Future Imperfect
Posted by Jeriaska on January 4th, 2008
In the not too distant future you may be able to buy an inexpensive video camera with the size and aerodynamic characteristics of a mosquito. Even earlier, we will see the proliferation of cameras on lamp posts designed to deter crime. Ultimately, this could lead to a society where nothing is private. At the 2007 Foresight Vision Weekend, Professor David D. Friedman gave an overview of three destabilizing technologies that could usher in a world of ubiquitous surveillance in a preview of his upcoming book Future Imperfect.
The following transcription of the Foresight Vision Weekend presentation by David D. Friedman entitled “Future Imperfect” has not been approved by the speaker. An audio recording is also available.
Future Imperfect
My most recent book, which ought to be out in the not too distant future, meaning a year or so, is called Future Imperfect. It deals with a whole bunch of different technologies, each of which might revolutionize the world in the next few decades. For almost any audience other than this, I can pick topics at random and assume that they will be new and exciting to everybody in the audience. I don’t think that would work here, which is a good thing, because although I do tend to speak too fast, managing a 114,000 word manuscript in forty minutes might be a bit of a stretch. If people are curious, my webpage is www.daviddfriedman.com and an almost final draft of Future Imperfect is linked to that webpage. If you have comments, please get them in fast.
What I decided to do instead was to take three chunks of what I discuss in that book that I think ought to be of interest to many of you and hopefully will be novel to many, perhaps most, of you. The first one involves privacy, and it involves two technologies, one of which could give us more privacy than humans have ever had, and one of which could give us less privacy than humans have ever had. The question is what happens if we get both. It turns out that the net effect, I’m willing to argue, depends on yet a third technology.
The second subject I want to discuss is the argument against government regulation of nanotechnology. This part of the talk is dedicated to Ralph Merkle, for reasons that will become I hope obvious in a little while. The third is on a very general but very important issue, which is, Can technological progress make us worse off? And if so, how? It sort of seems at first glance as though technological progress means we can do the things we want to do better. It ought to be unambiguously good. A lot of people worry it could be bad. I’m going to say they could be right, not that they are right, but that there are reasons that they could be right.
Let me start with the first bit, and the first bit is privacy, of course. The two technologies are public key encryption and surveillance. Public key encryption I expect many of you are familiar with. It provides a way in which it is possible for people to communicate with reasonable confidence that nobody else is listening. It also, and perhaps equally important, provides a way in which I can establish my online identity without giving away my realspace identity. By using digital signatures I can establish an online persona, prove in the future that I am that online persona, while the link between the online persona and my physical body is controlled by me. This was a point that was made about online things I think even before the invention of public key encryption, though I won’t swear to that, by Vernor Vinge in the novel True Names, a very interesting and insightful work, like much of Vernor’s work.
In addition, the same technology makes possible, though unfortunately we don’t have it yet, fully anonymous digital cash. That is a way in which I can transfer value to you by sending you a message without my having to know who you are, your having to know who I am, or the bank holding the value having to know who either of us is. Again, as we only have a limited amount of time, I am not going into the details. Half of you know them already, and the other half are welcome to read the more detailed explanation both in Future Imperfect and in other things I’ve got on my webpage.
So, what does that set of technology mean? It means in particular, a world where you can combine anonymity with reputation, which sounds like a very odd idea. But if I have an online persona that online persona can establish its identity, it can build a reputation, and yet it is anonymous in that the physical body that people can actually shoot or lock up is linked to that in a way that only I control.
I’m not sure it will happen. I’m not sure any of these will happen. As I say in the beginning of the book, any of the technologies might turn out to be a wet firecracker. But there are enough things with radical effects so that it is quite likely that the future will not be much like the past. What can you get with these technologies? You get freedom of speech that no longer depends on the current views of the Supreme Court, because they can’t arrest you if they don’t know who you are. But you also get criminal firms with brand names and reputation. I’m not going to run through here my business plan for Murder Inc. Again, I think it’s there. It’s certainly in other things I’ve published, but I think it would be fully practical to have a firm selling the services of hitmen to ordinary customers, not the insiders in the criminal marketplace, but to people like you and me who feel this is a service we might want to occasionally purchase. And that’s somewhat disturbing.
You could think of a variety of other ways in which the technology does indeed promote privacy, but privacy is a double-edged sword. It can be used for both good and bad purposes. The other technology, the one that sort of goes along with this, is surveillance. Many of you have probably read David Brin‘s book The Transparent Society, the first chapter of which is very, very good. It raises a lot of interesting and important points.
The basic argument is that there are plausible reasons why people want to put video cameras up on poles in public places, because this looks like a way of deterring crime. There are some arguments about how well it works in practice, but it’s a pretty obvious strategy. By itself, it isn’t all that dangerous to privacy. After all, there could be a policeman standing in a public place, too. The problem arises when you add two more technologies. One of them is face recognition and one of them is databases. Once you have your video hooked to a good computer, it identifies who the people are, that all goes into a database, and the judge instead of asking the defendant “Where we you at 10:00 AM on January 17th,” pushes a few buttons and sees the video of what the defendant was doing at 10:00 AM on January 17th. So that has the potential to give a world where everything done in public places is not really recordable, but findable by person, which is a level of privacy a little bit lower then primitive societies with crowded conditions and no walls had. Low privacy societies are historically not uncommon, but this is an extreme case.
Carry it a little bit further and one could imagine video cameras either belonging to police or to private people with the size and aerodynamic characteristics of a mosquito, at which point the condition in public spaces more or less goes away because pretty much everything is a public space. So that’s a possible technology. Again, I’m not saying it will happen. You could think of various conflicts, which gives you instead of more privacy than we have ever had, less privacy. Brin’s solution to this problem is to say we can’t stop this, so let’s have universal transparency. The cops can watch us, but we can watch the cops.
There is something to be said for that. There have been a number of high profile cases where cops got into trouble by doing things with video cameras pointed at them, but there is a certain asymmetry in this situation, because they can arrest us and we can’t arrest them. So given differences in ability to use force against people, universal transparency maybe is better than one-sided transparency, but it’s not really a very satisfactory solution.
What if we get both? That’s the fun question. What if we have both a world with a well distributed public key infrastructure, everyone has a private and public key, a digital signature, anonymity online, all that stuff, and at the same time there are these video mosquitoes buzzing around and everything you do is watched.
What the sum result is depends on two different questions. One of them is whether the interface between realspace and cyberspace is in realspace or in cyberspace. That is to say, it does me no good to have the strongest encryption in the world if there is a video mosquito watching me type. You could imagine having a hood that you type under. That would sort of be a low-tech solution to this, and hopefully it’s got good mosquito netting, or electronic interference to keep the video mosquito from broadcasting. You can think of alternative arms races going either way.
If you want a higher tech idea, you can imagine that we’ve got a mind-to-machine link of some sort. There are things in that direction beginning to become possible. Therefore, hopefully, the mosquito can’t see anything. One question is, Who controls that link? You could imagine at one extreme if the link is entirely in realspace, meaning that it can be observed by your surveillance system, then your cyberspace privacy is useless. On the other hand, if it’s really easy to keep the link secret, then when we have a conversation we don’t do it by this crude old fashioned method of vibrating air in between us, because that’s observable. The video mosquito’s got ears as well as eyes. I am in some sense talking inside my head, or someplace inside my private sphere. The information is being encrypted by my little computer, maybe it’s embedded in me or something, and trasnsmitted by bluetooth to you. Your computer decrypts it. It is projected by a little laser beam on your retina. We hope the inside of your eyeball is still private. You could then undercut a good deal, not all (you could still see me if I shoot somebody), but a good deal of the surveillance.
The second question, however, is the relative importance of cyberspace and realspace. That ultimately depends on a third and very important technology, the one which I argue in the book is the first one I listed as actually appearing as an important player. And that is virtual reality. Those of you who spend hours in World of Warcraft realize that virtual reality is here now. Ten million customers for that product I think in the U.S., I’m not sure. Something like a hundred million customers for all such products in the world.
However, a world with really good virtual reality is a world where your body is lying in a convenient locker somewhere with maybe some mechanical devices to exercise your muscles and some goop that you eat, but it doesn’t matter because it all tastes like sushi and fillet mignon. Tasting is an information transaction, not a physical transaction. At one extreme you can imagine a world where the video mosquitos are really bored because nothing interesting is happening in realspace. It’s all happening in cyberspace. That’s the limiting case with really good VR. That seems to be the interesting tension between these three technologies. I’m not promising we’ll get all three of them. It would be fun to see what happens if we do, but it’s an interesting issue that I wanted to bring up.
My next topic, this is the part of the topic that’s dedicated to Ralph Merkle. Ralph is a clever fellow. This book came out of a seminar I taught at the University of Santa Clara Law School called legal issues of the 21st century. Ralph consented to be a guest lecturer. I’m not sure if it was only on nanotech or if I got him on cryonics, too. Whenever Ralph talks to my students about nanotech, he talks about how dangerous nanotech is and why we need government regulation of nanotech. I have a suspicious mind, and it occurred to me that this might be strategic behavior on his part, because he knows what my political prejudices are, and if you want someone to invent arguments for you on one side of a controversy the obvious way is to take the argument on the other side of the controversy. This was my sneaking suspicion and that’s why my discussion of why we shouldn’t have government regulation of nanotech is dedicated to Ralph Merkle.
The argument for government regulation is obvious, and sufficiently compelling that even people at Foresight, who generally tend to be libertarian, have some sympathy for it. Nanotech makes possible extraordinarily deadly weapons that might be produced either deliberately or by accident, and therefore you need control. That control, obviously, is going to need to be enforced by the government. That’s who controls dangerous activities. I start my book by saying, “Look, all this global warming stuff might be right, it might be wrong, but it’s small potatoes. I’ve got three different technologies here that can wipe out our species before the end of the century.” Nanotech is one of them. Guess what the other two are.
The fundamental problem is that the solution to the perils of nanotech has this odd air of setting the fox to guard the hen house. In the world as we know it, with technologies we observe, there are generally both offensive and defensive technologies. That is typically true. What is the sole human institution that spends very large sums of money on offensive technology? It’s government, of course. Ain’t nobody else in that business. Rugers spends small sums of money designing things that can be used to kill people one at a time, but for the large scale project it’s all government.
It seems to me reasonable to suppose that in this technology, the more resources you spend on either an offensive or defensive technology, the more effective you will be. On average, if you spend much more on offense, probably it will win out. If you spend a lot more on defense, maybe it will win out. It may not be even, it may not take the same number of dollars, but you would expect at least that generally spending more resources is going to get you better, more powerful technologies. ]
Defensive technologies you can sell. If we live in a world where people are worried about diseases, you can indeed sell vaccines. If we live in a world where people are worried about burglars, we can sell burglar alarms. Well, if we live in a world where as nanotech develops it is clear that there are dangers, and the small dangers are probably going to come before the grey goo scale dangers, then there is going to be a market incentive for people to try to develop ways of defending against this technology.
In a world where nanotech is in effect a government monopoly, or a government regulated field where the government having a great deal of control over what everybody else can do, you would expect that there will be governments designing offensive nanotech and we will be seriously handicapping the private efforts to develop defensive nanotech. That does not strike me as an attractive solution to the problem. I think it is useful to think in terms of two different worlds, which are really the same world a decade or two apart. The early world is the one where it takes billions of dollars to do anything. We haven’t gotten even to that world yet.
I assume everybody in this room realizes that when I say “nanotech” I mean Drexlerian nanotech. I mean building machines, not just one-molecule thick layers of stuff. That’s nanotechnology, too, but a much less exciting one. I’m talking about the stuff that Eric‘s been pushing for a very long time. In the early stage, it’s going to be very, very hard to do it. It will cost billions of dollars. Firms that spend billions of dollars developing things have pretty strong incentives not to do things that result in killing lots of people, getting them sued, and making them unpopular.
If you look out on our experience in the past, chemical firms, refiners, airlines (not that no accidents ever occur) but they spend an awful lot to make sure that their planes don’t crash. When things go wrong there’s this odd pattern that you have this terrible Bhopal accident and it at least partly seems to have been due to the fact that the government of India was restricting the precautions that the chemical company could take, because there were lots of people who were voters who wanted to live close to the plant and keeping them away was unpopular. We had a second rather large accident which did a good deal of damage. That, of course, was Chernobyl. That was not a private firm running that reactor. You have lots of reactors running by private firms and we had one hiccup at Three Mile Island, and no disasters.
You can never be certain, but I think that in the early stage where it’s huge-scale things, you can reasonably expect people doing them are going to take precautions, if they are allowed to. But you can also expect that one of the really profitable products, when we can do it, is going to be those nice cell repair machines that I think Nanomedicine and other places have discussed. Wouldn’t it be nice to have a cure for all that ails you, including of course aging? That’s what a cell repair machine offers. Even before that I sort of want those little microscopic compressed air tanks as extra red blood cells, so that when I have my heart attack I can call up the doc, and he can make an appointment in an hour or two, and I can come in and get the problem solved. It turns out red blood cells just don’t store very much oxygen compared to what a higher tech version of the same size could do.
Now, let’s think about the late version. The late version is the scary one. That’s the one where the kid in his basement can end the world. That is where nanotech is readily available, and of course, you would expect that to come eventually, because one of the things that assemblers can do is to make assemblers. Therefore, eventually it gets cheap, and if you’ve got lots of computing power you can figure out how to design stuff. As you may have noticed, computing power keeps increasingly getting cheaper. That’s the stage in which we are worried about the internet worm, about computer viruses, only these are realspace versions of those things, and we happen to live in realspace and are therefore vulnerable to it.
It’s going to be extremely hard to prevent that. It’s going to be hard to prevent that with or without government regulations. Once you’ve got that world, it’s like the problems we see with preventing file sharing and stuff like that. Once it’s really easy to do, lots of people will do it. It is especially difficult to do in a world of many different governments with different agendas and different abilities to enforce. So I think you really have to take it for granted that in that world if grey goo can be made, grey goo will be made. If the nanotech equivalent of the virus that kills everybody who is over 25 can be made, some 24 year-old will try to make it, along with the antidote that wipes it out before he reaches 25. You can think of a lot of things of that sort.
What that means, if we are going to survive in that world, we’re going to survive by defensive technologies. We are going to survive not by the fact that nobody can make anything deadly, but that when the grey goo starts to propagate 73 different varieties of anti-grey goo eat it. When the deadly virus comes, my cell repair miniature submarines say there’s something here that shouldn’t be and eliminate it.
So, we are now talking about a war between offensive and defensive technologies. As long as there isn’t an inherent big advantage to the offense, which I’ll come back to in a minute, the defense wins. The reason the defense wins is that a kid designing a virus in his basement doesn’t spend the sort of resources that Apple and Macintosh spend on their operating systems. Similarly, the kind of offensive stuff, we have to worry a bit about governments. Government regulation doesn’t solve the government problem. But as far as private things are concerned, we will have cheap offensive things against very effective defenses.
We might still lose. It might turn out that with this technology, offense is much stronger than defense. On the other hand, we have run the experiment already in carbon. We observe that both bacteria and whales survive. We therefore observe that large, complicated nanotech devices like us can indeed fight at least even against small, simple nanotech devices like viruses and bacteria that are trying to turn the world into copies of them. After all, that’s what life is. Grey goo is not a new idea, right? We’re gray goo, too. We don’t make exact copies, we try to improve with each generation. I’ll avoid telling you how wonderful my children are. But, basically, living creatures are trying to turn the environment into copies of themselves. That’s just what we worry about nanotech stuff doing. In that struggle, in the limited subset of devices that evolution produces, we observe that indeed the defensive technologies seem to be able to make a half-way decent stand against the offensive.
One advantage that we have is that there is a certain sense in which the offensive technologies are not really offensive. The virus does not want to kill you. It’s a parasite, it likes to have its hosts around. It just wants to duplicate itself, and sometimes kills you in the process, which is what you would expect from nanotech as well.
That to me suggests that perhaps government regulation is not the right answer. When I was thinking about this it occurred to me that there is an example of a similar problem that we have already seen. We are really talking about centralized versus decentralized solutions to these perils. Generally most of us like decentralized solutions. Teenagers who can end the world in their basement may make us rethink that question.
It occurred to me there has only been one time in my life when I have been seriously worried about the survival of myself and my family. That was a little bit after 9-11 when somebody raised the issue of smallpox. They pointed out that there were a few smallpox cultures in a few government labs, that it might be possible to excavate arctic graves with people who had died from smallpox, and maybe some of it would still be there. Well, I said, terrorism is sort of scary, and bioterrorism is sort of scarier than crashing airplanes into big buildings. I better get everyone in the family vaccinated. And I discovered that I couldn’t. Why couldn’t I? Because vaccination is an activity controlled by the government, and the government had decided we didn’t need the vaccines anymore because everybody knew that smallpox had been wiped out, and the result as far as I can tell was that the existing stores of vaccines were enough for the hospital workers, the president, and other vital people, but not enough to cover any substantial part of the population, and not enough so that I could just go out on the market and arrange for vaccinations.
Fortunately, we’re still around, but it did suggest that centralized solutions can be rather hazardous, because they make one decision and one person’s mistakes can lead to catastrophic results. Whereas, if vaccination is just an ordinary, private activity and the government has nothing to do with it, I will give you odds that there would have been a speculator somewhere with a big chest full of vaccine kept at appropriately cold temperatures just so he could make a killing saving lives, if there happened to be a resurgence of smallpox.
Now let me get to my final point, which is the general issue of whether technological progress can be banned. This requires a brief digression to talk about the impossible world we already live in. The problem is what economists call the coordination problem. In order to do anything at all in a complicated society, you have to have some millions of people all doing the right things to go along. The classic example is the pencil. You hold up a pencil and you say, this needs wood. What does wood need? Somebody to grow forests, but also somebody to make chainsaws. Chainsaws require steel. Steel requires blast furnaces. Blast furnaces require gas or coal. Gas or coal requires mining or drilling technology. Meanwhile, we also need gasoline for those chainsaws. And that’s just the wood. We haven’t even gotten into the eraser and the metal band that connects the two, not to mention the graphite that’s called “lead.”
If you think about it for awhile, you realize that this is an enormously interdependent system where everybody depends on everyone else. In order for it all to work, we have to get some way for people to write orders. We have got to have some way of making sure that when I want to make steel, someone else is mining the necessary ore, and so forth, and so on.
There are really only two solutions to this problem. One of them doesn’t work. The one that doesn’t work is centralized control, and “doesn’t work” is unfair. It works on very small scales. It works for a football team, a small firm, maybe a family… though I can’t say it works very well for my family. We’ve never tried it. We’re strong on the decentralized method. It doesn’t scale. As you increase the number of people involved, you’re basically funneling information through a very small bottleneck of whoever is controlling it, and you run into incentive problems with people who control and have their own interests.
The solution that works is the decentralized solution. You can think of that, at least if you are an economist, as a sort of simple property rights/markets/trade version. That’s not the only version. I’m sure that there were markets in the Soviet Union, even if it was sort of influence and favors. Some version of the decentralized mechanism is the one that scales up with larger sizes.
However, in order for it to work, there has got to be some way of dividing up the world such that most of what I do affects only my piece of it. If you think of physical divisions in geography, it doesn’t matter to you if I’m playing loud music as long as your house is at least a hundred yards away from my house. It doesn’t matter to you if I’m watching porn, you can’t see it from your house. You could think of all sorts of ways in which physical geographical divisions ideally let you say what I do affects me, what you do affects you. If you want me to do something for you, we can talk about it. But as sort of a first approximation, as it were, we can optimize for everybody by having each person optimize for himself.
That’s a very oversimplified explanation of price theory. There’s a lot more out there. I have a complete price theory textbook webbed on my site for free if you want to read it. Or you can buy a shorter version, which was a textbook turned into a book for the intelligent layman. It’s called The Hidden Order, and Amazon has it. But it depends on your being able to do that. Whether you can do that depends on the range of the effects of human action. If lots of the things I do have effects for a thousand miles around, it gets real hard to set up any system of legal rules in which I have the right incentives. We use tort law as a sort of a second best. If what I do affects me, and it occasionally affects you and you can sue me, it’s not nearly as good as markets but it’s better than nothing. But if what I do, 10% of the effect is me and 90% is spread around 60 million people and we’re not quite sure who they are, now we’re down from one solution to zero solutions. You now have a system where the decentralized solution works badly, the centralized solution works badly, and the answer is we have a problem we haven’t solved.
I want to take as my example global warming. I’m not interested in the question of whether anthropogenic global warming is true or whether it’s a really serious problem. My own opinion is that it probably is true and that is probably is not a very serious problem. But that doesn’t matter. For the purposes of the next five minutes, assume that it is both true and catastrophic. My point is, there isn’t a solution. Why isn’t there a solution? Because at the individual level, when I produce carbon dioxide, I get a negligible fraction of the cost of the resulting global warming. I might even live in Minnesota or Northern Canada and benefit by it. Therefore I’m not going to control it.
Well, what about tort law? Right. I can just imagine the whole population of let’s say Bangladesh trying to sue everybody who drives a car in America. Well, government regulatory solutions. That’s getting even worse. That’s sliding us back toward that centralized method that does not work very well. It’s unlikely to do the right things. But even that won’t solve it. Not only is global warming a problem at the individual level, but it’s a public good problem at the national level, because the carbon dioxide we put in the air, assuming that global warming is a serious problem, is mostly going to affect people in Bangladesh and in tropical and low lying areas where getting hotter is bad instead of good.
Now, it’s true, global warming provides a useful excuse for raising taxes. It might be used for that purpose. It’s not likely to actually stop global warming. There is a solution. The solution, of course, is world government. But this is what we like to describe as a cure that’s worse than the disease. It seems to me that that demonstrates the reason why potentially technological advancements could have problems. One of the things that technological improvement does is increase our power. One of the results of that can be that the range of the effects of what we do becomes larger and larger. We weren’t capable of creating global warming 200 years ago. We just didn’t have large enough scale activities. Similarly, of course, for nanotech, viruses, and that sort of thing, if you increase the range over which people do things, it is at least possible that you will make it harder and harder for the one solution.
Now, it isn’t in any sense inevitable. Some technological improvement cuts the other way. If my nanites are patrolling the border of my territory, turning everything noxious into something that they have some use for, I don’t have to worry about your air pollution. But if an atomic bomb goes off, probably not even good nanotech will save me. Short of that, you can imagine technologies which result in reducing these problems, and I could give other examples, but you could also imagine technologies that result in making things much worse. That is my short explanation of why it is possible that technological improvements could make us worse off as well as better off. There are possibly other reasons. I discuss a whole bunch of other technologies in the book that aren’t here, but I think that I’ve at least reached the end of my talk.
