19 ms·

The Case Against Quantum Computing

138 points by niccl 8y ago

zeroxfe 8y ago
Had to read to the bottom to get to the (really weak) "case against Quantum Computing": > I believe that, appearances to the contrary, the quantum computing fervor is nearing its end. That’s because a few decades is the maximum lifetime of any big bubble in technology or science. After a certain period, too many unfulfilled promises have been made, and anyone who has been following the topic starts to get annoyed by further announcements of impending breakthroughs. What’s more, by that time all the tenured faculty positions in the field are already occupied. The proponents have grown older and less zealous, while the younger generation seeks something completely new and more likely to succeed. I was really hoping for a real technical argument, not speculation about bubbles or politics in academia. (Especially given that most of the real research is done in industry.)
- resource0x 8y ago
  > I was really hoping for a real technical argument Please read the body of the article, not only the bottom. There's plenty of technical arguments in there.
- setquk 8y ago
  Sounds like Fusion. That’s still going on after 60 odd years.
  - powera 8y ago
    With one major difference. We know that fusion in some form is possible (for example, in the sun, or in a hydrogen-bomb). The difficulty is doing it at a reasonable scale. It's less clear that quantum computing is possible.
    - sp332 8y ago
      I mean we already have quantum computers too. https://quantumexperience.ng.bluemix.net/qx/experience https://quantumexperience.ng.bluemix.net/qx/experience The trouble here is also scaling, but in the upward direction.
      - pavelrub 8y ago
        The "quantum computers" we have are classical computers with very weird hardware. The thing everybody is trying to achieve is quantum supremacy - this hasn't been achieved or demonstrated yet, and might never will be.
        TallGuyShort 8y ago
        For at least a very, very long time, I see Quantum Computers as simply being "accelerators" that do very niche tasks. Like GPUs or FPGAs today. I'm sure one day we'll have a whole other paradigm, but I doubt Quantum Computing as we envision it today will be the thing that replaces x86-compatible silicon.
        sp332 8y ago
        They are not classical computers. We've had actual quantum logic gates since 1995. https://en.wikipedia.org/wiki/Timeline_of_quantum_computing https://en.wikipedia.org/wiki/Timeline_of_quantum_computing And as linked above, IBM is renting out time on real quantum computers up to 17 qubits now. Intel, Google, and IBM had all announced plans for 49 or 50 qubit computers, which would have achieved quantum supremacy at the time, but IBM improved their simulator to 56 qubits last year. https://en.wikipedia.org/wiki/Quantum_supremacy https://en.wikipedia.org/wiki/Quantum_supremacy So we're only one or two hardware generations away now.
        pavelrub 8y ago
        They are classical computers in the sense they are are no stronger than classical computers - they provide a different physical implementation, but computationally can't achieve anything that a classical computer can't. Everything those "quantum computers" do can be simulated with a classical computer. The entire motivation behind the development of QC is supremacy, and the existence of those "quantum computers" alone doesn't provide any proof that supremacy is possible - so it is definitely not just a question of scale. Google already has a 72 qubit processor (Bristlecone), but whether they or anybody will be able to demonstrate supremacy in the current, next, or some future generation is entirely conjectural. The main limit is error correction, and there are theoretical arguments (Gil Kalai) for why achieving supremacy might be impossible.
    - core-questions 8y ago
      You can literally sign up for D-Wave Leap and run problems on a quantum computer immediately. It's not gate-model, but it's a real quantum annealer that functions and can compute useful things.
      - krastanov 8y ago
        Quantum annealers are not quantum computers. They are not more powerful than classical computers. You either need the "gate model" or the equivalent "adiabatic quantum computing model" to achieve something infeasible on classical computers. It is a common misconception to think that adiabatic quantum computing (which is equivalent to the gate model) and quantum annealing is the same thing. It is not, and quantum annealing is mostly a buzzword that D-Wave use to sell their devices.
  - simonh 8y ago
    If something sounds too good to be true, it probably is.
    - tarboreus 8y ago
      Sulfa. White swans do come along, they're just incredibly rare.
      - WillPostForFood 8y ago
        Are they rare, or do they just happen over longer timeframes than we want? Just about any mainstream technology today would appear to be too good to be true to someone ~100 years ago. Air travel, television, computers, phones.
        tarboreus 8y ago
        The telephone was invented in 1876, practically old hat a hundred years ago. Air travel was 1903, was around longer than Facebook has been around in 2018. I guess it comes down to your definition of rare, but it's certainly rare enough.
      - effie 8y ago
        Black swans are rare.
  - Animats 8y ago
    We know that fusion is possible. We have H-bombs. We have various things that do small scale fusion. Getting out more power than is put in remains way out of reach.
    - hazeii 8y ago
      I take the point, while noting H-bombs put out more power than is put in. Perhaps "getting more useful power out..."?
- TheOtherHobbes 8y ago
  The technical argument is absolutely clear: quantum computing cannot work because it relies on manipulating and measuring an absolutely astronomical number of continuous variables with near-infinite precision. The argument may or may not be correct, but it deserves something more thoughtful than a dismissive response that doesn't even recognise the basic point the author is making.
  - krastanov 8y ago
    This is one of the standard complaints (the gist of it being that quantum computing is some form of analog computing, i.e. requiring near-infinite precision). For researchers in the field it becomes rather frustrating to have to repeat the same response without being heard, so I can understand the annoyance expressed in the parent comment. For what is worth, here is a good explanation of how this complaint misrepresents the whole premise of quantum computation (in particular, point 6): https://www.scottaaronson.com/democritus/lec14.html https://www.scottaaronson.com/democritus/lec14.html
    - core-questions 8y ago
      > the gist of it being that quantum computing is some form of analog computing It absolutely is - at least, in the only practical, real-today, working instantiation of it which is in the form of quantum annealing.
      - krastanov 8y ago
        This is an extremely misleading statement. Quantum annealing is most certainly not what quantum computing is about (it is not particularly "quantum" either). At most, you can argue that it is an important first step (which is also doubtful). Most of the interesting hardware currently being developed has nothing to do with quantum annealing and most researchers are fairly annoyed at D-wave for originally pushing this nonsense (D-wave is the company that started selling quantum annealing machines, which are little more than a classical analog computer). See the ion traps developed in the UK/Maryland or the transmon qubits at Google/IBM/Yale for some well known prototypes of quantum computing hardware (which are admittedly quite far from being practical or usable).
- scottlocklin 8y ago
  There's certainly a bubble here; trivially obvious, but there are also very good technical arguments in that article. FWIIW Dyakonov has been saying the same thing since the bubble got started in 2000 or so.
nixpulvis 8y ago
I still need to gain the intuition for why a quantum computer can operate on some kinds of things "faster". I've read some of the math, but that did little to satisfy me (I need to study it more clearly). But all of this seems in a tragic state at the moment. Allowing rampant misinformation and hype as to what these machines are actually capable of.
- msclrhd 8y ago
  It's not that they can execute a computation faster in the traditional sense. It is that you can exploit the quantum behaviour in some types of computation (e.g. factoring primes) to make the answer more probable, and thus more likely to be "observed", when you read the state of the quantum computer and thus make it easy to locate the correct answer compared to traditional algorithms. PBS Infinite Series has several videos on quantum computing and Shor's algorithm: 1. https://www.youtube.com/watch?v=IrbJYsep45E https://www.youtube.com/watch?v=IrbJYsep45E "The Mathematics of Quantum Computers" 2. https://www.youtube.com/watch?v=12Q3Mrh03Gk https://www.youtube.com/watch?v=12Q3Mrh03Gk "How to Break Cryptography" 3. https://www.youtube.com/watch?v=wUwZZaI5u0c https://www.youtube.com/watch?v=wUwZZaI5u0c "Hacking at Quantum Speed with Shor's Algorthm"
- krastanov 8y ago
  If you are specifically interested in building up the mathematical intuition, check out "Quantum Computing since Democritus" by Aaronson - it excels at presenting the intuition. For less comp-sci and more physics look at the standard references: Nielsen and Chuang's book and Preskill's lecture notes.
  - nixpulvis 8y ago
    Thanks for the links. I'll start here: http://www.theory.caltech.edu/%7Epreskill/ph219/index.html#lecture http://www.theory.caltech.edu/%7Epreskill/ph219/index.html#l... and probably get the books too (hey, they're cheap on amazon... god I hate myself).
    - krastanov 8y ago
usnigg 8y ago
Having worked on quantum computing in some of the leading labs for nearly a decade I do share some of the author's pessimism. My feeling is that theory once again is way ahead of experiment here. Also many theoretician who work in the field of quantum information are all too happily ignoring all the details that actually make it so darn hard to maintain quantum coherence and scale. Having said that I still think/hope that some day we will manage to harness nature at the level of individual quanta with the level of precision required to perform some non-trivial quantum computation. We're not there yet and I think some major technological breakthroughs need to happen before that. I am afraid there's too much hype around QC and when it fades it will bring down the field with it.
catpolice 8y ago
The core argument here, concerning the number of variables that need to be manipulated, isn't very clear to me - I can't tell if that's because it's an unclear argument in general, or if the author tried to simplify it for laypeople to the extent that it became unconvincingly vague for people who know a thing or two about the subject area. Given the status of the author, I'm willing to give him the benefit of the doubt (and also assume that some clarity was lost in translation). Here's an easier-to-follow argument I've heard before (also simplified for laypeople) that might be close enough to the one the author intends to make to be useful: Quantum algorithms are supposed to be much more efficient than conventional algorithms at certain kinds of tasks. That efficiency is calculated with the assumption that each system starts in a state where it's prepared to execute the algorithm. For conventional computers, that makes sense for a number of reasons (e.g. data/program input is linear). For quantum computers, this assumption doesn't necessarily make sense because it takes a significant amount of work (in terms of measurements, etc.) to get the system (the machinery and physical representation of the initial data) into the right quantum state. In fact, the amount of work required scales horribly with the size of the input (in terms of qubits). If you think of those state preparation steps as part of the cost of the algorithm, the quantum algorithms won't, in general, outperform the classical ones. I've been out of the quantum computation loop for nearly a decade, so I have no idea whether that's a fair way of presenting the problem, but hopefully it helps someone understand the claims being made in the article.
- ThePhysicist 8y ago
  Gate-based quantum computers are similar to classical computers in that you can decompose any operation into a combination of a small set of universal gates (like single-bit NOT and two-bit NAND for classical computers). Typically you combine arbitrary single-qubit rotations along the X, Y, Z axes with a two-qubit gate like CNOT or iSWAP to obtain a full set of universal gates. You can then decompose your algorithm into these gates for running it, which is often doable efficiently (e.g. the Grover algorithm can be implemented easily using the gate set described above). You can also efficiently implement any classical gate operations on a quantum computer using a set of universal, reversible gates (e.g. single qubit NOT and classical CNOT). Finally, you need to read qubit states so you have to provide a readout of individual qubits, typically you implement the projection operator along the Z axis. Now, for each of these operations you can characterize your error rate. For the processor I built during my PhD those error rates were around 9 % for qubit readout, 10 % for the universal two-qubit gate and around 1 % for single qubit gates. These errors stemmed from qubit decoherence (dephasing and relaxation), limitations of the readout and limitations of the signal precision for the manipulation. Errors are often different for each qubit and depend on the operation and the state the qubits are in (there’s crosstalk and different decoherence depending on the qubit frequencies as well), but you can usually give an upper bound for the error. Systems today are much better already btw. Given the different error rates you can very easily see how much error you will accumulate during a given gate sequence and what algorithms you will be able to run. We were for example able to perform the 2-qubit Grover algorithm with more than 50 % fidelity, which demonstrated quantum speed up for that very simple case. For a small subset of your qubits you can also perform quantum state tomography and quantum process tomography to statistically characterize their state and the quantum evolution they went through (including errors and decoherence), which allows you to characterize the errors in your system experimentally in a more precise way. Again this technique is limited by the fidelity of single qubit operations and readout, but it can nevertheless provide very good data for error analysis. For larger qubit systems tomography becomes intractable but there are other techniques that can allow you to characterize errors for these systems as well (e.g. randomized benchmarking). So please stop telling people that it’s not possible to quantify errors, it is simply not true. Quantum computers are like classical computers in the sense that they obey the laws of physics and have various error sources that can be measured and characterized. The precision that’s required is higher than for classical computers but there’s no fundamental reason why quantum computers can’t work, it’s just a though engineering problem.
adamnemecek 8y ago
This read a lot like Clifford Stoll’s 1995 essay "Why web won't be nirvana" https://www.newsweek.com/clifford-stoll-why-web-wont-be-nirvana-185306 https://www.newsweek.com/clifford-stoll-why-web-wont-be-nirv...
tezthenerd 8y ago
Here is a way to see the fallacy of the OMG, its 10^300 variables, thats crazy style of “argument”. Consider a probabilistic classical algorithm on 500 bits. Perhaps a Monte Carlo simulation of an Ising model for example. Note first that the most general probability distribution over the 500 classical bits takes 2^500 real numbers to specify. (You have to specify P(000…0) and P(000…1) and… P(111…1)). [You should compare this to the 2^501 real parameters it takes to specify the quantum state of 500 qubits.] To generate the most general such distribution perhaps you are restricted to using circuits where the gates act on at most n-bits at a time. Each gate can be described by a 2^n x 2^n bistochastic matrix comprised of (n-1)^2 real parameters. [You should compare this to the n^2 real parameters it takes to specify a 2^n x 2^n unitary matrix for a quantum gate acting on n qubits.] Obviously its nuts to imagine you can generate all classical probability distributions over the 2^500 real parameters, particularly if you’re so mad as to think you are going to do it using a circuit comprised only of these n-bit gates! Therefore useful classical monte carlo computing is obviously decades away. (Oh and please trust me, I'm well know in stuff that isn't quantum computing.)
- atq2119 8y ago
  Monte Carlo simulation doesn't generate all classical probability distributions though. Attempting to do that would be nuts. Monte Carlo simulation only samples from a single, fixed distribution (or maybe a small number of distribution). I'm not super convinced by the argument based on number of parameters either, but your analogy doesn't refute it at all.
  - tezthenerd 8y ago
    Its not meant to be an analogy. We also will not attempt to generate all quantum states on a quantum computer. As with classical monte carlo, we will only ever generate a tiny fraction of the possible quantum states/distributions, and will also sample from a fixed distribution (whatever the quantum circuit outputs, we measure it in a fixed basis and always draw samples from that single, fixed distribution). We will also achieve robustness against the tyranny of the real numbers in our gate parameters in a very similar way that a classical computer does when it approximates some idealized Monte Carlo algorithm.
amelius 8y ago
I'm wondering what will be first: practical fusion reactors or practical QC. Also, will we at some point be able to linearly convert energy into compute power. In that case: would we still need QC? (You might say that this conversion is already possible, but it requires human interaction)
- TomMarius 8y ago
  Fusion reactor development was slowed down by the problem of superconductivity, which seems to be close to solved now (theoretically, of course).
mathgenius 8y ago
This is a lame article, bordering on click-bait. It is totally reasonable to predict a quantum computing winter, but all the supporting arguments here seem to be about how "hard-headed engineers" know best. > So the number of continuous parameters describing the state of such a useful quantum computer at any given moment must be at least 2^1000, which is to say about 10^300. That’s a very big number indeed. Yes this is the whole point: if it wasn't this big, we could just simulate it on a regular computer. > To repeat: A useful quantum computer needs to process a set of continuous parameters that is larger than the number of subatomic particles in the observable universe. Yep. > it’s absolutely unimaginable how to keep errors under control for the 10^300 continuous parameters that must be processed by a useful quantum computer. I do imagine it, every day. But I'm a professional, so maybe that doesn't count? > How many physical qubits would be required for each logical qubit? No one really knows... There are plenty of such estimates around. Obviously it depends on the noise rate of each qubit. > all of the assumptions that theorists make about the preparation of qubits into a given state, the operation of the quantum gates, the reliability of the measurements, and so forth, cannot be fulfilled exactly. They can only be approached with some limited precision. So, the real question is: What precision is required? ... Amazingly, not only are there no clear answers to these crucial questions, but they were never even discussed! There are boatloads of papers (and conferences) that discuss exactly this. Obviously it is a vital question to answer!
- kisstheblade 8y ago
  I think your reply is knee-jerk and "lame". Why don't you enlighten us, instead of taking it so personally when somebody criticizes your field. I thought the article was excellent and made good arguments.
  - eigenspace 8y ago
    How much working knowledge of quantum computing do you have? I can see why someone not connected to the field might think the article makes good arguments but if you are connected at all to the literature, one can see the article is bullshit. Many of the criticisms they’re levying here are either made up or misunderstood.
mcguire 8y ago
"While a conventional computer with N bits at any given moment must be in one of its 2N possible states, the state of a quantum computer with N qubits is described by the values of the 2N quantum amplitudes, which are continuous parameters (ones that can take on any value, not just a 0 or a 1)." If nothing else, that seems to me to be the clearest description of quantum computing I've seen.
- tezthenerd 8y ago
  [typo note its 2^N quantum amplitudes] A clear description but potentially very misleading and one that will certainly screw up your intuition about what to expect from quantum computers. As pointed out below, a classical probability distribution of N bits is a 2^N dimensional vector of real numbers. There are many very good reasons to think of the quantum state as "more like" this classical vector than the physical state of the "N classical bits" themselves. Here is one: - If I send you the physical systems which encode N classical bits then sure enough, I can communicate to you only N classical bits of information despite it taking exponentially many real parameters to specify the distribution/state I prepared them in. - If I send you the physical systems which encode N quantum bits then sure enough, I can communicate to you only N classical bits of information despite it taking exponentially many real parameters to specify the distribution/state I prepared them in. There are many other reasons to think of quantum states as not "inherently real" and more like the classical probability distribution. A key one is that both "instantaneously collapse" when you get information about the outcome of an observation. The issue of course is that while we know the "real states of the world are" of a conventional computer, nobody agrees on what (if any) they are for quantum systems (though many constraints on such purported real states are known, I write about some of them in _Q is for Quantum_)
  - wjhuggins 8y ago
    Actually, if you send a physical system which encodes N quantum bits you can send up to 2N classical bits. (check out https://en.wikipedia.org/wiki/Superdense_coding https://en.wikipedia.org/wiki/Superdense_coding)
calibas 8y ago
Has Google said what the error rate is yet for Bristlecone?
jvanderbot 8y ago
In addition to the technical arguments against this article, I say: "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." https://en.wikipedia.org/wiki/Clarke%27s_three_laws https://en.wikipedia.org/wiki/Clarke%27s_three_laws
maththrowawtays 8y ago
Many of the comments have pointed out this flaw in the argument. I will try to post my own summary of this flaw: The problem is that this case against quantum computing due to the sheer scale of the number of parameters is also a case against randomized computing, but randomized computing certainly exists in the real world (a computer with a coin flipper). A 'state' in randomized computing over N bits is specified by a probability being assigned to each of the 2^N possible states of the N bits. For example, to describe the distribution after flipping 100 coins, you require 2^100 real numbers to specify the probabilities for each outcome, since there are 2^100 combinations of heads and tails you can get. But it is very clear that we can do this in real life. The point is, you don't need 2^100 parameters with full precision to flip a coin 100 times. To be more precise, we can write the state as a 2^N length vector, and our operations (think logic gates) on the vector are called stochastic matrices (this is the type of matrix required to so that the operation maps states to states). Quantum computing is no different, except that instead of using real numbers for probabilities, complex numbers are used and they are called amplitudes instead. Also, instead of stochastic matrices for operations, they are now unitary matrices. The point is, the scale of the number of parameters is the same, the only thing that's changed is that it is complex numbers instead of real numbers. Now an argument might be that it is not possible to create gates with a very specific probability distribution with just a discrete and not very precise coin flipper, but it's actually possible to show that for randomized computing, you only need a constant factor more flips to get exponentially close to the distributions you want. A similar result was proved pretty early for the quantum case, and it is a foundational result. So this is not really a problem either.
lewisinc 8y ago
Shame on this article. We should be researching everything within moral limits. Everything. Research all of it so hard that there's no long any debate about whether this point or that has any standing because it's so well understood that we can just fucking move on. Regardless of the topic, if we aren't sure if it's of value - we learn as much as we can until we can allow ourselves as a species to move on. That's the only way to prove any point in the realm of curiosity that satisfies me. Articles that argue the validity of investgating an unknown for any reason other than "it would cause significant harm" don't hold any weight and should be disregarded as either being based in fear, self interest, or simple ignorance.
- kisstheblade 8y ago
  He said that the experimental research was useful, so shame on you. What kind of a silly strawman are you trying to build. I can see from the responses here that the field is obviously quite volatile and full of big egos with practicioners taking criticisms so personally :)
- carlmr 8y ago
  Completely agree. Imagine people had given up on the transistor. It took many attempts until there was a workable transistor. And then manufacturing still wasn't clear.
- ur-whale 8y ago
  The article isn't entirely worthless, the illustrations are really nice.

138 points by niccl 8y ago

Had to read to the bottom to get to the (really weak) "case against Quantum Computing": > I believe that, appearances to the contrary, the quantum computing fervor is nearing its end. That’s because a few decades is the maximum lifetime of any big bubble in technology or science. After a certain period, too many unfulfilled promises have been made, and anyone who has been following the topic starts to get annoyed by further announcements of impending breakthroughs. What’s more, by that time all the tenured faculty positions in the field are already occupied. The proponents have grown older and less zealous, while the younger generation seeks something completely new and more likely to succeed. I was really hoping for a real technical argument, not speculation about bubbles or politics in academia. (Especially given that most of the real research is done in industry.)

resource0x 8y ago
> I was really hoping for a real technical argument Please read the body of the article, not only the bottom. There's plenty of technical arguments in there.
setquk 8y ago
Sounds like Fusion. That’s still going on after 60 odd years.
- powera 8y ago
  With one major difference. We know that fusion in some form is possible (for example, in the sun, or in a hydrogen-bomb). The difficulty is doing it at a reasonable scale. It's less clear that quantum computing is possible.
  - sp332 8y ago
    I mean we already have quantum computers too. https://quantumexperience.ng.bluemix.net/qx/experience https://quantumexperience.ng.bluemix.net/qx/experience The trouble here is also scaling, but in the upward direction.
    - pavelrub 8y ago
      The "quantum computers" we have are classical computers with very weird hardware. The thing everybody is trying to achieve is quantum supremacy - this hasn't been achieved or demonstrated yet, and might never will be.
      - TallGuyShort 8y ago
        For at least a very, very long time, I see Quantum Computers as simply being "accelerators" that do very niche tasks. Like GPUs or FPGAs today. I'm sure one day we'll have a whole other paradigm, but I doubt Quantum Computing as we envision it today will be the thing that replaces x86-compatible silicon.
      - sp332 8y ago
        They are not classical computers. We've had actual quantum logic gates since 1995. https://en.wikipedia.org/wiki/Timeline_of_quantum_computing https://en.wikipedia.org/wiki/Timeline_of_quantum_computing And as linked above, IBM is renting out time on real quantum computers up to 17 qubits now. Intel, Google, and IBM had all announced plans for 49 or 50 qubit computers, which would have achieved quantum supremacy at the time, but IBM improved their simulator to 56 qubits last year. https://en.wikipedia.org/wiki/Quantum_supremacy https://en.wikipedia.org/wiki/Quantum_supremacy So we're only one or two hardware generations away now.
        pavelrub 8y ago
        They are classical computers in the sense they are are no stronger than classical computers - they provide a different physical implementation, but computationally can't achieve anything that a classical computer can't. Everything those "quantum computers" do can be simulated with a classical computer. The entire motivation behind the development of QC is supremacy, and the existence of those "quantum computers" alone doesn't provide any proof that supremacy is possible - so it is definitely not just a question of scale. Google already has a 72 qubit processor (Bristlecone), but whether they or anybody will be able to demonstrate supremacy in the current, next, or some future generation is entirely conjectural. The main limit is error correction, and there are theoretical arguments (Gil Kalai) for why achieving supremacy might be impossible.
  - core-questions 8y ago
    You can literally sign up for D-Wave Leap and run problems on a quantum computer immediately. It's not gate-model, but it's a real quantum annealer that functions and can compute useful things.
    - krastanov 8y ago
      Quantum annealers are not quantum computers. They are not more powerful than classical computers. You either need the "gate model" or the equivalent "adiabatic quantum computing model" to achieve something infeasible on classical computers. It is a common misconception to think that adiabatic quantum computing (which is equivalent to the gate model) and quantum annealing is the same thing. It is not, and quantum annealing is mostly a buzzword that D-Wave use to sell their devices.
- simonh 8y ago
  If something sounds too good to be true, it probably is.
  - tarboreus 8y ago
    Sulfa. White swans do come along, they're just incredibly rare.
    - WillPostForFood 8y ago
      Are they rare, or do they just happen over longer timeframes than we want? Just about any mainstream technology today would appear to be too good to be true to someone ~100 years ago. Air travel, television, computers, phones.
      - tarboreus 8y ago
        The telephone was invented in 1876, practically old hat a hundred years ago. Air travel was 1903, was around longer than Facebook has been around in 2018. I guess it comes down to your definition of rare, but it's certainly rare enough.
    - effie 8y ago
      Black swans are rare.
- Animats 8y ago
  We know that fusion is possible. We have H-bombs. We have various things that do small scale fusion. Getting out more power than is put in remains way out of reach.
  - hazeii 8y ago
    I take the point, while noting H-bombs put out more power than is put in. Perhaps "getting more useful power out..."?
TheOtherHobbes 8y ago
The technical argument is absolutely clear: quantum computing cannot work because it relies on manipulating and measuring an absolutely astronomical number of continuous variables with near-infinite precision. The argument may or may not be correct, but it deserves something more thoughtful than a dismissive response that doesn't even recognise the basic point the author is making.
- krastanov 8y ago
  This is one of the standard complaints (the gist of it being that quantum computing is some form of analog computing, i.e. requiring near-infinite precision). For researchers in the field it becomes rather frustrating to have to repeat the same response without being heard, so I can understand the annoyance expressed in the parent comment. For what is worth, here is a good explanation of how this complaint misrepresents the whole premise of quantum computation (in particular, point 6): https://www.scottaaronson.com/democritus/lec14.html https://www.scottaaronson.com/democritus/lec14.html
  - core-questions 8y ago
    > the gist of it being that quantum computing is some form of analog computing It absolutely is - at least, in the only practical, real-today, working instantiation of it which is in the form of quantum annealing.
    - krastanov 8y ago
      This is an extremely misleading statement. Quantum annealing is most certainly not what quantum computing is about (it is not particularly "quantum" either). At most, you can argue that it is an important first step (which is also doubtful). Most of the interesting hardware currently being developed has nothing to do with quantum annealing and most researchers are fairly annoyed at D-wave for originally pushing this nonsense (D-wave is the company that started selling quantum annealing machines, which are little more than a classical analog computer). See the ion traps developed in the UK/Maryland or the transmon qubits at Google/IBM/Yale for some well known prototypes of quantum computing hardware (which are admittedly quite far from being practical or usable).
scottlocklin 8y ago
There's certainly a bubble here; trivially obvious, but there are also very good technical arguments in that article. FWIIW Dyakonov has been saying the same thing since the bubble got started in 2000 or so.

nixpulvis 8y ago

I still need to gain the intuition for why a quantum computer can operate on some kinds of things "faster". I've read some of the math, but that did little to satisfy me (I need to study it more clearly). But all of this seems in a tragic state at the moment. Allowing rampant misinformation and hype as to what these machines are actually capable of.

msclrhd 8y ago
It's not that they can execute a computation faster in the traditional sense. It is that you can exploit the quantum behaviour in some types of computation (e.g. factoring primes) to make the answer more probable, and thus more likely to be "observed", when you read the state of the quantum computer and thus make it easy to locate the correct answer compared to traditional algorithms. PBS Infinite Series has several videos on quantum computing and Shor's algorithm: 1. https://www.youtube.com/watch?v=IrbJYsep45E https://www.youtube.com/watch?v=IrbJYsep45E "The Mathematics of Quantum Computers" 2. https://www.youtube.com/watch?v=12Q3Mrh03Gk https://www.youtube.com/watch?v=12Q3Mrh03Gk "How to Break Cryptography" 3. https://www.youtube.com/watch?v=wUwZZaI5u0c https://www.youtube.com/watch?v=wUwZZaI5u0c "Hacking at Quantum Speed with Shor's Algorthm"
krastanov 8y ago
If you are specifically interested in building up the mathematical intuition, check out "Quantum Computing since Democritus" by Aaronson - it excels at presenting the intuition. For less comp-sci and more physics look at the standard references: Nielsen and Chuang's book and Preskill's lecture notes.
- nixpulvis 8y ago
  Thanks for the links. I'll start here: http://www.theory.caltech.edu/%7Epreskill/ph219/index.html#lecture http://www.theory.caltech.edu/%7Epreskill/ph219/index.html#l... and probably get the books too (hey, they're cheap on amazon... god I hate myself).
  - krastanov 8y ago

usnigg 8y ago

Having worked on quantum computing in some of the leading labs for nearly a decade I do share some of the author's pessimism. My feeling is that theory once again is way ahead of experiment here. Also many theoretician who work in the field of quantum information are all too happily ignoring all the details that actually make it so darn hard to maintain quantum coherence and scale. Having said that I still think/hope that some day we will manage to harness nature at the level of individual quanta with the level of precision required to perform some non-trivial quantum computation. We're not there yet and I think some major technological breakthroughs need to happen before that. I am afraid there's too much hype around QC and when it fades it will bring down the field with it.

catpolice 8y ago

The core argument here, concerning the number of variables that need to be manipulated, isn't very clear to me - I can't tell if that's because it's an unclear argument in general, or if the author tried to simplify it for laypeople to the extent that it became unconvincingly vague for people who know a thing or two about the subject area. Given the status of the author, I'm willing to give him the benefit of the doubt (and also assume that some clarity was lost in translation). Here's an easier-to-follow argument I've heard before (also simplified for laypeople) that might be close enough to the one the author intends to make to be useful: Quantum algorithms are supposed to be much more efficient than conventional algorithms at certain kinds of tasks. That efficiency is calculated with the assumption that each system starts in a state where it's prepared to execute the algorithm. For conventional computers, that makes sense for a number of reasons (e.g. data/program input is linear). For quantum computers, this assumption doesn't necessarily make sense because it takes a significant amount of work (in terms of measurements, etc.) to get the system (the machinery and physical representation of the initial data) into the right quantum state. In fact, the amount of work required scales horribly with the size of the input (in terms of qubits). If you think of those state preparation steps as part of the cost of the algorithm, the quantum algorithms won't, in general, outperform the classical ones. I've been out of the quantum computation loop for nearly a decade, so I have no idea whether that's a fair way of presenting the problem, but hopefully it helps someone understand the claims being made in the article.

ThePhysicist 8y ago
Gate-based quantum computers are similar to classical computers in that you can decompose any operation into a combination of a small set of universal gates (like single-bit NOT and two-bit NAND for classical computers). Typically you combine arbitrary single-qubit rotations along the X, Y, Z axes with a two-qubit gate like CNOT or iSWAP to obtain a full set of universal gates. You can then decompose your algorithm into these gates for running it, which is often doable efficiently (e.g. the Grover algorithm can be implemented easily using the gate set described above). You can also efficiently implement any classical gate operations on a quantum computer using a set of universal, reversible gates (e.g. single qubit NOT and classical CNOT). Finally, you need to read qubit states so you have to provide a readout of individual qubits, typically you implement the projection operator along the Z axis. Now, for each of these operations you can characterize your error rate. For the processor I built during my PhD those error rates were around 9 % for qubit readout, 10 % for the universal two-qubit gate and around 1 % for single qubit gates. These errors stemmed from qubit decoherence (dephasing and relaxation), limitations of the readout and limitations of the signal precision for the manipulation. Errors are often different for each qubit and depend on the operation and the state the qubits are in (there’s crosstalk and different decoherence depending on the qubit frequencies as well), but you can usually give an upper bound for the error. Systems today are much better already btw. Given the different error rates you can very easily see how much error you will accumulate during a given gate sequence and what algorithms you will be able to run. We were for example able to perform the 2-qubit Grover algorithm with more than 50 % fidelity, which demonstrated quantum speed up for that very simple case. For a small subset of your qubits you can also perform quantum state tomography and quantum process tomography to statistically characterize their state and the quantum evolution they went through (including errors and decoherence), which allows you to characterize the errors in your system experimentally in a more precise way. Again this technique is limited by the fidelity of single qubit operations and readout, but it can nevertheless provide very good data for error analysis. For larger qubit systems tomography becomes intractable but there are other techniques that can allow you to characterize errors for these systems as well (e.g. randomized benchmarking). So please stop telling people that it’s not possible to quantify errors, it is simply not true. Quantum computers are like classical computers in the sense that they obey the laws of physics and have various error sources that can be measured and characterized. The precision that’s required is higher than for classical computers but there’s no fundamental reason why quantum computers can’t work, it’s just a though engineering problem.

adamnemecek 8y ago

This read a lot like Clifford Stoll’s 1995 essay "Why web won't be nirvana" https://www.newsweek.com/clifford-stoll-why-web-wont-be-nirvana-185306 https://www.newsweek.com/clifford-stoll-why-web-wont-be-nirv...

tezthenerd 8y ago

Here is a way to see the fallacy of the OMG, its 10^300 variables, thats crazy style of “argument”. Consider a probabilistic classical algorithm on 500 bits. Perhaps a Monte Carlo simulation of an Ising model for example. Note first that the most general probability distribution over the 500 classical bits takes 2^500 real numbers to specify. (You have to specify P(000…0) and P(000…1) and… P(111…1)). [You should compare this to the 2^501 real parameters it takes to specify the quantum state of 500 qubits.] To generate the most general such distribution perhaps you are restricted to using circuits where the gates act on at most n-bits at a time. Each gate can be described by a 2^n x 2^n bistochastic matrix comprised of (n-1)^2 real parameters. [You should compare this to the n^2 real parameters it takes to specify a 2^n x 2^n unitary matrix for a quantum gate acting on n qubits.] Obviously its nuts to imagine you can generate all classical probability distributions over the 2^500 real parameters, particularly if you’re so mad as to think you are going to do it using a circuit comprised only of these n-bit gates! Therefore useful classical monte carlo computing is obviously decades away. (Oh and please trust me, I'm well know in stuff that isn't quantum computing.)

atq2119 8y ago
Monte Carlo simulation doesn't generate all classical probability distributions though. Attempting to do that would be nuts. Monte Carlo simulation only samples from a single, fixed distribution (or maybe a small number of distribution). I'm not super convinced by the argument based on number of parameters either, but your analogy doesn't refute it at all.
- tezthenerd 8y ago
  Its not meant to be an analogy. We also will not attempt to generate all quantum states on a quantum computer. As with classical monte carlo, we will only ever generate a tiny fraction of the possible quantum states/distributions, and will also sample from a fixed distribution (whatever the quantum circuit outputs, we measure it in a fixed basis and always draw samples from that single, fixed distribution). We will also achieve robustness against the tyranny of the real numbers in our gate parameters in a very similar way that a classical computer does when it approximates some idealized Monte Carlo algorithm.

amelius 8y ago

I'm wondering what will be first: practical fusion reactors or practical QC. Also, will we at some point be able to linearly convert energy into compute power. In that case: would we still need QC? (You might say that this conversion is already possible, but it requires human interaction)

TomMarius 8y ago
Fusion reactor development was slowed down by the problem of superconductivity, which seems to be close to solved now (theoretically, of course).

mathgenius 8y ago

This is a lame article, bordering on click-bait. It is totally reasonable to predict a quantum computing winter, but all the supporting arguments here seem to be about how "hard-headed engineers" know best. > So the number of continuous parameters describing the state of such a useful quantum computer at any given moment must be at least 2^1000, which is to say about 10^300. That’s a very big number indeed. Yes this is the whole point: if it wasn't this big, we could just simulate it on a regular computer. > To repeat: A useful quantum computer needs to process a set of continuous parameters that is larger than the number of subatomic particles in the observable universe. Yep. > it’s absolutely unimaginable how to keep errors under control for the 10^300 continuous parameters that must be processed by a useful quantum computer. I do imagine it, every day. But I'm a professional, so maybe that doesn't count? > How many physical qubits would be required for each logical qubit? No one really knows... There are plenty of such estimates around. Obviously it depends on the noise rate of each qubit. > all of the assumptions that theorists make about the preparation of qubits into a given state, the operation of the quantum gates, the reliability of the measurements, and so forth, cannot be fulfilled exactly. They can only be approached with some limited precision. So, the real question is: What precision is required? ... Amazingly, not only are there no clear answers to these crucial questions, but they were never even discussed! There are boatloads of papers (and conferences) that discuss exactly this. Obviously it is a vital question to answer!

kisstheblade 8y ago
I think your reply is knee-jerk and "lame". Why don't you enlighten us, instead of taking it so personally when somebody criticizes your field. I thought the article was excellent and made good arguments.
- eigenspace 8y ago
  How much working knowledge of quantum computing do you have? I can see why someone not connected to the field might think the article makes good arguments but if you are connected at all to the literature, one can see the article is bullshit. Many of the criticisms they’re levying here are either made up or misunderstood.

mcguire 8y ago

"While a conventional computer with N bits at any given moment must be in one of its 2N possible states, the state of a quantum computer with N qubits is described by the values of the 2N quantum amplitudes, which are continuous parameters (ones that can take on any value, not just a 0 or a 1)." If nothing else, that seems to me to be the clearest description of quantum computing I've seen.

tezthenerd 8y ago
[typo note its 2^N quantum amplitudes] A clear description but potentially very misleading and one that will certainly screw up your intuition about what to expect from quantum computers. As pointed out below, a classical probability distribution of N bits is a 2^N dimensional vector of real numbers. There are many very good reasons to think of the quantum state as "more like" this classical vector than the physical state of the "N classical bits" themselves. Here is one: - If I send you the physical systems which encode N classical bits then sure enough, I can communicate to you only N classical bits of information despite it taking exponentially many real parameters to specify the distribution/state I prepared them in. - If I send you the physical systems which encode N quantum bits then sure enough, I can communicate to you only N classical bits of information despite it taking exponentially many real parameters to specify the distribution/state I prepared them in. There are many other reasons to think of quantum states as not "inherently real" and more like the classical probability distribution. A key one is that both "instantaneously collapse" when you get information about the outcome of an observation. The issue of course is that while we know the "real states of the world are" of a conventional computer, nobody agrees on what (if any) they are for quantum systems (though many constraints on such purported real states are known, I write about some of them in _Q is for Quantum_)
- wjhuggins 8y ago
  Actually, if you send a physical system which encodes N quantum bits you can send up to 2N classical bits. (check out https://en.wikipedia.org/wiki/Superdense_coding https://en.wikipedia.org/wiki/Superdense_coding)

calibas 8y ago

Has Google said what the error rate is yet for Bristlecone?

jvanderbot 8y ago

In addition to the technical arguments against this article, I say: "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." https://en.wikipedia.org/wiki/Clarke%27s_three_laws https://en.wikipedia.org/wiki/Clarke%27s_three_laws

maththrowawtays 8y ago

Many of the comments have pointed out this flaw in the argument. I will try to post my own summary of this flaw: The problem is that this case against quantum computing due to the sheer scale of the number of parameters is also a case against randomized computing, but randomized computing certainly exists in the real world (a computer with a coin flipper). A 'state' in randomized computing over N bits is specified by a probability being assigned to each of the 2^N possible states of the N bits. For example, to describe the distribution after flipping 100 coins, you require 2^100 real numbers to specify the probabilities for each outcome, since there are 2^100 combinations of heads and tails you can get. But it is very clear that we can do this in real life. The point is, you don't need 2^100 parameters with full precision to flip a coin 100 times. To be more precise, we can write the state as a 2^N length vector, and our operations (think logic gates) on the vector are called stochastic matrices (this is the type of matrix required to so that the operation maps states to states). Quantum computing is no different, except that instead of using real numbers for probabilities, complex numbers are used and they are called amplitudes instead. Also, instead of stochastic matrices for operations, they are now unitary matrices. The point is, the scale of the number of parameters is the same, the only thing that's changed is that it is complex numbers instead of real numbers. Now an argument might be that it is not possible to create gates with a very specific probability distribution with just a discrete and not very precise coin flipper, but it's actually possible to show that for randomized computing, you only need a constant factor more flips to get exponentially close to the distributions you want. A similar result was proved pretty early for the quantum case, and it is a foundational result. So this is not really a problem either.

lewisinc 8y ago

Shame on this article. We should be researching everything within moral limits. Everything. Research all of it so hard that there's no long any debate about whether this point or that has any standing because it's so well understood that we can just fucking move on. Regardless of the topic, if we aren't sure if it's of value - we learn as much as we can until we can allow ourselves as a species to move on. That's the only way to prove any point in the realm of curiosity that satisfies me. Articles that argue the validity of investgating an unknown for any reason other than "it would cause significant harm" don't hold any weight and should be disregarded as either being based in fear, self interest, or simple ignorance.

kisstheblade 8y ago
He said that the experimental research was useful, so shame on you. What kind of a silly strawman are you trying to build. I can see from the responses here that the field is obviously quite volatile and full of big egos with practicioners taking criticisms so personally :)
carlmr 8y ago
Completely agree. Imagine people had given up on the transistor. It took many attempts until there was a workable transistor. And then manufacturing still wasn't clear.
ur-whale 8y ago
The article isn't entirely worthless, the illustrations are really nice.