More Science Talk
Glenn Zorpette, executive editor of IEEE Spectrum magazine, and journalist John Horgan discuss various ideas related to what some call "the coming singularity," a point where computers will allegedly attain consciousness and superintelligence. Or not. Plus, we'll test your knowledge of some recent science in the news. Web sites mentioned on this episode include www.spectrum.ieee.org/singularity
Tyson: NOVA scienceNOW kicks off a brand new season on PBS, Wednesday July 25th at 9 A.M., 8 P.M. central, with me your host, Neil deGrasse Tyson.
Steve: Welcome to Science Talk, the weekly podcast of Scientific American for the seven days starting June 18th, 2008. I'm Steve Mirsky. This week on the podcast: why Ray Kurzweil's days are numbered. We will get to that in the course of a conversation about what a lot of people refer to as "the coming singularity". Our buddies at the magazine IEEE Spectrum devoted their June issue to this subject. The issue is called, "The Rapture of the Geeks: Separating Science from Fiction in the Technological Singularity". Last week I went over to Spectrum's offices here in New York City and spoke to two former Scientific American staffers, Glenn Zorpette and writer John Horgan, who contributed the lead article to the issue.
Steve: Glenn, you are the executive editor here—why don't you give us an overview. Tell us what this singularity is supposed to be for anybody who is coming into the story without a background, and then tell us what the whole overall kind of theme of your special issue is.
Zorpette: This whole issue originated with [a] very talented journalist here named Harry Goldstein, who actually runs our Web site, and he noticed a year or two ago that 2008 would be the 15th anniversary of a famous essay that was written by an author and retired computer scientist named Vernor Vinge; and Vinge wrote what is widely regarded as the manifesto that gave rise to this idea of the singularity. As John noted in his piece for us, the basic concept of the singularity in one form or other has been kicking around for decades, if not centuries. But Vinge laid down the modern version, which was that consciousness—there is nothing mystical about it—that consciousness as it occurs in human brains is something that is the result of physical processes and biological processes and chemical processes and because of that it will be reproducible. And this is where the singularity theory gets contentious, because Vinge as well as some of the other people who believe in this—we call them "singularitarians"—didn't think what is going to happen in 500 years or 600 years or 300 years, he and others, notably Raymond Kurzweil, and there are a few others, firmly believe it this is going to happen in 20 or 30 years.
Steve: What is going to happen?
Zorpette: Okay, that's a very good question. Different people believe different things but I am going to give you a sort of smorgasbord here. The elements of the singularity are: machines become super intelligent. This is kind of the seminal event. What I mean by super intelligent is, far smarter than human beings. So what happens then? Well they're smarter than us, so they kind of don't need us. There is a self-replicating aspect, in other words, okay? They're very [much] smarter than human beings, these machines, and they're going to start reproducing themselves and they won't need us at that point. But it is even better than that because to go from generation to generation in the human world is either nine months or 21 years, depending on who you talk to; you've got to get pregnant; you've got to buy baby clothes, the whole thing; the kid goes to school and then college and a couple of decades later, you've got a new generation. A computer generation could be a week; a computer generation could be two or three days. The point is that you can go from generation to generation and every generation being much smarter than the one before it in a few days. So in a year, you can conceivably have, say a 100 generations, each one, even if the improvements aren't huge, you've got the 100 generations in a year. The machines are stupendously intelligent at this point. The idea of consciousness gets folded in here. And consciousness is kind of different from intelligence, but consciousness is what makes life meaningful for all of us. Consciousness is hard to define and describe, but it is basically, it's what happens in your brain as a result of sensation, memory and other stuff. In other words, you take a bite of watermelon and there is a bunch of chemical stuff that happens in your mouth and nose, but you have this wonderful sensation in [the] taste of watermelon: It is sweet, it's wet, it's got this certain flavor—that's consciousness. It's different from the objective thing that's happening which is chemicals and stuff happening in your tongue. Consciousness is that experience and it could be the sound of an oboe, it could be kissing your girlfriend, you know, it could be watching a movie, this is human consciousness. Now consciousness arises because of these physical and biochemical processes in the brain. Because there is nothing mystical or magical about it, at some point, and because many people believe that it arises somehow as a result of processing which we believe is more or less what the brain does, not necessarily the kind of processing that happens in computers, but processing of some sort, there is this belief that consciousness will arise artificially, when computers are powerful enough and when we get the software right. And there is also a sort of assumption that
in this runaway breakthrough in intelligence , that is going to occur shortly after we build the first smarter-than-human machine, consciousness will fall out of that. How, no one really knows. I mean that is one of the things—it is just sort of widely assumed that it will happen. If you try to pin them down, well, this is one of those things, it's one of those things that people debate, shout at each other.
Horgan: Can I just add a couple of elaborations on the idea of a singularity? Glenn is really focusing in on artificial intelligence and that was what Vinge focused on in his original essay; but over the last 15 years that the basic vision has got more complicated. So one part of it still is that computers are going to become more and more intelligent and then there is going to be this runaway positive feedback factors that keep creating more intelligent versions of themselves; but there is also the possibility that advances in nanotechnology and genetic engineering and neuroscience will allow humans and machine to become hybrid cyborg[s], so through genetic engineering we can enhance our intelligence, we can also put electronic chips in our brain, as this is already happening.
Zorpette: So we won't be completely left in the dust, in other words.
Horgan: Right. So there might be a transitional period where humans are enhanced. We still have our flesh and blood cells, but we soup them up with nanotech brain chips, genetic engineering and these sorts of things. Another possibility that I don't think you mentioned is that we're going to become immortal. So this is a big part of Ray Kurzweil's vision in particular; and in some of his writings, he says that this might happen without artificial intelligence at all. So through genetic engineering or even pharmaceuticals we can extend our lives by hundreds or even possibly thousands of years to become an effective immortal.
Steve: And the holy grail is once we understand the brain well enough, we could actually upload all the contents of our brain when our bodies start to fail and insert those contents into an artificial body and for all intent and purposes achieve immortality.
Zorpette: Right, there are two versions. There is an artificial body which would be a super-sophisticated mobile robot or [of] some kind. But there is another idea that it might be some elaborate virtual paradise. In other words, if you'
are consciousness is just bits, it doesn't matter if they're running around in the...
Steve: So, now you are in the matrix and you know you are leaving the stake and you don't care that your body is actually in storage somewhere.
Steve: So, okay, we've talked about all this for, you know, at least five minutes now. But the whole point of this issue of IEEE Spectrum is to say that all that stuff is garbage.
Zorpette: Yeah, we do call it...
Horgan: Is that the point, Glenn?
Zorpette: Well, we had a lot of problems with a number of the scenarios.
Horgan: It's interesting because this is a high-tech magazine that is deflating the ultimate high-tech vision.
Zorpette: Yeah! Well we thought, "Who better than us?" Really, many of the people who you would identify as singularitarians come from the technology world; and you ably pointed out in your article John, they don't really understand the brain and they don't really understand the brain because nobody really understands the brain, and they tend to underestimate it.
Steve: It is an excellent point; I mean, John, you quote Eric Kandel in your article and Eric Kandel won the Nobel prize for his groundbreaking research into memory and that work was done with a sea slug and basically they have teased out the most basic workings of memory in an invertebrate and these other folks like Kurzweil think that within his lifetime, you're going to be able to understand all the workings of the human brain to the point where you can basically replicate it.
Horgan: The human brain is the most complicated object that science has ever confronted and there is this, sort of, paradox in neuroscience. Now neurosciences is the fastest growing field in science. I think there are 40,000 people who show up at its annual meeting every year now. There are tremendous advances in brain scanning technologies. You've powerful computers that can analyze all the data that come from MRI machines and from research and which you put electrodes into brains; but the paradox is that the goal of creating a unified theory of the brain that explains how a brain makes a mind seems more distant than ever.
Zorpette: And basically when you say a brain makes a mind, [another]
in other way [of] putting that is how a brain gives rise to consciousness.
Horgan: Right, but anything, just, you know, you can leave consciousness—it's such a complicated part of this—you can leave aside consciousness, which brings up all sorts of philosophical issues and just ask how we can carry out a conversation like the one we are having right here. How you recognize a can of Coke as a can of Coke; how I can remember what I did this morning before I got here. All these sorts of things are also profoundly mysterious. So you're right. Eric Kandel has focused on one tiny, little part of the puzzle and there have also been other pieces of research that explain one tiny, little cog in this giant machine that the brain represents. Nobody has a clue how it all fits together and that's what I was trying to do in my contribution to this issue is to remind people how fantastically complicated and mysterious the brain is; and there is nothing mystical about it. It is not as though the brain is some kind of special supernatural substance or that the mind is, that we will never be able to explain; for those sorts of reasons, it is just fantastically complicated.
Steve: Not to mention that this idea that if you uploaded the contents of your brain and put it into some kind of an artificial body, it seems to be so stuck in a Cartesian outlook, [a] philosophical outlook, that there is a mind-body split; which[while it] seems like all of modern biological sciences [are] pointing to the idea that, no, the body is completely integral to consciousness, because we're informed sensorially through our toes—you're your toes as much as your brain.
Horgan: You're right. I think one of the assumptions in artificial intelligence has always been
that, you know, the brain as this kind of digital computer and the mind is a software program and you can just extract that software program from this flesh-and-blood, three-pound mass in our skulls and put it in a laptop or some kind of silicon-based machine; and that's a tremendous assumption, it might be that you can say, maybe, the mind is software but it might be a software that can only run in this particular stuff within which it evolved.
Zorpette: That's like Rodney Brooks said in his article: I do firmly believe that the brain is a machine but whether this machine is a computer is another question.
Steve: You know, for...
Zorpette: But one, before we go too
much far, I should say that there is one other small—maybe not so small—aspect of singularity that we left out. We have pretty much covered the basics—intelligent machines, brain boosting nanotech; the other dimension of this, for completeness' sake, was covered in our issue by Robin Hanson, and that is an economic aspect. Because there is a theory—that is mostly Hanson's theory, but I guess other people subscribe to this—that the limiting factor holding back a capitalistic economy is brain power; at this point nothing else really holds it back, well resources holds it back, but brain power is a big one; and a[the] more advanced an economy becomes, the more brainpower becomes a bottleneck.
Steve: Right and...
Horgan: And when we...
Steve: Brain power and resources are not completely separate because, you know, we got our oil situation. If the brainpower is applied well enough that resource may not be necessary.
Horgan: ...and our resources last longer.
Zorpette: Our resources last longer, yeah! So Hanson's theory is that when you can have much greater than human intelligence and reproduce it, amazingly mass produce it, and
I[o]n an incredibly cheap scale, running of essentially copies of human brains that would essentially eliminate that bottleneck to economic expansion and that would cause runaway economic growth to go along with your other runaway advancement. So in this issue Hanson follows that through to a conclusion coming up with tiny insect-like robots with greater than human level intelligence living by the billions in skyscrapers and sort of doing their virtual work at the equivalent of pennies per day and what this leads to, there are two different ideas about what this kind of economic runaway advancement would ultimately lead to. One is Hanson, more of excited at the idea that there would be this paradise where we would be, where the computers and machines would be our servants and essentially they would just, for whatever benevolent reasons, we would live this life of luxury in this kind of paradise, because there wouldn't be much for us to do because we couldn't compete with these machines; but nevertheless these machines would feel benevolently towards us and we would have it made. The other idea which Hanson, and to some extent Bill Joy, subscribe to is that there kind of wouldn't be any place for us; we would be in the way, we would be stupid and inefficient and, you know, slimy and we really wouldn't be necessary.
Steve: Don't we already have massive quantities of these tiny little entities that do incredible amounts of work on pennies a day? We just call them photosynthetic bacteria.
Zorpette: Yeah, but they're not as smart as these creatures envisioned by Hanson. These creatures envisioned by Hanson would you know have, quadruple digit IQs, and I don't how, one would assume that if they were that smart, they'll tire of living by the billions in matchboxes and skyscraper[s], but...
Horgan: Let me just make a point;
I forget few at Spectrum in 1984, and we did a special issue...
Zorpette: With a bald lady on the cover.
Horgan: ... with a bald lady on the cover. And it was about all these amazing things that technology was going to bring about and I actually ...
Zorpette: We have a bald robot on our current cover.
Horgan: Is that right?
Horgan: Yeah! That issue was filled with predictions about what expert systems would do.
Zorpette: None of them came true.
Horgan: None of them came true. So I edited an article by a guy named Frederick Hayes-Roth—I think he was a big shot in the field of artificial intelligence then—and he said that basically computers were going to become doctors and businessman and lawyers and air traffic controllers and pilots; they're going to take over all the major professions. And he was sort of at a loss to explain what we might be. He thought that humans might be psychotherapists. They would help each other deal with all the problems created by the super-smart machines. None of that stuff has happened. Artificial intelligence has gone through these cycles of optimism and hype followed by disillusion and collapse. That's one of the reasons why I am so skeptical [of some of these]
if somebody claims about the singularity. Obviously, computers have advanced tremendously in certain ways; we've got the internet. I started at Spectrum, I had an IBM Selectric typewriter to write stories. I mean probably a lot of our listeners won't even know what that is. Now we have got these little, skinny laptop computers. So there has been tremendous progress in certain areas. But in other areas like psychiatry or cancer research, fusion energy, remember that that was going to give us energy too cheap to meter and[in] certain areas. And this is what the singularitarians don't seem to recognize: Science bumps up against limits and it makes very little progress, and for some of these visions that the singularity would require, those sorts of things aren't going to happen.
Zorpette: Well, they're bumping up the greatest arguably the greatest limit of all, which is understanding how the brain makes the mind, which we known nothing about—nothing.
Horgan: Or senescence, aging, why our bodies, after a certain period of time, just tend to fail; I mean there is really no sign that that is going to be solved.
Steve: And I became quite convinced by reading Antonio Damasio, neuroscientist, that if you did upload the contents of your brain into this artificial body, you would be mad, you would be literally insane.
Steve: Well I'm glad I'm I threw that [in there].
Horgan: (laughs) Good, that's [a] good contribution Steve!
Steve: So basically the bottom line here is sorry, Ray, you're going to die.
Zorpette: Yeah! Sorry Ray....
Steve: You can check out the IEEE Spectrum issue on the singularity or lack thereof at www.spectrum.ieee.org/singularity.
Now it's time to play TOTALL....... Y BOGUS. Here are four science stories; only three are true. See if you know which story is TOTALL....... Y BOGUS
Story number 1: An ad for a snack food was beamed last week in the direction of a solar system 42 light years away.
Story number 2: As this podcast went to press, a science book was expected to sell for about a million dollars.
Story number 3: You know how when you want to wrap old chocolate, it has whitish specks on it? Well, you don't want to eat that chocolate because those specs are of fungus that turns the chocolate from sweet to sour.
And story number 4: The typical face you make when scared actually provides some survival advantages.
Time is up.
Story number 1 is true. Last week an ad for a snack food was indeed beamed out in the direction of a solar system 42 light-years from Earth. I am not giving that snack food free advertising here on Earth, so I will refer to it as "blank," so the ad for blank was pulsed out for over six hours from high-power radars at the EISCAT European space station in the Arctic Circle. The head of the project said, "We are constantly looking to push the boundaries of advertising." But what we do if the aliens who receive the ad come here and say, "Take us to your dip"? Pretty much the same thing as if they say, "Take us to your leader."
But anyway, story number 2 is true. A collection of rare science books was being auctioned in New York City on June 17th; a 1543 edition of Copernicus is expected to fetch about a billion bucks. Also on the
blog [block] is the first-ever phone book; oddly the book had no numbers. They didn't exist yet, just names and addresses of the people who actually had phones. You have to tell the operator to connect you with whoever you wanted to ask about Prince Albert in a can.
And story number 4 is true. That standard scared face—wide open eyes, raised eyebrows—can help you survive a scary situation. That facial posturing slightly improves eyesight and allows you to increase the amount of air you bring into your lungs. For more, check out the June 16th episode of the daily SciAm podcast, 60-Second Science.
All of which means that story number 3, about the white specs on old chocolate being a sour tasting fungus, is TOTALL....... Y BOGUS, because what is true is that the white specs are merely the fat in the chocolate recrystallized after melting. Researchers reported the finding in a journal called Soft Matter published by the Royal Society of Chemistry. The white dots are actually called fat blooms and the chocolate is still perfectly safe to eat.
The [A] quick note: On the May 21st podcast, I made an offhand comment about how all the treadmills and Stairmasters in health clubs should be wired to generate electricity to solve our energy needs.Apparently, a writer in The New York Times made a similar suggestion in early June and on June 12th, the Times published a letter by a physicist that noted that eight million New Yorkers on treadmills could produce about 600 million watts an hour; but we couldn't do it more than an hour or two a day and we consumed 3.4 billion watts a day. So clearly what we need is solar-powered treadmills.
Well that's it for this edition of the weekly SciAm podcast. Visit www.SciAm.com for the latest science news, blogs and all our podcasts. And sign up for the daily digest at www.sciam.com/daily. For Science Talk, the weekly podcast of Scientific American, I'm Steve Mirsky. Thanks for clicking on us.