?

Re:looking sideways

[SIGFPE]

That your own code doesn't find optimal solutions to a problem has little to contribute to the issue of how optimal evolution is.

quantum computer power requirements

[nobody/incognito]

i know there is some concern about the power requirements of quantum computers, but my physicist friends assure me that cold fusion reactors will provide all the power they need.

nobody

To all my dead hommies

[QuantumG]

As sung by the infamous Mc Hawking.

Entropy, how can I explain it? I'll take it frame by frame it,
to have you all jumping, shouting saying it.
Let's just say that it's a measure of disorder,
in a system that is closed, like with a border.
It's sorta, like a, well a measurement of randomness,
proposed in 1850 by a German, but wait I digress.
"What the fuck is entropy?", I here the people still exclaiming,
it seems I gotta start the explaining.

You ever drop an egg and on the floor you see it break?
You go and get a mop so you can clean up your mistake.
But did you ever stop to ponder why we know it's true,
if you drop a broken egg you will not get an egg that's new.

That's entropy or E-N-T-R-O to the P to the Y,
the reason why the sun will one day all burn out and die.
Order from disorder is a scientific rarity,
allow me to explain it with a little bit more clarity.
Did I say rarity? I meant impossibility,
at least in a closed system there will always be more entropy.
That's entropy and I hope that you're all down with it,
if you are here's your membership.

You down with entropy?
Yeah, you know me!
You down with entropy?
Yeah, you know me!
You down with entropy?
Yeah, you know me!
Who's down with entropy?
Every last homey!


Defining entropy as disorder's not complete,
'cause disorder as a definition doesn't cover heat.
So my first definition I would now like to withdraw,
and offer one that fits thermodynamics second law.
First we need to understand that entropy is energy,
energy that can't be used to state it more specifically.
In a closed system entropy always goes up,
that's the second law, now you know what's up.

You can't win, you can't break even, you can't leave the game,
'cause entropy will take it all 'though it seems a shame.
The second law, as we now know, is quite clear to state,
that entropy must increase and not dissipate.

Creationists always try to use the second law,
to disprove evolution, but their theory has a flaw.
The second law is quite precise about where it applies,
only in a closed system must the entropy count rise.
The earth's not a closed system' it's powered by the sun,
so fuck the damn creationists, Doomsday get my gun!
That, in a nutshell, is what entropy's about,
you're now down with a discount.


You down with entropy?
Yeah, you know me!
You down with entropy?
Yeah, you know me!
You down with entropy?
Yeah, you know me!
Who's down with entropy?
Every last homey!

Re:looking sideways

[SIGFPE]

God is omniscient => God can predict the behaviour of a person => the behavior of a person is predictable