Interesting Times

May you live in interesting times.

Ancient Chinese curse, apparently invented by Eric Frank Russell c. 1950


A writer looking forward from 1900 might have anticipated rockets. He might have anticipated nuclear explosives. The nuclear balance of terror, one of the central facts of the second half of the century, required both. Through most of this book, I have taken futures one at a time. They will not come that way.

One interaction among technologies was discussed in Chapter 5. If cyberspace is private and realspace public, how much privacy we have depends on how much of our lives is lived in each. That in turn depends on another technology: virtual reality. In the limit of deep VR, everything important is happening in cyberspace, leaving the automated cameras of the transparent society very little to watch.

Another example appeared in Chapter 21. How large a role space plays in our lives over the next century depends on how expensive it is to get there. That, in turn, depends on the strength-to-weight ratio of the materials available to us. With sufficiently strong and light materials, it becomes possible to build a space elevator, drastically reducing the cost of getting off earth. Short of that, better materials make possible launch vehicles with a much higher payload and much lower costs. One way of getting very strong and light materials, such as single molecule carbon fibers, is nanotechnology.

In some cases, one technology eliminates problems raised by another. Genetic testing makes genetic risks uninsurable. But with sufficiently advanced genetic engineering that does not matter, since there will be no genetic risks left to insure. Biometric identification technologies can impose on everyone on earth a built-in, unforgeable, ID card – until nanotechnology makes it possible, even easy, to revise your fingerprints or the pattern of blood vessels in your retinas.1 Cryonic suspension raises puzzles connected with adequately punishing criminals who choose to get frozen while their sentence runs, but if surveillance technology produces a world where criminals face a near certainty of conviction, adequate punishment may not be a problem. If we know enough about how the human brain works to emulate it in silicon, we may know enough to rehabilitate criminals by methods less crude than the penitentiary. Of course, that knowledge, and that power, could create other problems compared to which problems of crime and punishment may be insignificant.

 Future Imperfect

In the past twenty-one chapters we have considered a wide variety of possible futures, some attractive, some frightening, few dull. Most offer both problems and promises. By avoiding the former so far as possible and taking advantage of the latter, we can in most cases make ourselves better off than if the technology did not exist. Consider a few examples.

Encryption, ecash, and widely available computer networks will make certain crimes easier; make it easier, for example, to collect the payoff from kidnapping or extortion without being caught in the process. But those technologies also provide new and powerful ways of protecting ourselves from crime. They also make it much harder for governments to control people. Governments, on the historical evidence, are a great deal more dangerous than private criminals are; over the past century, according to one estimate, governments killed more than 200 million of the people they ruled, wars excluded.2 In my view, at least, the benefits of weakening the power of governments are greater than the costs.

Human reproductive technology, the ability of parents to choose which of the children they might have they do have or to use genetic engineering to provide their children characteristics that nature failed to give them, raises potential problems, since important decisions will be made for people before they are born, necessarily by other people. But those decisions are no different in kind from decisions already being made for children by other people, such as the decision to bring a child into existence and the decision of how to rear him or her.

A government with a sufficiently long time horizon could use such technologies to try to breed superior warriors, scientists, or bureaucrats. But such a government could do the same thing using techniques of selective breeding that we have been applying to other species for several thousand years. Few or none have, perhaps because governments rarely have that long a time horizon.3

These technologies may also give us the ability, within one or two generations, to eliminate both genetic diseases and a wide range of other hereditary disadvantages, such as the bad heart that killed my paternal grandfather at a young age and, at a much more advanced age thanks to modern medicine, my father. They also carry the potential to increase the average intelligence of our species, which might be an improvement and will surely be interesting.

The implications of other technologies are more ambiguous. Obvious examples are nanotechnology and artificial intelligence. Either might lead to an enormously more attractive world. Both contain the potential for catastrophe on a scale that has not been seen on earth at least since the elimination of the dinosaurs some sixty-five million years ago.

Technological progress means that we can do more things. We do things because we think doing them benefits us, hence one might expect that we would always be better off with new technologies than without them. If there are exceptions to that conclusion, why? Is there some logic by which, in some possible futures, technological progress could make us worse off?

There is. The first step toward understanding it is to think a little more carefully about the impossible world in which we now live, how it works, and how it might stop working.


Our world is inhabited by a very large number of individuals. Each has his own objectives, beliefs, and abilities. For any moderately complicated society to function, those individuals have to find some way of coordinating their efforts. In order for me to build a sword or a plough blade, someone has to smelt iron. In order for him to smelt iron, someone has to mine iron ore and produce charcoal. Each of those in turn requires inputs produced by other activities carried out by other people. And once we have a plough blade, it is of little use unless there is a farmer to plow the land, seeds to sow, and much else.

In a modern society the problem is even harder. For an example not original with me,4 consider a pencil. The wood it is made from requires, tracing the chain back, trees and sawmills and chainsaws and gasoline and batteries and steel and electric generators and blast furnaces and iron ore and coal and …. Follow out all the lines and you have millions of people coordinating their activities to produce a pencil, a shirt, a computer.

How can we do it? How can we make sure that the mine produces the amount of ore needed to make the amount of steel needed to …? How can we solve that problem a millionfold, when every piece depends on every other piece?

There are only two known solutions, and one of them doesn’t work. That, the obvious one, is central control. Someone has the job of figuring out what everyone else should do, telling them to do it, and making sure they obey. On the scale of a family, a firm, a football team it may come close to working, although even there a careful examination is likely to find people doing quite a lot of things they have not been told to do. But the centralized solution does not scale. As the enterprise gets bigger, more and more information has to funnel through the central authority, much of it getting lost on the way, and the problem of figuring out what everyone should do becomes unworkably complicated.

The solution that does work – most important, that scales – is the decentralized one. Everything belongs to someone. People are free to trade. If the value of what you own – including, most importantly, the use of your body to do things you know how to do – is worth more to me than to you, there is some offer I can make that you will accept. Each of us has different ends. But each can offer, through an exchange of goods, services, or money, to help achieve another person’s ends in exchange for that person’s help with his. Working through the full logic of that system requires a semester or so of price theory, but the basic logic is fairly simple.5

The decentralized solution works best if there is some definition of property rights, some way of dividing up the world into pieces, such that each individual’s use of his piece has significant effects only on himself and a few other people – ideally, effects that occur only with the mutual consent of both parties. My use of my ability to tell stories affects only myself and those who choose to listen to them; if you do not offer me acceptable terms, I will not choose to tell stories to you. Similarly for most goods or services. As long as that condition is met, each of us can decide what to do in terms of its value to him – I like telling stories – and its value to other people, measured by what they are willing to offer him to do it.

Unfortunately, in any reasonably complicated world, there is no definition of property rights that entirely meets this condition. My storytelling is a voluntary interaction between myself and my willing audience, but when I get to the loud parts it becomes an involuntary interaction between myself and my neighbor. My driving my car is the result of voluntary trades with the people who made the car, refined the gasoline, sold the maps. But it imposes costs involuntarily on people I might run into and people downwind breathing my exhaust.6

In the simpler cases, when such external costs become significant, we can and do deal with them through tort law. If I make too much noise too late at night, my neighbor may be able to get an injunction. If I drive negligently and my car ends up in your living room I will owe you payment for damage done. Litigation is a clumsier and less efficient mechanism than trade; on average, of every dollar spent by a defendant only about fifty cents ends up with a plaintiff, the rest going to lawyers, court costs, and the like.7 But it does provide a mechanism for forcing individuals to take account of costs they impose on others when deciding what to do with their property.

As external effects become more dispersed – not a large injury to one person whose house I drove through but a tiny injury to each of ten million people with the bad luck to be breathing downwind of me – tort law becomes less and less useful. The damage from any particular polluter is in most cases difficult or impossible to measure. Although legal mechanisms such as class actions designed to combine large numbers of victims into one for purposes of litigation exist, they work very poorly. In such cases the typical response is either to put up with the problem or to try to solve it by government regulation.

Government regulation is a reversion to the centralized solution, someone at the top deciding what everyone should do and making them do it. For reasons suggested earlier and explored in greater detail elsewhere, by myself and others,8 it works very badly, especially when applied to large societies. We have no good way of making regulatory agencies try to act in our interest and they have no good way of figuring out how to do it.

When the scale of the effect becomes international, the problem gets worse. Regulatory agencies of the U.S. government, such the FDA or the FCC, may be politically motivated, incompetent, or both, but at least they have some interest in doing things that U.S. citizens approve of. They have very little reason to care about the effect of their policies on the citizens of Bangladesh or the Maldive Islands.

For an example of the sort of problem that neither decentralized nor centralized solutions can be trusted to solve, consider global warming due to human activities. While there is quite a lot of legitimate disagreement over its likely scale and consequences, there is good reason to believe that it exists and that it is at least in part due to increases in the amount of carbon dioxide in the atmosphere.9

When I use electricity generated by the burning of coal, or light a fire in my fireplace, or breathe, I am producing carbon dioxide (CO2). One result may be a tiny increase in the temperature of the earth. One result of that may be a very small increase in sea level. Predicted effects, even after combining the results of everyone’s activity, are not very large – current estimates10 suggest a temperature increase of about two degrees centigrade and a rise of a foot or two over the next century. Where I live, on the U.S. West Coast, that will be a very minor inconvenience; beaches will get a little narrower. It will be more of a problem for people living in low-lying territories routinely threatened by floods. And if it turns out that the rate of global warming and sea-level change are substantially larger than present estimates suggest, it could be a very serious problem indeed.

How might we deal with that problem? We cannot very well define property rights in a way that gives every inhabitant of Bangladesh a veto over the power use, home heating, and breathing of every American. We cannot expect the international legal system to honor and enforce tort claims by a citizen of Bangladesh against a citizen of the United States for contributing one part in ten billion to a flood fifty years in the future. We could imagine the government of Bangladesh trying to use international law to force the U.S. government to severely regulate activities of its citizens that are predicted to increase the risk of future floods in Bangladesh but, if the regulations are costly for Americans, why would the American government comply?

We can imagine, indeed have observed, attempts to negotiate international treaties for the same purpose. But while arguments about global warming may sometimes be useful to help governments persuade their citizens to put up with actions that the governments already favor – additional energy taxes, say, that are desired as a source of revenue and can be defended as ways of holding down energy consumption and so CO2 production – it is hard to see their doing much more than that. The United States, in controlling its citizens’ activities to avoid costs imposed on the citizens of the rest of the world, is in the same situation I am in keeping my voice down when it bothers my neighbors. Short of a world government, there is no legal mechanism analogous to tort law to make them do it. And a world government, in my view at least, is a cure worse than the disease.

 The Dark Side of the Force

I have discussed the case of global warming not because I think it is a particularly serious threat – for reasons I will return to shortly I think it is probably not, at least over the next fifty years or so – but because it is a problem that most of my readers will be familiar with and whose nature is fairly easy to explain.

The general problem of which it is one example is the breakdown of the conditions that make possible the decentralized solution to the coordination problem. Technological progress increases our ability to do things. Often, although not always, that means increasing the scale and range of the effects of human action. As scale and range increase, it becomes harder and harder to define property rights in a way that satisfies the requirements of decentralized coordination, a way in which the effects of my actions are mostly confined to my property and the property of those who have agreed to permit those actions. The result is to push us away from the one workable system for coordinating human action – private property and trade – toward the less workable alternatives of tort law and regulation.

As the scale of effects expands beyond the boundaries of the individual nation, regulation by national governments becomes even less capable of dealing with the problem. We are left with the unattractive choice of either putting up with whatever problems result from individuals ignoring the distant and dispersed costs imposed by their actions or creating a world government. If we choose the latter, we find ourselves trying to use a centralized regulatory mechanism to deal with problems far beyond the scale for which such mechanisms are workable.

Not all technological progress raises these issues. Encryption and virtual reality make possible societies that come closer to the market ideal than what we now have, since the interactions they enable are entirely voluntary. In virtual reality, the problem of trespass, the more general problem of conflicting use, vanish; we can each enjoy his own version of California’s Pacific coast, one unspoiled and natural, one with all the benefits of modern commercial culture. The same may turn out to be true of other technologies as well.

Nanotechnology is an interesting case because its effects go both ways. On the one hand, the creator of a self-reproducing nanite successfully designed to turn the entire biosphere, us included, into copies of itself imposes external costs vastly larger than any plausible level of global warming. On the other hand, good enough defensive nanotechnology could eliminate a wide range of current externality problems. I do not have to worry about what my neighbor puts in his air if my boundary is patrolled by molecular machines capable of disassembling any noxious gases that waft across. I do not even have to worry very much about radiation sleeting through my body from his experimental nuclear reactor if my body is patrolled by microscopic cell repair machines capable of repairing any damage in real time. I still have to worry about what happens if the reactor explodes – even nanotech machinery has its limits – but not about much short of that.

I conclude that there is a general problem that we can expect to be produced by some forms of technological progress. It occurs when human abilities change in ways that make it harder to define property rights in a workable fashion, a way that lets each of us go about his own business without having to worry too much about distant effects on dispersed and anonymous people. We can expect some technologies to change human abilities in that direction, others to have the opposite effect.

  Growing Pains

The problems I have been discussing are produced by a technology in place. Other sorts of problems are associated with getting there. One example is the problem of inverting the hierarchy of age and expertise. In a slowly changing society, which is to say in almost all of human history, older people, while they may not run as fast or even think as fast as younger people, know more. So it makes sense to have institutional structures in which, on average, older people have authority over younger people.

As the rate of change increases, so does the rate at which knowledge depreciates. The head of the research department knows much more about vacuum tubes than the young engineers whose work she supervises, but they are not researching vacuum tubes. The judge in a software patent infringement case knows a great deal about patent law, but he knows much less about software than either plaintiff or defendant. To some extent he can lean on the knowledge of other people, such as his law clerks, or make himself a temporary expert with the help of briefs provided by both sides; the problem of judicial ignorance of the substance of what is being litigated over is not a new one. But the faster the world is changing, the more ignorant the people in authority are likely to be, hence the more likely to make serious errors in their decisions.

One cannot solve the problem by simply inverting the age/authority hierarchy, appointing students fresh out of law school as judges, hiring the newest graduates of Cal Tech and Harvey Mudd to supervise research laboratories. Judges have to know things about both the law and the legal system that a newly hatched lawyer has not yet learned, and managing a team of engineers requires knowledge of managing as well as knowledge of engineering. Some of the skills required for the job are in slowly changing fields, where the traditional pattern makes sense, some in rapidly changing fields, where it does not.

One result of the situation is to reinforce the natural tendency of employees to ignore or evade the instructions of their superiors. That, in my view, is what was really going on between Randy Schwartz and Intel. Schwartz believed, probably correctly, that he knew a great deal more about networked computers than the Intel executives who were giving him orders. So he followed their instructions when he thought they were watching and the rest of the time did the job the way he thought it should be done, not the way he had been told to do it.

The inverted hierarchy of expertise not only encourages employees to believe that they know how to do their job better than the managers they report to, it also encourages them to believe that they can get away with saying “Yes, sir” and then doing what they please. No old fogey of forty-five is going to figure out what they are really doing. It is an attitude especially likely in employees with the sort of personality typical among bright young technophiles.


This book is about possible futures, their perils and promises. Over the past fifty years, quite a lot has been written about future perils by people loosely describable as environmentalists. Their view that the world is threatened by increasing population, depletion of natural resources, and growing pollution has received generally favorable press and is quite widely accepted, despite its extraordinary record of false prophecies.

Thirty-five years ago, the widely popular Limits to Growth warned of a future where the only way of avoiding one catastrophe was to dive into another. That future has now arrived and, while there are still problems in the world, it is a world that, by historical standards, is strikingly lacking in the sorts of catastrophes predicted. At about the same time, Paul Ehrlich predicted that between 1968 and 1977 there would be acute food shortages that would result in the death of one-seventh of the world’s population. From the time that The Population Bomb was published to now, world food production per capita has trended up, not down; the very occasional famines of recent decades have been the result, not of a global lack of food, but of conditions, usually civil war, that prevented food from getting to the people who needed it. While some environmentalists have responded to the failure of their predictions by changing their views, many have not, arguing instead that they merely had the timing a little off.11

One reason their predictions were wrong was their failure to take account of basic economic principles. Limits to Growth consisted largely of working out the implications of large computer models describing a set of interacting systems supposed to represent the world. What struck me when I read the book was that the authors, in modeling the world, had left out the role played by human rationality. It was as if they were trying to predict what would happen on a highway by extrapolating the paths cars were following while ignoring the fact that each car had a driver with good reasons to avoid colliding with other cars.

Consider a simple example: soil depletion. In their models, if food became scarce and expensive, farmers tried to produce more on the same land. Doing so wore out the soil, so future productivity fell. If that were true, Japan, where food has long been expensive, should by now be unable to grow anything at all; in fact, Japanese agriculture is extraordinarily productive. When food is expensive and expected to stay expensive, farmers value not only present productivity but future productivity and so are willing to go to a great deal of trouble to maintain or increase the fertility of their land.

A similar analysis applies to other resources whose allocation is handled through the ordinary mechanisms of the market. If energy sources are scarce, that gives users of energy an incentive to find ways of getting by with less, producers an incentive to find new ways of producing, and so too with raw materials, crop lands, and the like.

In a world of private property, the same argument applies to population. When I have a child, he does not arrive with a claim to his per capita share of the world’s resources clutched in his fist. In order to get things – food to eat, land to live on – he, or I, or someone has to offer the owner of those resources something in exchange, something that owner considers at least as valuable. My having a child does not automatically make everyone else’s children poorer, the implicit assumption of much environmentalist literature.

So far I have been considering only interactions within markets. When we go beyond that, the situation gets more complicated. My child might end up on welfare, imposing costs on your tax-paying child. My child might, in the process of producing and consuming, generate pollution that makes your child worse off. In these cases and many others, resources are being allocated in ways other than voluntary exchange, so there is no longer any reason to assume that interactions leave both parties at least as well off as they would be without the interaction.

Before concluding that the opponents of population growth are right after all, remember that such effects can go in either direction. My child might end up as the taxpayer who is supporting your child on welfare, or sharing with your children the tax burden of supporting other people’s children on welfare. My child might invent the medicine that saves your child’s life. Once we abandon the simple framework of voluntary exchange, we can no longer be confident that I am paying all of the costs of my decision to have a child. But it requires a fairly detailed, and in large part conjectural, accounting to figure out whether, on net, the existence of my child makes yours worse off or better off.

My first piece of published economics, written more than thirty years ago, was an attempt to answer that question. I concluded that there were substantial negative effects on others from producing a child, there were substantial positive effects, and the size of the effects could not be estimated accurately enough to decide which was larger. I have seen no reason since to change that conclusion.

The environmental movement was and is wrong, but not entirely wrong. The part of their predictions that involves ordinary private goods produced and sold in markets – natural resources, energy, food – is wrong because it ignores the economic mechanisms that allocate goods across both space and time.12 That is why Julian Simon, the most visible critic of the environmentalist position, won his famous bet with Paul Ehrlich over future prices of raw materials. But the part of their predictions that involves problems such as pollution, where one person imposes costs on others without requiring their consent, and where it may be impossible to redefine property rights in any workable way for which that is not true, could be right.

Which brings us back to global warming, the newest and most widely accepted version of that problem. So far as the physics are concerned, the logic seems straightforward, although the complications of the interacting systems of atmosphere, land, and ocean make measurements of past changes and predictions of future changes difficult. The economics are equally straightforward; when I decide whether to drive, or burn, or exhale, I take account of the effects on me but not the effects on you. If the latter are negative and, multiplied by the population of the world, substantial, I will produce carbon dioxide even when, taking into account the effects on everyone, I ought not to.

Nonetheless, I do not think we ought to be doing large and expensive things at present to reduce our output of carbon dioxide. One reason is that change is not always bad. Just as in the case of population increase, it is necessary to look at both positive and negative effects. A slight increase in global temperature is probably a bad thing if you live in the tropics or on land very slightly above sea level. It may be a good thing if you live in Siberia or Norway. Judged by where most of its people live, Canada is more than 2,000 miles long and, on average, less than 100 miles wide; a moderate increase in global temperature might double its effective area.13

Temperature aside, an increase in carbon dioxide is probably a good thing. It is, after all, a major input to photosynthesis, so on average we would expect crops to grow better in an atmosphere with more of it.14

Hence, while global warming might make us substantially worse off, it is not at all clear that it would do so. Most of what I see written on the subject is pretty clearly written by people, on one side or the other, who knew what answer they wanted to get before they started, which makes it hard for careful readers to form a confident opinion. Insofar as I can extract an objective estimate from what I read, it looks as though global warming is likely to occur but unlikely to be catastrophic.15

The second reason I do not think we should be doing much about global warming at present should be clear from the first twenty chapters of this book. If current estimates are correct – and they may not be – substantial problems due to global warming are decades into the future. Even a century from now the effect on sea level, by current estimates, will be a rise of less than a meter. It is only when you go further than that that effects become really big.

We live in a radically uncertain world. It is entirely possible that, fifty years from now, our species will no longer exist. It is also possible that, fifty years from now, we will have powers enormously greater than at present. Even if we end up between those two extremes, the odds are high that in 50 or 100 years we will be living very differently than at present.

That might mean a drastic reduction in power consumption; you can have a lot of fun with very little energy in a world of deep VR. It might mean a shift to power sources such as nuclear or orbital solar that generate no greenhouse gases.16It might mean a world of low-cost space transport, with population expanding through the solar system. It might mean low-cost ways of reducing the earth’s absorption of heat from the sun, such as a very large array of orbiting mirrors. More modestly still it might, in my view probably will, mean a world sufficiently wealthy, and with sufficiently advanced engineering, to make the diking of Bangladesh a considerably less difficult project than the diking of the Netherlands was a few centuries back.

If all of this seems like wild-eyed, pie-in-the-sky speculation, consider how much the world has changed over the past century. A hundred years ago medicine could cure very nearly nothing; with rare exceptions, all a competent doctor could do was tell the patient whether he should be taking a few days off work or making his will.17 The usual form of individual transportation was walking, riding a horse, or riding in a horse-drawn carriage. The only form of rapid communication available to ordinary people was the telegraph, sending short messages in Morse code at a fairly high price. The adding machine was a recent invention; 18 the only sorts of mechanical calculator in common use were the slide rule and abacus. The first powered heavier-than-air craft capable of carrying a human being had flown just a few years earlier – for 12 seconds and 120 feet.

The changes from then to now were in large part changes in technology, in what human beings knew how to do. Those changes are continuing. Arguably their rate is accelerating, as developments in one field make easier developments in another. To make plans for the world of a century hence today based on today’s technology and practice makes no more sense than it did in 1900, when a man with a prudent eye to the future might have worried about avoiding a collapse of the transportation system due to a shortage of hay and oats.

Global warming is a problem that at some point we may want to deal with, but not a problem we ought to be dealing with now. We do not know enough. Working that far ahead risks wasting valuable resources solving problems that will solve themselves sometime between now and then or, worse, spending our resources pushing the world in what will turn out to be the wrong direction.

ext Necessarily the figures are uncertain, but it is quite possible that in only seventy years our population will amount to about eleven millions, over half of whom will be Old Age Pensioners.

ES George Orwell, 1946, discussing Britain’s looming depopulation problem. As of 2006, the population of the United Kingdom was 60,587,000. The current projection for 2016 is about 65 million.


1 Biometrics based on DNA raise a more difficult problem, as well as an interesting philosophical issue: If you use advanced nanotech or biotech to revise the DNA in every cell in your body, are you still you?

2 The numbers are based on caculations by R. J. Rummel.

3 One could argue that the Chinese civil service system – exams open to all, with the high scorers awarded positions in the civil service that carried with them status and income in a polygenous society – was such a system of selective breeding. I do not know of any evidence that it was intended to be.

4 Read, 1958.

5 Friedman (1986, 1990). Also see David Friedman, Hidden Order: The Economics of Everyday Life, Harper-Collins, 1996.

6 For a more detailed presentation of these ideas see Friedman, 2001, chapters 3 and 4.

7 An old estimate from Viscusi, 1991.

8 Friedman, 2000, p. 30. Also see Buchanan and Tullock, 1962, Mueller, 2003, and other works in the public choice literature.

9 For a critical but not unreasonable account of the evidence, see Lomborg, The Skeptical Environmentalist, pp. 258-324, and Lomborg, Cool It.

10 The upper end of the range of estimated sea level rise in the 2001 IPCC report was 80 cm. The 2007 report gives a variety of scenarios, predicting rises ranging from .18 to .59 meters while refusing to describe any of them as a maximum, and mentioning the possibility of much larger rises over periods of thousands of years and the inability to be certain that those will not occur over mere centuries. The Cato Institute has a critique of the view of global warming as a massive catastrophe that must be stopped.

More drastic scenarios have been proposed that could lead to much larger increases in world sea levels, however, including some involving the collapse of the Antarctic ice sheet.

11 Future Pundit provides a pessimistic view of the future by a thoughtful non-environmentalist, based on the implications of Darwinian selection for future human fertility.

12 For details of how markets allocate across time, see Friedman, 1996, chapter 12, or Friedman, 1986, chapter 12.

13 It is often forgotten that global warming has potential benefits as well as potential costs—also that reducing CO2 output may not be the only, or the best, way of dealing with it. There have been some interesting proposals for very large scale engineering projects to prevent global warming.

14 “Globally, the potential for food production is expect to increase with increases in average temperature over a range of 1-3 degrees centigrade, but above this it is projected to decrease.” (From the 2007 IPCC reportSince the IPCC predicts a temperature increase of .2 degrees/decade, 1-3 degrees corresponds to the next 50-150 years.

15 My blog has a sketch of my views and links to additional information. One problem is that the science of global warming inevitably gets tangled up with the politics. “The latest results from the Geophysical Fluid Dynamics Laboratory (Knutson and Tuleya, Journal of Climate, 2004) suggest that by around 2080, hurricanes may have winds and rainfall about 5% more intense than today. It has been proposed that even this tiny change may be an exaggeration as to what may happen by the end of the 21st Century.” (Chris Landsea, one of the authors of an earlier IPCC report, in his webbed explanation of why he publicly resigned from the IPCC in protest over politically motivated misstatements by the lead author of the section to which he had been asked to contribute.)

16 More modestly, it might mean terrestrial solar power largely replacing fossil fuels.

17 see Lewis Thomas, The Youngest Science.

18 By 1900 sales by Burroughs, the company founded by its inventor, were up to 972 machines worldwide.