If Zephod Beeblebrox shows up at the University of Illinois Urbana-Champaign and makes off with the interferometers, don’t say I didn’t warn you.

by John Holbo on February 24, 2006

I must confess: since I can’t really tell the difference between the method these folks used and the method these folks used, I should probably just stop having intuitions about the universe since “often deviates from intuitive reasoning, leading to some surprising effects” isn’t the half of it. Because, granting that they did what they did, my intution is that they can go on to develop infinite improbability computing, relying on the fact that their experiment cannot be scaled up to cause the scaled-up algorithm not to run, thereby producing the answer. Am I right (or am I right or am I right?)

Here’s a link to the Nature paper. (above link via boingboing.)

{ 3 trackbacks }

Constructive Interference » Some real (bizarre) constructive interference
02.26.06 at 2:21 pm
Quantum interrogation | Cosmic Variance
02.27.06 at 3:33 am
John Quiggin » Blog Archive » What I’ve been reading
03.05.06 at 8:10 am



dunno 02.24.06 at 12:34 am

Arch heresiarch!

(it’s Zaphod).

(Memory, sad but true).


Paul Orwin 02.24.06 at 12:35 am

Not to be a skeptic, but is there really a “disruptive technologies office?” It sounds fishy (much like the rest of quantum mechanics). It’s too early for an April Fool’s joke, though…


John Holbo 02.24.06 at 12:41 am

I thought exactly the same thing, Paul. I couldn’t find a disruptive technologies office, but there’s an actual Nature paper. And the people named in the article check out as real.


Daniel 02.24.06 at 1:45 am

c’mon our John. I always get depressed when I see intelligent arts people throwing up their hands in the face of science stuff. Remember these people aren’t demigods; they’re just called “physicists” because they did a physics degree.

It’s not very much more difficult to understand how a photon can simultaneously run and not run an algorithm than to understand how it can simultaneously interfere and not interfere with itself, and you can understand the two-slit experiment from reading the Feynman lectures which are perfectly comprehensible to an intelligent layman who is prepared to apply himself. Being able to perform the actual calculations would be tedious and difficult but the underlying principles are not beyond the ken of man.


Daniel 02.24.06 at 1:50 am

physicists; yadda yadda yadda two slit experiment not really anything to do with superposition. try to take in the broad historical sweep here will you?


John Holbo 02.24.06 at 1:55 am

I can understand the historical sweep – that is, I could write a screenplay that included an appropriate chase scene.


John Holbo 02.24.06 at 2:04 am

To put it another way: I think I understand the principles behind infinite improbability drive rather well. It’s just that I can’t either do the calculations that would actually build a working drive, nor can I conceive of how the principles could be true. (It is just a two-slit problem. But somehow more baroque, don’t you find?)


NPCurmudgeon 02.24.06 at 2:04 am

The “Disruptive Technologies Office” is the new name for an old agency: ARDA.

It’s worth noting that the computer in question must be turned on, and properly programmed/ready to go, so the savings in labor and energy are pretty minimal. In the past, it was thought that the probability that a final measurement of the system would yield the state |Program not run>|Correct answer obtained> was so small that a random guess would be more likely to succeed. The authors used a neat trick (a “Super Zeno Booster”!) to amplify the “interesting” probability above the random-guess threshold.

Daniel is absolutely right, of course: the calculations needed to explain these results are laborious but straightforward. What is not straightforward is the ingenuity needed to (1) think up paradoxical experiments that will probe the limits of quantum mechanics, and then (2) design and build the experiments themselves. This is why the field of quantum information science is so much fun.


neil morrison 02.24.06 at 3:16 am

Very clever, given this can’t be scaled up, the algorithm cannot be run so will therefore give an answer in a, zeno-enhanced, large probabilty of cases, and if by answer we mean “develop infinite improbability computing” then we arrive at Zaphod’s gadget.

But since we have no equiment by which we have scaled up we will never know where this gadget is. They, probably, lie all around us. But since I haven’t been looking…


karri 02.24.06 at 3:35 am

Daniel, from my (physics undergrad) perspective going through the Feynman lectures would seem to be actually harder than learning to calculate some elementary stuff with two-state quantum systems since that’s just linear algebra. Two-state systems allow a fairly rigorous but still easy formulation of quantum mechanics with a simple finite dimensional Hilbert-space and I sometimes wonder why they are not the starting point for teaching QM.


agm 02.24.06 at 4:19 am

Kwiat is quite real. Besides moonlighting as the Free Money For Your Life guy, he gives quite the entertaining talk for undergrad interns (which I once was).


Doug 02.24.06 at 4:49 am

And do you know why I want a cup of tea?


constablesavage 02.24.06 at 5:49 am

If you can run a program to its conclusion without starting it at all, doesn’t that mean you can run an infinitely long program in a finite time?

I’m feeling a bit worried here.


Dave F 02.24.06 at 6:36 am

If you use multiverse theory, all this run/not run, simultaneity of position etc becomes perfectly plain. There’s nothing mysterious about it.


Daniel 02.24.06 at 7:14 am

If you use multiverse theory, all this run/not run, simultaneity of position etc becomes perfectly plain. There’s nothing mysterious about it.

It doesn’t really; it explains things away rather than explaining them, by postulating a monstrous coincidence with respect to an utterly mysterious probability measure. It’s not really much of an advance on the theory that God wills it so.


Doug T 02.24.06 at 8:56 am

I understand the basic physics principle without any problem. But I’m a bit confused about the actual math here–from the newspaper blurb it sounded like they were doing something akin to the double slit, interfere with itself sort of thing, as Daniel points out.

But then it goes on to imply that they also know which state it was in (in this case, the “don’t run the algorithm” state.) And if you’ve measured to determine the particle’s path (in the two slit problem), then it no longer interferes with itself. So there’s something different going on here. I assume it’s explained in the Nature article (or in the references or the references of the references, since high profile physics articles are now constrained to being 3 pages long, which is much too short to actually explain any work starting from basics.), but that’s behind a subscription wall.

This in addition to the general weirdness of quantum computing. Of course, regular computing is pretty weird, too–I can tell you exactly how transistors work, at the electron/semiconductor physics level. But there’s a mysterious “then a miracle occurs” step two in my understanding going from that to an actual chip doing calculations. I have a similar problem transitioning in my mind from well understand prinsiples of quantum mechanics to performing calculations with them.

Just one more in the ever expanding list of fascinating things I don’t have time to learn.


Jonathan 02.24.06 at 11:36 am

The search algorithm and Zeno effect were the two things I found unusually mysterious out of the host of treasures.

Take the average humanities PhD as an example. Have that worthy read the Nature paper. Then read all of the references. Then read of the references’ references. Stop only when you get to something you’ve read before. What’s the average nesting distance?


garymar 02.24.06 at 12:04 pm

Experimenting with two slits and a chained Zeno effect? Count me in!


dipnut 02.24.06 at 3:00 pm

What the…? I did NOT write this comment!


Todd 02.24.06 at 7:16 pm

I can tell you exactly how transistors work, at the electron/semiconductor physics level. But there’s a mysterious “then a miracle occurs” step two in my understanding going from that to an actual chip doing calculations.

Really? That’s when it gets easy. At that point, it’s boolean logic all the way up.


Dan 02.24.06 at 10:28 pm

People who understand this experiment: where would you point us baffled non-physicists to get the hang of it?

I have a hard time finding anything between newspaper articles and popular science books (comprehensible, but only because they skip all the awkward bits, and make up with analogies and hand-waving), and the papers explaining the details (too much specialised math for the likes of me).


NPCurmudgeon 02.25.06 at 6:07 pm

People who understand this experiment: where would you point us baffled non-physicists to get the hang of it?

Dan, that’s such a good question that I scooted over to the Stanford bookstore to try to find a book that might be a good place to start. Unfortunately, I didn’t find any unambiguously good candidates for you: they were heavy on analogies and hand-waving, and light on the details. For example, I think that the book by Aczel on Entanglement would not be to your liking.

There are probably two reasonable places to start:

1. Wikipedia has a pretty good page on quantum mechanics, and another page on just the basics.

2. Daniel’s suggestion above is actually a pretty good one: try reading the third volume of the Feynman lectures. Feynman doesn’t spend much time on entanglement, but he does an excellent job on the two-slit experiment. There is math, but I bet that you could just skim over it on the first reading. (It’s mostly just algebra — even if it looks more abstract than you saw in high school — and the first year of calculus.)

If I were going to write a book on quantum mechanics for the layman, I think that I’d describe one experiment per chapter, anticipate the result using “classical” intuition, and then describe the actual result. The apparent paradox would then be a vehicle for explaining what seems to be so odd about the universe. (For example, after a discussion of the two-slit experiment and a Bell-inequality experiment, I’d discuss the quantum eraser.) After ten or so chapters, you could probably gain a pretty good understanding of the subject in a general way. I’d forget the history completely; there are enough books on that subject already. Maybe there’s a book available that’s like this, but I didn’t see it at the bookstore.


Dylan 02.25.06 at 9:46 pm

Wikipedia has a good page explaining a simpler version of this experiment: the Elitzur-Vaidman bomb testing problem. That describes it all quite well, I thought, in an essentially correct way.

I seem to remember there was a good Scientific American article a few years ago, as well.


John Quiggin 02.26.06 at 10:43 pm

I’m happy enough with Feynmann’s explanation of the two-slit experiment, and unlike Daniel I also like the multiverse representation.

But I’d like a decent reference on quantum computation focusing on the question implied by the post: can you get something for nothing here, and if so, how and in what sense?


John Quiggin 02.26.06 at 10:49 pm

Actually, the bomb testing example is a pretty good start, Dylan. If you’res till reading and can reference the SciAm article, that would be great.


art 03.01.06 at 2:27 am

Don’t panic! It’s as simple as…oh, damn, where’s my towel!

Comments on this entry are closed.