TIME’s cover story on D-Wave: A case study in the conventions of modern journalism
This morning, commenter rrtucci pointed me to TIME Magazine’s cover story about D-Wave (yes, in today’s digital media environment, I need Shtetl-Optimized readers to tell me what’s on the cover of TIME…). rrtucci predicted that, soon after reading the article, I’d be hospitalized with a severe stress-induced bleeding ulcer. Undeterred, I grit my teeth, paid the $5 to go behind the paywall, and read the article.
The article, by Lev Grossman, could certainly be a lot worse. If you get to the end, it discusses the experiments by Matthias Troyer’s group, and it makes clear the lack of any practically-relevant speedup today from the D-Wave devices. It also includes a few skeptical quotes:
“In quantum computing, we have to be careful what we mean by ‘utilizing quantum effects,'” says Monroe, the University of Maryland scientist, who’s among the doubters. “This generally means that we are able to store superpositions of information in such a way that the system retains its ‘fuzziness,’ or quantum coherence, so that it can perform tasks that are impossible otherwise. And by that token there is no evidence that the D-Wave machine is utilizing quantum effects.”
One of the closest observers of the controversy has been Scott Aaronson, an associate professor at MIT and the author of a highly influential quantum-computing blog [aww, shucks –SA]. He remains, at best, cautious. “I’m convinced … that interesting quantum effects are probably present in D-Wave’s devices,” he wrote in an email. “But I’m not convinced that those effects, right now, are playing any causal role in solving any problems faster than we could solve them with a classical computer. Nor do I think there’s any good argument that D-Wave’s current approach, scaled up, will lead to such a speedup in the future. It might, but there’s currently no good reason to think so.”
Happily, the quote from me is something that I actually agreed with at the time I said it! Today, having read the Shin et al. paper—which hadn’t yet come out when Grossman emailed me—I might tone down the statement “I’m convinced … that interesting quantum effects are probably present” to something like: “there’s pretty good evidence for quantum effects like entanglement at a ‘local’ level, but at the ‘global’ level we really have no idea.”
Alas, ultimately I regard this article as another victim (through no fault of the writer, possibly) of the strange conventions of modern journalism. Maybe I can best explain those conventions with a quickie illustration:
MAGIC 8-BALL: THE RENEGADE MATH WHIZ WHO COULD CHANGE NUMBERS FOREVER
An eccentric billionaire, whose fascinating hobbies include nude skydiving and shark-taming, has been shaking up the scientific world lately with his controversial claim that 8+0 equals 17 [… six more pages about the billionaire redacted …] It must be said that mathematicians, who we reached for comment because we’re diligent reporters, have tended to be miffed, skeptical, and sometimes even sarcastic about the billionaire’s claims. Not surprisingly, though, the billionaire and his supporters have had some dismissive comments of their own about the mathematicians. So, which side is right? Or is the truth somewhere in the middle? At this early stage, it’s hard for an outsider to say. In the meantime, the raging controversy itself is reason enough for us to be covering this story using this story template. Stay tuned for more!
As shown (for example) by Will Bourne’s story in Inc. magazine, it’s possible for a popular magazine to break out of the above template when covering D-Wave, or at least bend it more toward reality. But it’s not easy.
More detailed comments:
- The article gets off on a weird foot in the very first paragraph, describing the insides of D-Wave’s devices as “the coldest place in the universe.” Err, 20mK is pretty cold, but colder temperatures are routinely achieved in many other physics experiments. (Are D-Wave’s the coldest current, continuously-operating experiments, or something like that?
I dunno: counterexamples, anyone?I’ve learned from experts that they’re not, not even close. I heard from someone who had a bunch of dilution fridges running at 10mK in the lab he was emailing me from…)
- The article jumps enthusiastically into the standard Quantum Computing = Exponential Parallelism Fallacy (the QC=EPF), which is so common to QC journalism that I don’t know if it’s even worth pointing it out anymore (but here I am doing so).
- Commendably, the article states clearly that QCs would offer speedups only for certain specific problems, not others; that D-Wave’s devices are designed only for adiabatic optimization, and wouldn’t be useful (e.g.) for codebreaking; and that even for optimization, “D-Wave’s hardware isn’t powerful enough or well enough understood to show serious quantum speedup yet.” But there’s a crucial further point that the article doesn’t make: namely, that we have no idea yet whether adiabatic optimization is something where quantum computers can give any practically-important speedup. In other words, even if you could implement adiabatic optimization perfectly—at zero temperature, with zero decoherence—we still don’t know whether there’s any quantum speedup to be had that way, for any of the nifty applications that the article mentions: “software design, tumor treatments, logistical planning, the stock market, airlines schedules, the search for Earth-like planets in other solar systems, and in particular machine learning.” In that respect, adiabatic optimization is extremely different from (e.g.) Shor’s factoring algorithm or quantum simulation: things where we know how much speedup we could get, at least compared to the best currently-known classical algorithms. But I better stop now, since I feel myself entering an infinite loop (and I didn’t even need the adiabatic algorithm to detect it).
Follow
Comment #1 February 6th, 2014 at 1:45 pm
Dang! I’m too cheap to spend $5 on yet another sensational D-Wave hype article. If anyone notices a bootleg version online (or is generous & unethical enough to upload one) please post a link! 🙂
Comment #2 February 6th, 2014 at 1:57 pm
This comment may be construed ad hominem, but it’s funny / ironic that TIME assigned this article to a guy with a PhD in Comparative Literature who otherwise seems to cover the video games, viral videos & consumer electronics beat (courtesy Wikipedia). And whose past credentials include literary criticism & writing three books in the Fantasy genre.
Certainly, these aren’t disqualifications, but couldn’t TIME find anyone more suitable?
I hope I didn’t google up the wrong Lev Grossman.
Comment #3 February 6th, 2014 at 2:25 pm
Scott:
The lowest temperatures ever achieved by scientists can be found here: http://hypertextbook.com/facts/2001/NehemieCange.shtml
Comment #4 February 6th, 2014 at 2:51 pm
Here dwave implies that their fridges are off-the-shelf:
I thought I read some other place where they claimed that the turnkey aspect was their contribution. I find it a lot more plausible that they have a record for turnkey systems than the continuous operation record you proposed. This link reconciles the two claims by saying that they buy the fridges from others, but they are driving the development.
Comment #5 February 6th, 2014 at 3:06 pm
Rahul, I think of Lev Grossman as a science journalist, but I don’t remember why. His bio claims that his beat at Time is books first and technology second. Looking at his archive, it’s a pretty distant second, but it’s there and he’s been writing about it for more than a decade.
Comment #6 February 6th, 2014 at 3:19 pm
Oh, Come on Rahul, I think the artist that designed that TIME cover picture is a GENIUS. That cover is precisely optimized for maximum Scott annoyance. The Infinity Machine. I also would have thrown in the caption “The Perfect D Wave”
Comment #7 February 6th, 2014 at 3:25 pm
“Quantum computing uses strange subatomic behavior to exponentially speed up processing. It could be a revolution, or it could be wishful thinking”
You see: If the D-Wave Machine doesn’t fullfill its promise it proves that quantum computing is wishful thinking!
Comment #8 February 6th, 2014 at 3:30 pm
Michael Marthaler:
You see: If the D-Wave Machine doesn’t fullfill its promise it proves that quantum computing is wishful thinking!
Yes, thanks! That implication—never quite stated explicitly—sort of hung over the whole article, and the likelihood that readers would come away with that impression was one of the main things that bothered me.
Comment #9 February 6th, 2014 at 4:16 pm
Hey scott,
I commented on the time article with the below msg.
“dwave has proven nothing yet with its machines. read the excellent blog below:
https://scottaaronson-production.mystagingwebsite.com/?p=1400
“
Comment #10 February 6th, 2014 at 4:54 pm
Sol’s link is over a decade old, but it appears to me that the temperature it reports, 0.1nK, is still the record. The coldest known sustained natural temperature is 1K, in some dust cloud. I’ve heard it claimed that dust clouds produced by supernovas get down to µK, but I can’t find a source for it. In fact, googling for “cold supernova” turns up papers about 10K dust clouds – hotter than the universal average.
It’s easy to find discussion of plans for gravitational wave detectors operating at tens of milliKelvin (or even 1mK), but I’m not sure they were actually built.
Comment #11 February 6th, 2014 at 5:39 pm
Scott. Two things:
(1) I agree that DWave has shown no signs of speed-up. I have also not read the Time magazine article. That being said, I think your comparison with the eccentric billionaire is unfair. In this case, DWave is run by professional physicists, computer scientists and engineers with PhDs. Further, they didn’t have a bunch of money lying around to spend, instead they gathered funding from reputed sources. Thus, in the eyes of the layman, this is strong evidence that they are doing something useful. Of course, as we realize now, this evidence is not sufficient to reach the conclusion that DWave is useful, and I agree that journalists should update on this. But it is not unreasonable to reach such a conclusion.
(2) I think one of the reasons that the Exponential Parallelism Fallacy (EPF) is so common is that there seem to be no other good alternatives. The EPF appeals so well to intuition (doesn’t mean we should use it). The other day a friend asked me to explain why QC gives us a speedup, and I stuttered: “Interference, entanglement and such…”. He was not convinced. Indeed, there doesn’t seem to be any clear intuition for a speedup, is there one?
Comment #12 February 6th, 2014 at 6:45 pm
Millikelvin temps for matter in a dilution fridge are routine.
Atomic physics experiments on dilute gases of cold atoms (for Bose-Einstein condensation, etc) routinely go down to well below 100 nK.
The number below 1 nK quoted in another comment is presumably for nuclei in a solid; the remainder of the solid is not at that low temperature, and only weakly thermally coupled to the nuclei.
Comment #13 February 6th, 2014 at 9:19 pm
Sid #11:
(1) Point taken. I wasn’t trying to satirize D-Wave itself, but rather the “report-the-controversy” format that so many media outlets feel like they have to adhere to, with its implicit postmodern attitude that there’s no discernible truth beyond the competing claims and counterclaims of corporations, the government, academic scientists, and other interested parties. (When in fact, in case of disagreement, everyone should just listen to the academic scientists. 😉 ) A decade ago, the report-the-controversy format was a large part of how the Bush administration was able to get away with preposterous claims in election years and make the case for the Iraq war.
(2) The intuition for quantum speedup is that, for a few special problems, you can choreograph a pattern of quantum interference where the amplitudes corresponding to each wrong answer cancel each other out, while the amplitudes corresponding to the right answer reinforce. I tried my best to explain the intuition here, here, here, here, and here (among other places); you can judge for yourself to what extent I succeeded. A huge part of the problem is that people want a 30-second explanation, analogizing QC to something they already know, when if such an explanation were possible, then QC wouldn’t be so interesting. Give me 30 minutes, and I think I can convey some accurate impression of the end result of a century’s worth of insights about physics and computer science (which is really what we’re talking about here). But starting from zero, it’s hard to do it faster than that!
Comment #14 February 6th, 2014 at 10:53 pm
Let me just say that Grossman’s The Magicians and Magician Kings were very enjoyable. Kind of a dark take, and a psychologically revealing one, on both the Narnia and Harry Potter forms of fantasy. It might also be of interest to some readers that in these books the wizards are basically geeks.
Comment #15 February 6th, 2014 at 11:59 pm
Douglas Knight #4, when I spoke with Geordie he told me that the off the shelve fridges were not at all good enough at the beginning, but rather then building their own they partnered with the fridge makers to improve them (I assume using some cross-licensing agreements).
He also hinted that he wants to lower the temp further.
Comment #16 February 7th, 2014 at 12:08 am
Scott, what do you make of the claim, put forward by Helmut Katzgraber et al., that the current benchmarks are using the wrong problem set to possibly find a quantum speed-up with D-Wave’s machine?
Comment #17 February 7th, 2014 at 12:28 am
@quax:
Is he proposing the benchmark that would be right & fair? I read the paper but wasn’t sure.
A lot of this sounds like a religious belief / faith that what we have here is definitely a hammer and only if we look hard enough we are sure to find the right nail.
Comment #18 February 7th, 2014 at 8:33 am
About the tormented relationship between journalism and quantum computing: I’ve recently found, on an italian newspaper, an article about quantum computers.
There is a picture of some galaxy on the top, and the first 2 sentences of the article are:
“Potrebbe essere una delle cinque scoperte capaci di trasformare il mondo. Rendere possibile contare le stelle dell’Universo in appena qualche ora…”
which means:
“It could be one of the five discoveries able to change the world. It would make possible to count the stars of the Universe in just few hours…”
Comment #19 February 7th, 2014 at 12:39 pm
It isn’t really important for this discussion, but dilution refrigerators have reached temperature just below 2 mK in continuously operating, macroscopic, solid state systems (if you are willing to ignore atom traps as being microscopic and out of equilibrium). Nuclear demagnatization refrigerators have reached 10s of microkelvin in thermal equilibrium, but in single shot mode. Below that temperature, the idea of ‘temperature of a block of metal’ breaks down: the lattice vibrations, electrion motion, and nuclear spins decouple. Nuclear demagnatization can reduce the nuclear spin temperature to below a nanokelvin, but that isn’t thermalized to the phonons and electrons. Of course atom traps reach comparable temperatures, and are likewise out of equilibrium.
The reason commercially available fridges weren’t suitable for D-wave is probably a number of factors: The fridges large enough to handle their wiring load didn’t used to get cold enough — 20-30 mK was more common, and it could be worse once you add the hundreds of wires needed to operate a complex circuit like their devices. Also, commercial fridges have usually come with zero or minimal magnetic and RF shielding, leaving that up to individual researchers to add in to their taste. The d-wave machine, like all squid circuits, is *extremely* sensitive to stray magentic flux, and I believe they use multiple layers of magnetic shielding and also active field cancellation.
Comment #20 February 7th, 2014 at 8:42 pm
Evan #19: Thanks so much for clarifying!
Comment #21 February 8th, 2014 at 4:18 am
Scott, just a question in general, and I’m curious where you stand on this; let’s say there’s a company (not Dwave) that succeeded in creating a quantum computer – for real, but this will result in some groups losing their funding in quantum research including yours. What’s your reaction to losing the research and projects even though the new technology has enormous benefits to humanity?
Comment #22 February 8th, 2014 at 8:15 am
ramsey66, your hypothetical scenario is completely bizarre for the following reason: if someone built a practical QC, that would increase funding for academic QC research, not decrease it! I mean, it’s not like classical CS research dried up once the first digital computers were built. Instead it exploded, from Alan Turing and a few friends to tens of thousands of people. After all, we still needed to figure out what classical computers were good for, and how to program them. Likewise, I expect that practical QCs would enormously increase funding agencies’ interest in what we can and can’t do with QCs, which is the main question that I’ve been interested in since the beginning.
Having said that: if, somehow, there would be “enormous benefits to humanity,” but only if I abandoned QC research forever, then I like to think that I’d be man enough to accept that trade, and spend the rest my career studying (say) classical complexity only. And if I had to abandon all research? Well, fortunately, I already have tenure… 😉
Comment #23 February 8th, 2014 at 5:29 pm
Thanks Scott, I think I have a better understanding of your work and what you do. You should continue being a chief Dwave critic or Quantum critic in general unless your busy that is. I have a feeling that this debate will only get intense down the road. My only concern sometimes is when it gets too intense and colourful (that’s speaking about everybody) that it distracts us from the real science involved and may deteriorate our ability to have a clear conversation. Not to mention the confusion in the media and general public. But I think it’s better now compared to how it all started 😉
Comment #24 February 11th, 2014 at 1:15 am
I suspect that part of the problem is that D-Wave gets covered by tech-beat journalists, rather than science/’ideas’ section journalists. If you’re a tech-beat reporter then you’re used to justifiably treating NASA and Google as pretty heavy-hitting epistemic authorities in the domain of your reporting, and might see a disagreement between the tech developers of a company backed by NASA and Google and top academics as a standoff.
Comment #25 February 11th, 2014 at 11:34 am
If quantum computing were to take place it would need to account for paradox… the second level of paradox at least. In order to accomplish that the device would need to ‘peer’ into a wave… not just respond to it. Unfortunately that would require awareness as only awareness can focus its attention ‘in’. Consciousness, even in a computer, is linear and as such unable to qualify for such attention.
For more on that see link below…
http://whoneedsthehiggs.blogspot.com/
Respectfuly
Robert McCoy
bommac123@gmail.com
Comment #26 February 11th, 2014 at 12:01 pm
That’s very interesting, Robert, you keep working on it and don’t get discouraged by people who may call you a crank. Why would they even say that?
Comment #27 February 11th, 2014 at 12:06 pm
Very good Vadim… but ‘personal’ attacks are not an argument. Arguments can only occur if an individual is willing to investigate that which they are arguing ‘against’… so to speak. So by all means read my ‘argument’ and then feel free to disassemble it with logic.
Comment #28 February 11th, 2014 at 1:07 pm
Robert,
I’ve looked at your site and it’s just metaphysical rambling; there’s nothing in there worth taking seriously enough to bother arguing. It’s a lot like the Time Cube thing but at least the text isn’t as big.
Comment #29 February 11th, 2014 at 2:33 pm
Actually there are several points worth taking on. One is that linear waves produce linear ‘effects’… therefore ‘non-linear’ effects would have to be the product of something of a ‘non-linear’ origin… IE: coming from a non-linear direction. So if an individual wishes to create a quantum ‘effect’ it stands to reason that they should be ‘looking’ in a non-linear direction. As such the blog has much to offer in that regard for the individual who is interested in a direct experience of such… in addition to the ‘argument’.
If it is true that non-linear effects come from a non-linear direction then direct knowledge of such would be profound… and they would have ‘serious’ effects on the discussion of quantum computing… IMO.
Comment #30 February 11th, 2014 at 3:19 pm
If there’s a fast-lane to getting banned on Shetl, Robert McCoy’s webpage link ought to be self-recommending. 🙂
Comment #31 February 11th, 2014 at 3:33 pm
Rahul… once again, ridicule is not an argument. The question is quite simple… if a non-linear quantum effect is desired should not ‘that’ which is effected be looking in a non-linear direction?… be it a computer or a human?
I would submit that if a quantum computer is unable to ‘look’ in a non-linear direction then just what is it doing?
Comment #32 February 11th, 2014 at 7:19 pm
If he is wrong, why exactly? Vadim? Rahul?
Comment #33 February 12th, 2014 at 9:48 am
Not so much wrong, just crazy. One example is his above leap-from-left-field on awareness. Another is his using the term “linear” to mean whatever he wants to it mean, but not the same thing it normally means. But I see no reason to hijack Scott’s blog any further to point out what everyone already knows to be nonsense.
Comment #34 February 12th, 2014 at 11:01 am
This is one of those things in life, where if you don’t already see where he’s wrong, then explaining it to you is probably futile.
Comment #35 February 25th, 2014 at 1:49 pm
In case the readers of this blog are wondering, D-Wave Systems does have supporters at MIT.
D-Wave named to MIT Technology Review’s 2014 List of the 50 Smartest Companies-
http://www.technologyreview.com/featuredstory/524671/50-smartest-companies-2014/
D-Wave is giving a Closing Keynote at the 2014 MIT Sloan Tech Conference –
http://disruptinglife.mitsloantech.com/