Official google.org Blog: Open Source Altruism

Sports Bookie

We occasionally joke about the looming robot menace, but seriously. Discoblog has picked out the Niftiest Robots of 2009, but “Scariest” would have been an equally appropriate appellation.

Yes, there is a robot that crawls around inside your colon, not to mention a Japanese emobot, but the one I would least like to meet in a dark alley is Chembot. It’s a blob-shaped thing that uses jamming in granular materials to make a robot that can alter its shape.

Still pretty primitive, but you can see where we’re headed here. Don’t say we didn’t warn you.

Here are the opening sentences of Chapter Five of my book:

When most people hear ?scientist,? they think ?Einstein.? Albert Einstein is an iconic figure; not many theoretical physicists attain a level of celebrity where their likeness appears regularly on T-shirts. But it?s an intimidating, distant celebrity. Unlike, say, Tiger Woods, the precise achievements that Einstein is actually famous for remain somewhat mysterious to many people who would easily recognize his name.

And now we can add, “and some achievements should really stay mysterious, thanks just the same.”

Last week I Twittered/Facebooked some provocative results from a poll of philosophers. In particular, this little tidbit:

Teletransporter (new matter): survival or death?

Accept or lean toward: survival	337 / 931 (36.1%)
Other	304 / 931 (32.6%)
Accept or lean toward: death	290 / 931 (31.1%)

Yes, that’s all the detail presented in the question: “Teletransporter (new matter): survival or death?” As a professional philosopher, you’re supposed to be familiar with the issue, which I reconstruct as follows. Imagine that someone has invented a working teleportation device. You step in the box, lights flash and sparks fly, and “you” rematerialize in another box, exactly the same in every way, but constructed out of a completely new collection of atoms. The original version of you is destroyed. Did you die? (And then, what if a million years passed in between the two events?)

It would probably be annoying to real philosophers, but I personally put this question in the category of “Not that hard.” And I would phrase my answer as: “Who cares?” What we should care about is how well the teleporter actually works — is the reconstructed person really in exactly the same quantum state as the original one was in? Same memories, feelings, etc? That’s an interesting technology question.

But there’s no interesting question associated with “Did you really die when you were teleported?”, or “Are you really the same person after being teleported?” These are just bad questions. They assume a certain way of looking at the world that ceases to be useful once we’ve invented teleportation. Namely, they assume that there’s a certain “essence of you-ness” that is (somehow) associated with your physical body and continues through time. That’s a perfectly sensible way of talking in the real world, where we don’t have access to duplicator devices or transporter machines. But if we did, that conception would no longer be very useful. There is a person who stepped into the first box, and a person who stepped out of the second box, and obviously they have a lot in common. But to sit down and demand that we decide whether they are “really” the same person is just a waste of time — there is no such “really.”

Which isn’t to say there aren’t interesting questions along these lines, but they are operational questions — how should I actually act, or what should I actually expect to happen, in these situations? — rather than arid metaphysical ones. What if you murdered someone, and then teleported — would the reconstructed person still be guilty of murder? That’s not quite the right question, because it still relies on the slippery essence of continuous personhood, but there’s a closely related sensible question — should we treat the reconstructed person as if they had committed murder? And it seems to me that the answer is clearly “yes” — whatever good reasons we had for treating the pre-teleportation person in a certain way, those reasons should still apply to the post-teleportation person.

The issue of duplication seems much thornier to me than the issue of teleportation. If someone made an exact copy of a known murderer, should we treat both the original and the copy as murderers? (I vote “yes.”) Fine, but what about the view from the inside? Let’s say you have an offer to get paid $100 if you let yourself be copied, with the proviso that after being copied one of the two of you will randomly be chosen for immediate painless execution. Do you take that deal?

I think problems like that are legitimately interesting, although to a great extent their mystery relies on the inadequacy of our conceptions of death. Most of us don’t want to die, at least not right away. But if we did die, we’d be gone, and wouldn’t have any wants or desires any more — but it’s very hard to consistently reason that way. Note that if we replaced “immediate painless execution” with “prolonged torture,” it seems like a much more straightforward question.

This showed up in our long-ago discussion of the quantum suicide experiment. In the many-worlds interpretation of quantum mechanics, you can make measurements that split the wave function of the universe into distinct branches. In some sense, then, you really do have a duplicator machine — it’s just that the whole universe gets duplicated, not just you. Some folks have tried to argue against this idea by pushing adherents into a logical cul de sac. You shouldn’t (to make a long story short) be averse to bargains that leave you dead with large probability, as long as there exist branches of the wave function where you are alive and flourishing — after all, in the branches where you are dead you don’t care any more, right?

My point in that earlier post — a point I somehow managed to completely obscure — was that these are misleading thought experiments, because very few of us would take seriously the corresponding classical suicide experiment. “Here, I’ll flip a coin, and give you $100 if it’s heads and shoot you instantly dead if it’s tails. Deal?” Very little temptation to take that offer. But the logic is essentially the same — if you’re dead you don’t care, right? (For purposes of these thought experiments we always assume you have no friends or loved ones who would miss you; it’s just part of the philosophical game, not a comment on your actual social situation.)

At some point in thinking about the many-worlds interpretation, issues like this inevitably do come up. That’s what David Albert and I talked about a bit on Bloggingheads. There might be a certain measurement that yields result A 10% of the time, and result B 90% of the time. But in the MWI, the measurement splits the universe into two branches, and you end up either in the branch where you saw A or the branch where you saw B. What does it mean to say that you had a “10% chance of measuring A”? You either did or you didn’t — there is no ensemble of millions of you all doing the same experiment. People have made progress on these questions — here’s a talk by David Wallace on his work with David Deutsch in attacking this problem. (Don’t ask me why everyone who thinks about these issues is named “David.”) I haven’t ever looked at this work closely enough to have an informed opinion.

All I know is that being able to teleport around would be really cool.

Northern New Mexico is an absolutely fabulous place to live. But, on occasion, I wish I had a teleporter handy. One of those occasions would be when the “Swell of the Century” hits the Hawaiian Islands, as it did last week. It turned out to be more like the “Swell of the Decade”, but apparently was nonetheless quite spectacular. Spectacular enough to hold the Eddie, a big-wave contest that can only be held in epic conditions.



There’s some good video here and here. Conditions were apparently squirrely, so most of the footage consists of fairly spectacular drops leading to gnarly wipeouts. Every now and then someone (often Kelly Slater [at left above]) emerges intact. Unless you’ve been out in overhead+ surf, I think it’s hard to fathom just how powerful these waves can be. You’ve bodysurfed three footers, and you think you have a clue. But you don’t. For a little perspective, here’s an excellent video from a (professional) surfer who wiped out at Jaws (on the north shore of Maui; it only fires on big days [like last week]). He survived (more-or-less).

While Hawaii was being slammed by water horizontally, we got a vertical contribution (in the form of a couple of feet of snow). It’s time to go dust off my snowboard.

In one of the comments to Daniel’s post on the stolen climate emails, techskeptic points to a wonderful chart at Information is Beautiful. The author did a great deal of gruntwork to lay out the various arguments of “The Global Warming Skeptics” vs. “The Scientific Consensus.” As far as I can tell, it’s a legitimately balanced view of both sides, complete with citations. If you’re confused about the various issues and accusations being bandied back and forth, there are worse places to start. This is a small piece of the full chart.

Of course, there is no such thing as a purely objective and judgment-free presentation of data, no matter how scrupulously the data itself may be collected; if nothing else, we make choices about what data to present. And a side-by-side comparison chart like this can’t help but give a slightly misleading impression of the relative merits of the arguments, by putting the conclusions of an overwhelming majority of honest scientists up against the arguments of a fringe collection of politically-motivated activists. But it’s certainly good to see the actual issues arrayed in point-counterpoint format.

Still, there remains a somewhat intractable problem: when people are arguing about issues that necessarily require expert knowledge that not everyone can possibly take the time to acquire for themselves, how do we make judgments about who to believe?

This problem has been brought home by the incredibly depressing news that James Randi has come out in favor of global-warming denialism (via PZ Myers). Randi is generally a hero among fans of reason and skepticism, so it’s especially embarrassing to see how incredibly weak his reasoning is here. It basically amounts to: “The climate is complicated. And scientists don’t know everything. And I admit I don’t know much about the field. Therefore … we have good reason to distrust the overwhelming majority of experts!” Why Randi chose not to apply his vaunted powers of skepticism to the motivations behind the denialists remains a mystery.

This gets to the heart of why I’ve always been skeptical of the valorization of “skepticism.” I don’t want to be skeptical for the sake of being skeptical — I want to be right. To maximize my chances of being right, I will try to collect what information I can and evaluate it rationally. But part of that information has to include the nature of the people making arguments on either side of a debate. If one side consists of scientists who have spent years trying to understand a complicated system, and the other is a ragtag collection of individuals with perfectly obvious vested interests in the outcome, it makes sense to evaluate their claims accordingly.

By all means, we should apply our own powers of reason to every interesting problem. But when our reasoning leads to some conclusion at odds with the apparent consensus of a lot of smart people who seem to know what they’re talking about — whether it’s on the nature of dark energy, the best way to quantize gravity, the most effective route to health care reform, or the state of the environment — the burden is on us to understand the nature of that difference and try to reconcile it, not to take refuge in “experts don’t know everything” and related anti-intellectual piffle.

… not yet released, but we’ll find out in just a bit. 2:00 p.m. Pacific time, to be exact.

Last week we mongered the rumor that the CDMS experiment was going to announce an exciting new result soon — and that time is now. (My guess remains: some interesting data that falls well short of “we’ve discovered dark matter!”) If you’d like to watch the talks online, here you go:

• Kavli Institute for Particle Astrophysics and Cosmology


• Fermilab

Further bulletins as events warrant.

1:10 PM:  As previously reported by Sean, the physics community has been all atwitter recently with rumors that the CDMS collaboration has observed dark matter events.  Personally, I have heard rumors that they have either 0, 1, 3, or 4 signal events above the 0.5 events expected from background, with the number being dependent on who is propagating the rumor.  The journal Nature has denied that a paper will be published, but everyone seems to agree that a CDMS paper will appear on the arXiv tonight.

And now, finally, the day of reckoning has arrived.  Members of the CDMS collaboration are giving at least 2 simultaneous seminars this afternoon (SLAC and Fermilab).  2:00 PM  PST.  The excitement in the air is palpable.  Not much work is being done ? everyone is pretty much talking in the hallways, trying to pass the time until 2:00.  David E. Kaplan is here from Johns Hopkins University to film the event for a documentary he is making.  Rumors are spreading that at least 3 groups of theorists have papers ready to submit to the ArXiv after the seminars are concluded.  The era of data has begun!

While we wait, let?s recall how CDMS detects dark matter.  The Cryogenic Dark Matter Search consists of Ge detectors, situated deep underground in the Soudan mine in northern Minnesota.  The depth is required to reduce background from cosmic rays.  As the earth travels through the sky, it presumably sweeps through clumps of dark matter which subsequently scatters with the particles in the Ge.  The scattering is expected to be elastic, meaning the initial and final scattering particles are the same.  The  detector measures the ionization and phonons from these interactions.  Background events mainly create electron recoils, and so a nuclear recoil gives a good DM signal.  The energy of this recoil gives information about the mass of the DM particle and the number of events compared to the total amount of data taken gives the interaction rate.

1:30:   So what does a theorist do while waiting for the seminar?  A quick calculation, of course!  What else?  The energy of the highest observed recoil event, with an assumed DM velocity distribution,

gives a bound on the mass of the DM particle.    Using some DM velocity assumptions from 0912.2358, one could do a quick and dirty calculation!  My student Randy did so, and here’s the plot (vertical axis is WIMP mass in units of GeV and horizontal axis is the recoil energy (in GeV):

Taking this to the talk, and then can immediately read off the WIMP mass bound!  Wanna bet how many theory papers appearing tonight have this plot?

1:45:  In my seat – front and center.  The auditorium is packed – SRO.  Folks from UC Santa Cruz and Berkeley are here.  The speaker is Jodi Cooley from Sounthern Medthodist University and is wired and ready to go.

1:54: JoAnne beat me to it, but I?ll join in on the live blogging with the view from the back of the Auditorium! ?Risa

1:55:  Learned that there’s a group of theorists at the KITP in Santa Barbara watching this event over skype.

1:58: You can watch it yourself — Live Streaming Video starting at 2 PM:

http://www-group.slac.stanford.edu/kipac/cdms_live.html

– Risa

2:00:  A hush just came over the room…JoA

2:01:  Jodi is showing her outline – starts with an intro, then describes the detector, then gives the results, and then discusses future plans for the collaboration.   My colleague sitting next to me just leaned over and whispered ”It looks like a signal talk.” - JoA

2;05: liveblogging appears to be deleting some of our comments, so you gotta pay close attention! –Risa

2:12: Quick comment on JoAnne’s nice plot while we are still in the intro. It’s important to remember that the actual value of the WIMP mass depends, as JoAnne says, on “some

DM velocity assumptions”. The paper she cites is eminently reasonable, but the values are still wildly uncertain (see e.g. arXiv:0906.5361), so if we do actually have a dark matter detection, understanding the expected density and velocity distribution of dark matter in our own galaxy becomes much more important. –Risa

2:15:  slide 14 discusses the direct detection event rate.  Expected signal event rate is < 0.01 events/kg-day.  kg-day is the unit used by the experiments to measure their sensitivity – kg denotes the size of the detector, and day is the number of days the detector was taking data.  -JoA

2:20:  Each detector consists of 230 g Ge and 100 g Si.  There are 30 detectors stacked into 5 towers.  -JoA

2:23 Experiment is designed to have ~0.5 background events.

– Risa

2:24 1/2 mile Underground in the Soudan mine. Note that this is in Minnesota, not the Sudan. –Risa

2:23:  Slide 24:  they can select or tune the background level as desired by their cuts.  They choose 0.5 background events in the signal box as a target.  The depth is at 6060 mwe (meters water equivalent) – JoA

2:27:  About to open the signal box… – JoA

2:30: November 5, 2009 opened the box in a big conference call. –Risa

2:31:  Measured surface event background in 0.6 plus/minus 0.1 (stat).  They use 3 techniques to determine this.  Slide 30.  Neutron bckgrnd is 0.04 plus 0.04 minus 0.03.  Radiogenic bckgrnd in 0.03-0.06 events. – JoA

2:32: Experiment has 194.1 kg-days WIMP equivalent @ 60 GeV/c^2 –Risa

2:33:  THE NUMBER IS TWO!!!! – JoA

2:34: I consistently heard 3 events, so interesting comment on the rumor mill — Risa

2:35: The two events happened at different times, different months, different detectors. — Risa

2:38: Were in inner region of the detector where there is better background rejection. — Risa

2:40 Probability of observing these two events is 23%. Clearly not the solid detection we were all hoping for! –Risa

2:41:  Refined calculation of the surfact background including timing,  post unblinding leakage estimate is now 0.8 plus/minus 0.1 (stat) plus/minus 0.2 (sys)  -JoA

2;42:  The 23% probability of observing the 2 events includes the post-unblinding leakage.  -JoA

2:45  Summarizing now:  cross section limit is < 3.8 x 10^{-44} for WIMP mass of 70 GeV at 90% CL.  -JoA

2:47:  More towers of detectors have recently been installed, with more to be installed in 2010.  -JoA

2:48:  arXiv number is 0912.3320  -JoA

2:50: Bottom line: “The results cannot be interpreted as significant evidence for WIMP interactions, but we cannot reject the possibility that either event is signal.” –Risa

2:51: Question about other unanalyzed data. Answer: next step is really SuperCDMS. — Risa

2:52: The events took place on 8/5/07 and 10/27/07. –Risa

2:55: Very solid, careful work, a new upper limit, and a new thing to add to “hints of dark matter”. But could easily be a statistical fluctuation. Going to be a bit longer until something solid. — Risa

First event was detected on 10/27/07 with recoil energy of roughly 12 keV, and the second was on 08/05/07 with roughly 15 keV.   A 3^rd event lies just outside their box with recoil of 12 keV.  -JoA

2:59:  The crowd is filing out amid much discussion, while Dimopoulos and crowd are being interviewed.  There will be much discussion all afternoon! – JoA

3:01:  Given the above figure, if these events are interpreted as signal, the lower bound on the WIMP mass for these recoil energies is roughly 2 GeV.  Not that constraining!  -JoA

3:18: Summary in CDMS’s own words here (pdf). -Sean

4:00: Corrected figure now posted. -JoA

6:40: Can’t find the paper on the arXiv and 0912.3320 is a nice paper (on DM) written by my good friend and collaborator Hooman Davoudiasl, but has nothing to do with CDMS…

Jodi told me it was submitted at 1, just before the talk, which means it won’t be out until the next round (Sunday night, I think). –Risa

My favorite example of a recent Hollywood blockbuster that scientists should like is Iron Man. Yes, it’s implausible that a prisoner in a cave in Afghanistan could build a lethal flying suit out of scrap metal, etc. But plausibility should never be the criterion for judging a science-fiction/fantasy scenario; sometimes you just have to bend the rules of the real world to get the required dramatic effects. Consistency, on the other hand, is crucial; the non-real world you invent should follow some set of rules, even if they veer away from the actual world. (Nobody complains that the Enterprise travels faster than light, but there are plenty of complaints about the bizarre use of time travel in the Star Trek franchise.)

Even better is when a film does a decent job at reflecting the practice of science. And that’s why I loved Iron Man — the whole second act revolves around Tony Stark in his lab, engineering designs and using trial-and-error to determine experimentally what works and what doesn’t. It makes for compelling viewing, which should be a lesson to people.

So we’re all excited about Iron Man 2, right?

The Science and Entertainment Exchange had a small hand in this one — apparently they needed a particle physicist to help get some of the scenes right. I don’t think it was the scene with the whips.