The programmers will know perfectly well how it was obtained, and they will have programmed the steps that will have obtained it. The fact that answers are obtained from a quantum computer that couldn't be obtained any other way will make people take seriously that the process that obtained them was objectively real.
Nothing more than that is needed to lead to the conclusion that there are parallel universes, because that is specifically how quantum computers work.
Because you can run a computer simulation of reality, that leads to the conclusion there are parallel universes? Uh, no.
A lot of people are still in love with the "many-worlds" hypothesis, possibly partly just because it makes for great sci-fi, but there's very little evidence for it.
If it works this will be the greatest advance in computing since transistors.
I am rather doubtful.* But at the sub-atomic level things are not as we perceive them with our senses. It has to do with transitions - when particles change position or energy there is an age old question 'what happens in between?'
Zeno used that problem two thousand years ago in puzzles such as arguing that a faster runner could never overtake a slower runner.
*And when an old guy doubts something new it isn't always a bad idea to bet the opposite way.
Whether the theory of multiple universes is valid or not--I have no idea--I would bear in mind that David Deutsch is the physicist who is considered the father of the entire field of quantum computing, and much of what they're doing rests on his work. Therefore I wouldn't be too quick to dismiss what he says merely because he's enamoured of the notion. He's laid the theoretical groundwork for what these guys are actually doing.
Reporters watched as the machine solved a Sudoku puzzle and a seating arrangements problem, and, most impressively, searched for molecules similar to the drug Prilosec from a database of molecules.
If it solved a Sudoku puzzle it got that right. I would think it therefore a given that if it had failed the other exercises it would have read, "...but it failed to solve a seating arrangements problem and, most disappointingly, failed to search for molecules similar to the drug Prilosec..."
I cannot imagine they would have even held the press conference if they didn't already have these things working, either.
As for what the the "multiple universe hypothesis" means: quantum theory nowadays posits that there are 11 dimensions, not just four, and that alternate universes are implicit as a result.
Deutsch isn't positing that as part of his theories at all by the way. It's not central to his mission. He's just taken that assumption, made by others since the 1970s, and drawn out a bunch of theoretical implications that should make it possible to design quantum computers.
If they work as he describes, then, that will be just further evidence that multiple universes are real.
If it solved a Sudoku puzzle it got that right. I would think it therefore a given that if it had failed the other exercises it would have read, "...but it failed to solve a seating arrangements problem and, most disappointingly, failed to search for molecules similar to the drug Prilosec..."
It's more complicated than that, though. The Sudoku puzzle has a known, provable answer. Seating arrangements and similar molecules are examples of the sorts of problems (commonly addressed with parallel processing, genetic algorithms, and other sorts of non-traditional programming) where there are no definitively correct answers. All you can really do is compare two answers and see that one is better than another; but you can never definitively say that you have found the "best" answer, unless the problem is limited and a proof by exhaustion is possible.
So what I would like to know for more clarity is which of the following is true?
1. The quantum computer succeeded in finding demonstrably good answers for these problems, possibly better than parallel and genetic techniques, or possibly faster, or both.
2. The quantum computer successfully found the best answers because its design allowed for a practical proof by exhaustion.
3. The quantum computer changes the rules, allowing the discovery of provably correct answers to these problems.
I'm no mathematician, but my math buddies tell me #3 ain't happening: short of proof by exhaustion, these sorts of answers are not amenable to proof, for mathematically proven reasons.
#2 doesn't seem right to me. #1 is what I expect: that the quantum computer discerns very good answers almost automatically, instead of the clever-but-cumbersome techniques of genetic algorithms.
I'm reminded of an article in Scientific American 20 years or so ago. It discussed holographic computers that weren't computers in any way we typically think of the term, but could solve interesting problems almost effortlessly. As far as I know, that effort never went anywhere; but what I was fascinated by was that the technique was entirely different from digital computing.
If you stop and think about it for a moment all he is really saying is that quantum flux of function and outcome, if it is useful for solving problems in this world, operates by an analysis of that uncertainty of outcome which by necessity posits several different solutions, one (although not necessarily only one) of which will work in this world.
But the observed correct one will work no matter how many variables influence the process by which that solution or state is derived.
But because any number of possible functionally variable solutions is viable in a given set of circumstances where different, or even the same variables are in action, then that necessitates the idea that multiple and functionally useful outcomes will occur eventually, even if we never observe or perceive them.
Hence for quantum computing to work it must work by a process which acknowledges the fact that different outcomes, some of which will be viable, even if we never perceive or detect those outcomes, will occur.
Therefore if such effects occur at both the sub-atomic level and in computational calculation then it stands to reason that the overall accumulation of such effects will eventually render other realities in which different effects occur, even if in such places the observers there cannot detect our variations of the same phenomena.
I'm not saying whether or not quantum computational calculation is really occurring or not, (and if you think about it really to the necessary conclusion, then if it is occurring it by necessity ain't, or at least you could never really be sure, but if you can't be sure, then it probably is - that is both the practical paradox and the inherent theoretical irony) merely that if it is then it does stand to reason that variations in computational effect are governed by the same or parallel theoretical structures which govern physical quantum operations, and that such operations may eventually produce a state of existence which renders variant outcomes to any number of state and force interactions, whether those interactions are observed or not.
This of course would mean that quantum effects would have to "accumulate" in an eventually stable enough manner to effect matter on the macro scale, of which I also have no real opinion. I've seen calculations which posit the possibility and those which seem to imply quantum effects degrade at certain scales and become impossible because of ubiquitous, or near ubiquitous forces and states like the influence and interference of gravity, the electromagnetic forces, the weak force, etc.
We'll probably never know some of these things and if quantum phenomena do indeed operate as they seem to operate then it will be impossible to even perceive (given our current limitations) particular outcomes. You would have to be literally not observing to observe the operation and outcome.
But be that as it may, it is interesting to think about.
And if it can be made to function in our world for useful purposes, then as far as I'm concerned it really doesn't matter what exactly is understood about it, and what is not. After all I'd wager 99% of the population watching TV doesn't really know how a TV works, yet they successfully employ one everyday, and although using Newtonian physics and Relativity we can calculate the effects and functions of gravity, maybe even some of the causes, we cannot really say what gravity actually is, or why it exists. Merely that it does, and that that is a good thing in most ordinary circumstances and bad if you fall off the 12th floor without a net.
Science will never be able to tell us everything that is, or why things are as they are, but it is extremely useful in many circumstances, and that's good enough for me.
Theory is fun, and one day may even prove useful, but to me if a thing works, that's good enough.
I'm never gonan be God and understand everything, I just work here. But if it works then that's worth a shot as far as I'm concerned, and Godspeed with his efforts.
It's just another example of God Technology to me, and maybe it will be really useful one day.
If so, good for him.
Martin: You're way, way, way, WAAAAAAAAY ahead of things.
It's doing the equivalent work of a simple 1950s era computer, nothing more.
It's going to have to go a lot further before it's capable of anything that current digital computers cannot do.
As the article makes clear, they need to get these beasties a lot more advanced than they are before they start to do anything you can't do with a very simple primitive microprocessor. That's not the point.
And personally, my interest in it is pretty much that they can compute things simply by manipulating subatomic particles. If nothing else it demonstrates the amazing storage potential in the universe.
By the way, K: It already does work. In fact they've been doing things with quantum computers for a while now. They were at the "single transistor" stage some time ago. Now they've got these puppies working. Still doing primitive stuff of course.
Dean: I was expressing doubt about whether quantum computing will, or has already, lead to useful devices.
Reports of single stage - and I have seen two stages claimed - logic have been coming out for roughly five years.
It appeared that coupling stages and eliminating noise* would be a long battle. Not to mention programming (they do say it is a special, not general, purpose machine).
Yet here a small company announces these obstacles are gone. And they can even search databases and figure out similarities.
That is astonishing to me. I didn't say it was impossible.
* What I call 'noise' corresponds roughly to what Deutsch calls 'errors and decoherence.' And what I term 'stages' would be 'qubits'. Deutsch points out a 'qubit' has no standard definition so D-Wave may not be using the term as he would.
But Dean, that's exactly why I think the linked article raises more questions than it answers. It leaves the impression that the quantum computer is behaving in a general purpose sense by saying it "solved" these problems, without really explaining the nature of the questions and answers. Being that it's in Wired, I honestly expected a more detailed explanation deep in the article. I was surprised there wasn't one. Without that, I think Chris is entirely right to say:
The article doesn't state whether the machine got the answers right, which is a real pity, as it's rather relevant.
Now here's a press release that indicates they claim to be solving -- not just findind good answers, but solving -- NP-complete problems. But then this article promises less, but seems more realistic (emphasis added):
However, quantum computers can be used to calculate approximate solutions to large NP-complete optimization problems more quickly than the best-known methods running on any supercomputer.
And then there's this claim, which may be just marketing hype, but is exciting nonetheless:
The company plans to deliver field-deployable systems in 2008.
The history of physics tells us that interpreting physical theories to yield understanding of the nature of reality, is pointless. Any fundamental physical theory can be "interpreted" as "meaning" opposite things.
Classical mechanics is deterministic (Newtonian form) and teleological (Hamiltonian form). Relativity perfects classical physics (what Einstein thought) and it totally trashes classical physics (what almost everyone else thinks). Everything in quantum mechanics is waves (Schroedinger) and matrices (Heisenberg).
What is described here as "multiple universes" can equally be described as less than one universe. Suppose we measure the spin of an electron as "up," when it was originally in a mixed state ["up"+"down"]/sqrt(2). Deutsch would say that the measurement created two universes, one in which the spin was "up," and one in which the spin was "down." We are in the one with the spin "up."
But you could also say that we started with a universe containing both "up" and "down," and we lost the half with "down." We achieved certainty about the spin of the electron, at the cost of throwing away the half of the universe that held the potential for the electron to be anything but "up."
String theory is its own pile of interesting. Have you read The Trouble with Physics by Lee Smolin? It's a really interesting overview of String theory, quantum gravity, etc. (Smolin is himself mostly in quantum gravity, though he has at times done string theory.) To say that quantum physics posits 11 dimensions is vastly over-simplifying things.
The reading that I've done on quantum tends to suggest that the multiple worlds things isn't very accurate (especially since these worlds have to interact with each other). I'm curious what Deustch means that quantum computers depend on multiple worlds. The electron double-slit experiment, for example, suggests that a multiple-independent-worlds interpretation of quantum is mistaken. Multiple interacting worlds is different, and difficult to describe. It's also unclear that multiple interacting worlds doesn't just mean a single world that's weirder than we expect it to be.
Anyhow, from what I gather simply producing a result at all on a quantum computer is no small feet. It wouldn't surprise me if merely managing that (on so large a scale as 16 qbits) would be press-worthy. Moreover, quantum computers are not exactly what you would call deterministic. Even apart from experimental error, reading error, etc. I'd much rather if they had been explicit about getting correct results.
And I don't see why you see this as having big implications for data storage. Quantum computers aren't exactly small; you're going to have to go through a lot of iterations of twice-the-performance before you can get something with more storage inside of a personal computer.
And that's leaving out of account, entirely, that the ability to manipulate a sub-atomic particle isn't at all the same thing as that sub-atomic particle staying that way over the course of days. Electrons are not exactly what you'd call docile things.
Bob - Wow, I guess I've always thought of it like that, but never quite realized that was a way to express that way of thinking. I'll mull over that some more. Very interesting, thanks.
I'm still not convinced that the whole situation is distinct from the fact that to measure something, you must interact with it, and interacting with it always affects its state. That seems to me to be a rational basis for the uncertainty principle, and can also explain why states "collapse" when they are "observed". Observation = interaction.
Chris: I'm aware of the book and its arguments, but I haven't read it. Just FYI, string theory has been upgraded, the dominant theory in the field is known as M-theory, and it posits a quite specific number of dimensions. 11.
In any case all I was doing was answering your question; Deutsch is merely suggesting that his theories of quantum computing are based entirely on those theories which assume multiple universes. It's not really "the multiple world theory," he's just saying that if these beasties work at all then it validates the fields of speculation upon which they are based.
As for using electrons as storage: one thing at a time. The potential is there.
This may be a form of crawling just shy of walking or it may be the first steps of a new technology, but it demonstrates a form of progress that suggests good things in the future. The article in Wired, assisted by a leading theoretical voice in the field, demonstrates that progress is being made in a new field.
Outcomes are uncertain, as they should be. They look to be positive in advancing quantum computing's entrance onto the technological stage. This is nothing more (or less) than a positive step for quantum computing.
My question is more practical: How can I call dibs on one of the other universes? I've always wanted my own universe.
2.19.2007 11:30am
Commenting on Dean's World is a privilege, not a right. Dean is your host, you are his guest, and you should behave in that fashion. Dean is not your babysitter, nor is he your punching bag. Please remember this. In general, you are free to disagree with anyone on any subject you wish, but abusive behavior will not be tolerated.
Of course we all lose our tempers now and then. Dean freely admits to being imperfect in this regard, which is why regulars to this establishment will generally be cut more slack than people who we don't know very well.
Still: behave like an adult, or go find somewhere else to play. Thanks.
Nothing more than that is needed to lead to the conclusion that there are parallel universes, because that is specifically how quantum computers work.
Because you can run a computer simulation of reality, that leads to the conclusion there are parallel universes? Uh, no.
A lot of people are still in love with the "many-worlds" hypothesis, possibly partly just because it makes for great sci-fi, but there's very little evidence for it.
I am rather doubtful.* But at the sub-atomic level things are not as we perceive them with our senses. It has to do with transitions - when particles change position or energy there is an age old question 'what happens in between?'
Zeno used that problem two thousand years ago in puzzles such as arguing that a faster runner could never overtake a slower runner.
*And when an old guy doubts something new it isn't always a bad idea to bet the opposite way.
I'd be very curious to know what the multiple-universe hypothesis really means, though.
Reporters watched as the machine solved a Sudoku puzzle and a seating arrangements problem, and, most impressively, searched for molecules similar to the drug Prilosec from a database of molecules.
If it solved a Sudoku puzzle it got that right. I would think it therefore a given that if it had failed the other exercises it would have read, "...but it failed to solve a seating arrangements problem and, most disappointingly, failed to search for molecules similar to the drug Prilosec..."
I cannot imagine they would have even held the press conference if they didn't already have these things working, either.
As for what the the "multiple universe hypothesis" means: quantum theory nowadays posits that there are 11 dimensions, not just four, and that alternate universes are implicit as a result.
Deutsch isn't positing that as part of his theories at all by the way. It's not central to his mission. He's just taken that assumption, made by others since the 1970s, and drawn out a bunch of theoretical implications that should make it possible to design quantum computers.
If they work as he describes, then, that will be just further evidence that multiple universes are real.
It's more complicated than that, though. The Sudoku puzzle has a known, provable answer. Seating arrangements and similar molecules are examples of the sorts of problems (commonly addressed with parallel processing, genetic algorithms, and other sorts of non-traditional programming) where there are no definitively correct answers. All you can really do is compare two answers and see that one is better than another; but you can never definitively say that you have found the "best" answer, unless the problem is limited and a proof by exhaustion is possible.
So what I would like to know for more clarity is which of the following is true?
1. The quantum computer succeeded in finding demonstrably good answers for these problems, possibly better than parallel and genetic techniques, or possibly faster, or both.
2. The quantum computer successfully found the best answers because its design allowed for a practical proof by exhaustion.
3. The quantum computer changes the rules, allowing the discovery of provably correct answers to these problems.
I'm no mathematician, but my math buddies tell me #3 ain't happening: short of proof by exhaustion, these sorts of answers are not amenable to proof, for mathematically proven reasons.
#2 doesn't seem right to me. #1 is what I expect: that the quantum computer discerns very good answers almost automatically, instead of the clever-but-cumbersome techniques of genetic algorithms.
I'm reminded of an article in Scientific American 20 years or so ago. It discussed holographic computers that weren't computers in any way we typically think of the term, but could solve interesting problems almost effortlessly. As far as I know, that effort never went anywhere; but what I was fascinated by was that the technique was entirely different from digital computing.
But the observed correct one will work no matter how many variables influence the process by which that solution or state is derived.
But because any number of possible functionally variable solutions is viable in a given set of circumstances where different, or even the same variables are in action, then that necessitates the idea that multiple and functionally useful outcomes will occur eventually, even if we never observe or perceive them.
Hence for quantum computing to work it must work by a process which acknowledges the fact that different outcomes, some of which will be viable, even if we never perceive or detect those outcomes, will occur.
Therefore if such effects occur at both the sub-atomic level and in computational calculation then it stands to reason that the overall accumulation of such effects will eventually render other realities in which different effects occur, even if in such places the observers there cannot detect our variations of the same phenomena.
I'm not saying whether or not quantum computational calculation is really occurring or not, (and if you think about it really to the necessary conclusion, then if it is occurring it by necessity ain't, or at least you could never really be sure, but if you can't be sure, then it probably is - that is both the practical paradox and the inherent theoretical irony) merely that if it is then it does stand to reason that variations in computational effect are governed by the same or parallel theoretical structures which govern physical quantum operations, and that such operations may eventually produce a state of existence which renders variant outcomes to any number of state and force interactions, whether those interactions are observed or not.
This of course would mean that quantum effects would have to "accumulate" in an eventually stable enough manner to effect matter on the macro scale, of which I also have no real opinion. I've seen calculations which posit the possibility and those which seem to imply quantum effects degrade at certain scales and become impossible because of ubiquitous, or near ubiquitous forces and states like the influence and interference of gravity, the electromagnetic forces, the weak force, etc.
We'll probably never know some of these things and if quantum phenomena do indeed operate as they seem to operate then it will be impossible to even perceive (given our current limitations) particular outcomes. You would have to be literally not observing to observe the operation and outcome.
But be that as it may, it is interesting to think about.
And if it can be made to function in our world for useful purposes, then as far as I'm concerned it really doesn't matter what exactly is understood about it, and what is not. After all I'd wager 99% of the population watching TV doesn't really know how a TV works, yet they successfully employ one everyday, and although using Newtonian physics and Relativity we can calculate the effects and functions of gravity, maybe even some of the causes, we cannot really say what gravity actually is, or why it exists. Merely that it does, and that that is a good thing in most ordinary circumstances and bad if you fall off the 12th floor without a net.
Science will never be able to tell us everything that is, or why things are as they are, but it is extremely useful in many circumstances, and that's good enough for me.
Theory is fun, and one day may even prove useful, but to me if a thing works, that's good enough.
I'm never gonan be God and understand everything, I just work here. But if it works then that's worth a shot as far as I'm concerned, and Godspeed with his efforts.
It's just another example of God Technology to me, and maybe it will be really useful one day.
If so, good for him.
It's doing the equivalent work of a simple 1950s era computer, nothing more.
It's going to have to go a lot further before it's capable of anything that current digital computers cannot do.
As the article makes clear, they need to get these beasties a lot more advanced than they are before they start to do anything you can't do with a very simple primitive microprocessor. That's not the point.
Reports of single stage - and I have seen two stages claimed - logic have been coming out for roughly five years.
It appeared that coupling stages and eliminating noise* would be a long battle. Not to mention programming (they do say it is a special, not general, purpose machine).
Yet here a small company announces these obstacles are gone. And they can even search databases and figure out similarities.
That is astonishing to me. I didn't say it was impossible.
* What I call 'noise' corresponds roughly to what Deutsch calls 'errors and decoherence.' And what I term 'stages' would be 'qubits'. Deutsch points out a 'qubit' has no standard definition so D-Wave may not be using the term as he would.
Now here's a press release that indicates they claim to be solving -- not just findind good answers, but solving -- NP-complete problems. But then this article promises less, but seems more realistic (emphasis added):
And then there's this claim, which may be just marketing hype, but is exciting nonetheless:
Classical mechanics is deterministic (Newtonian form) and teleological (Hamiltonian form). Relativity perfects classical physics (what Einstein thought) and it totally trashes classical physics (what almost everyone else thinks). Everything in quantum mechanics is waves (Schroedinger) and matrices (Heisenberg).
What is described here as "multiple universes" can equally be described as less than one universe. Suppose we measure the spin of an electron as "up," when it was originally in a mixed state ["up"+"down"]/sqrt(2). Deutsch would say that the measurement created two universes, one in which the spin was "up," and one in which the spin was "down." We are in the one with the spin "up."
But you could also say that we started with a universe containing both "up" and "down," and we lost the half with "down." We achieved certainty about the spin of the electron, at the cost of throwing away the half of the universe that held the potential for the electron to be anything but "up."
String theory is its own pile of interesting. Have you read The Trouble with Physics by Lee Smolin? It's a really interesting overview of String theory, quantum gravity, etc. (Smolin is himself mostly in quantum gravity, though he has at times done string theory.) To say that quantum physics posits 11 dimensions is vastly over-simplifying things.
The reading that I've done on quantum tends to suggest that the multiple worlds things isn't very accurate (especially since these worlds have to interact with each other). I'm curious what Deustch means that quantum computers depend on multiple worlds. The electron double-slit experiment, for example, suggests that a multiple-independent-worlds interpretation of quantum is mistaken. Multiple interacting worlds is different, and difficult to describe. It's also unclear that multiple interacting worlds doesn't just mean a single world that's weirder than we expect it to be.
Anyhow, from what I gather simply producing a result at all on a quantum computer is no small feet. It wouldn't surprise me if merely managing that (on so large a scale as 16 qbits) would be press-worthy. Moreover, quantum computers are not exactly what you would call deterministic. Even apart from experimental error, reading error, etc. I'd much rather if they had been explicit about getting correct results.
And I don't see why you see this as having big implications for data storage. Quantum computers aren't exactly small; you're going to have to go through a lot of iterations of twice-the-performance before you can get something with more storage inside of a personal computer.
And that's leaving out of account, entirely, that the ability to manipulate a sub-atomic particle isn't at all the same thing as that sub-atomic particle staying that way over the course of days. Electrons are not exactly what you'd call docile things.
I'm still not convinced that the whole situation is distinct from the fact that to measure something, you must interact with it, and interacting with it always affects its state. That seems to me to be a rational basis for the uncertainty principle, and can also explain why states "collapse" when they are "observed". Observation = interaction.
In any case all I was doing was answering your question; Deutsch is merely suggesting that his theories of quantum computing are based entirely on those theories which assume multiple universes. It's not really "the multiple world theory," he's just saying that if these beasties work at all then it validates the fields of speculation upon which they are based.
As for using electrons as storage: one thing at a time. The potential is there.
Outcomes are uncertain, as they should be. They look to be positive in advancing quantum computing's entrance onto the technological stage. This is nothing more (or less) than a positive step for quantum computing.
My question is more practical: How can I call dibs on one of the other universes? I've always wanted my own universe.
Of course we all lose our tempers now and then. Dean freely admits to being imperfect in this regard, which is why regulars to this establishment will generally be cut more slack than people who we don't know very well.
Still: behave like an adult, or go find somewhere else to play. Thanks.