Friday, December 2, 2016

10 Books You Should Read ASAP

(Or six important books that will make you realize that everything you thought you knew is wrong, as they say in other clickbaits.)

First published in 1944, Horkheimer & Adorno's The Dialectic of Enlightenment still echoes in debates about rationality and a scientific progress aiming at control over nature while it also alienates us from it. Adorno and Horkheimer's witty and erudite analyses about the entertainment industry, or totalitarian regimes and the naïve press and pundits who cannot foresee their rise to power, remain insightful today.

Humanity with its powerful brain capacity is not a deviation from the natural history, they write; with our machines, chemicals and organizational powers extending our physiology we are able to promote our own species at the cost of eradicating other species. In the bigger perspective, ideas about freedom and justice, religions and political persuasions are relevant only insofar as they increase or diminish our species' prospects of survival on Earth or in the universe.

As Adorno and Horkheimer presciently write:
The human capacity for destruction promises to become so great that―once this species has exhausted itself―a tabula rasa will have been created. Either the human species will tear itself to pieces or it will take all the earth's  fauna and flora down with it, and if the earth is still young enough, the whole procedure―to vary a  famous dictum―will have to start again on a much lower level.

The Sixth Extinction by Elizabeth Kolbert fills in the details about this on-going destruction. It begins with some history of science from a time when it was not even known or accepted that species could go extinct. Even as fossils of long extinct species were being displayed in curiosity cabinets, Carl von Linné "made no distinction between the living and the dead [species] because, in his view, none was required", Kolbert writes. By the late 18th century, the discovery of the mastodon with its bones and teeth that did not quite match those of any known animal led Cuvier to the idea of an animal still larger than the elephant. If such enormous animals were still around they surely would have been spotted somewhere, but since none had been seen, Cuvier concluded that they must have gone extinct.

Kolbert's book offers many glimpses into the process of establishing scientific theories, as popular science books should try to do. Habitat loss, the introduction of invasive species, climate change and various forms of pollution are some of the threats to the survival of many species. Building roads through rainforests is one way the habitat shrinks by trapping highly specialized species in small regions where they become more vulnerable to forest fires, hunting, diseases or other threats. Species are introduced on purpose, as we have done with some plants, or inadvertently, as with rats and snails. If no local predators feed on the introduced species, they may spread very efficiently and replace indigenous species. There is no single mechanism behind all this biodiversity loss, but there is no doubt that it is closely related to the growth of the human population and our exploitation of natural resources.

Although modern civilization has accelerated the threats to biodiversity,
there are many species (such as flightless birds) that were just too easy to hunt, which went extinct at various times when humans began to colonize the continents. The megafauna that existed reproduced very slowly and had no natural enemies. As prehistoric humans occasionally managed to hunt these animals, their population declined too gradually for anyone to notice, until there were none left.

Kolbert does not moralize, and although it is clear that humans are to blame for the current mass extinction, there are also many touching stories in the book about wildlife conservationists who go to exceptional measures trying to save endangered species. In fact, some species would not exist today were it not for the sustained efforts to keep them alive, as is the case with some frogs that have to be kept in "frog hotels" because they would not survive fungus infections in the wild.

And as the ecosystems break apart, eventually humans too will be at risk. One of the researchers Kolbert met while writing the book speculates that rats will be the most advanced species to remain.

The threat that biodiversity loss poses to humans is one of many complementary perspectives one needs to keep in mind when reading a book such as Nick Bostrom's Superintelligence. Bostrom argues that an artificial intelligence explosion is possible; it would happen when an AI discovers how to improve its own intelligence. At that point it will improve itself at an exponentially increasing rate, attaining an incomprehensible level of intelligence by far surpassing that of the most intelligent humans in the course of days or seconds. Once a superintelligence exists, the further fate of humanity would depend on it.

That part of Bostrom's argument is impeccable, but there are some serious omissions and weak spots in this very detailed book. First, there is not much of a discussion of what constitutes intelligence in humans, and what would qualify as intelligence in a machine. For example, it is not clear whether intelligence must be accompanied by consciousness or not, although Bostrom seems to think so. However, he suggests that a superintelligence is likely to take such an unfamiliar form that we might not even recognize it.

Second, since the superintelligence would develop on computers (on a single computer or a cluster), it would be enlightening to discuss the physical requirements for such an intelligence explosion. There must be physical constraints in terms of material and energy consumption. If the physical computing resources are limited any increase in an AI's capacities must come from software improvements, which must saturate at some point. Given the enormous data centers devoted to scientific number crunching, internet storage or mass surveillance, the computer resources needed for an artificial intelligence takeoff may already be there, waiting to be filled with the appropriate algorithms.

The belief in accelerating scientific and technological progress which Bostrom seems to share is far from the unfettered techno-optimism celebrated by transhumanists, but Bostrom still fails to consider the complexity of the current situation from all the relevant angles.

Although Elon Musk and Stephen Hawking have seconded his concerns for an AI takeover, many have warned about a more immediate cause of concern which is the effects of automation on the labour market. To be fair, Bostrom briefly mentions how the horse population dwindled as automobiles took over the work they performed, noting that a similar course of events could make a large part of the human labour force superfluous. And of course this is already happening. Some of the jobs perhaps could be replaced by more skilled workers, but jobs will be lost faster than they are replaced.

Bostrom is probably right about our inability to grasp the nature of superintelligence and our over-estimation of our intellectual capacities:
Far from being the smartest possible biological species, we are probably better thought of as the stupidest possible biological species capable of starting a technological civilization―a niche we filled because we got there first, not because we are in any sense optimally adapted to it.
But with some effort, we are able to discover our cognitive biases, many of which are discussed in Daniel Kahneman's Thinking, Fast and Slow. According to Kahneman and some other psychologists we have two cognitive systems. One is fast and intuitive. Relying on intuition and stereotypes it usually gets things right, but it checks no facts and does not stop to think critically, so occasionally it gets things really wrong. The slow system is used for hard problems, such as the mental multiplication of 18x23, checking the validity of a logical argument, or anything that demands prolonged attention.

A large part of Kahneman's book is devoted to behavioural economics and effects such as loss aversion. Experiments have shown that most people are reluctant to make an investment if there is a certain chance of a net loss even though the expected gain would be positive. We are not always as rational as we would like to believe, but at least we are able to demonstrate our own cognitive impediments through scientific experiments.

Behavioural economics is limited to the personal perspective, that is, to microeconomics. It prides itself on proving neoclassical theories of the rational agent wrong, but has nothing to say about the role of economics at the level of a society or among nations. Piketty shed light on the history of economic inequality during the last two centuries in Le capital au XXIe siècle (discussed elsewhere). In short, the rich get richer for various reasons, notably because at a certain level, the return on capital investment is greater than the economic growth rate. Wealth therefore accumulates among those who own stock or real estate. Technological inventions and automation also tend to increase inequality by increasing demand for well paid high-skilled workers as well as for low-skilled workers while the middle segment gets hollowed out.

Wilkinson & Pickett's The Spirit Dimension demonstrates the effects of economic inequality on societies and on individuals. They show how more equal countries fare better in several respects, including drug abuse, violence, health, literacy, imprisonment and social mobility. The Scandinavian countries and Japan range among the most equal, whereas USA and Great Britain are two of the most unequal of those considered. As a choice of method, Wilkinson & Pickett have restricted their study to developed countries in order to demonstrate the effects of economic equality separately from the general level of wealth in a country. Comparing the American states, they find the same pattern at that level.

Although the point of The Spirit Dimension is to display the negative consequences of inequality on health, social mobility and so on, each of the problem fields the book touches on has its own set of causes. For example, the high imprisonment rate in USA have other causes beside inequality, notably that prisons have become an industry run for profit.

Moreover, although several pieces of evidence point in the same direction―that inequality is causing the adverse effects in all the domains they describe―it would be unwarranted to rule out an opposite direction of causation in some cases, such as high imprisonment rates and illiteracy leading to increasing inequality. The findings of Wilkinson & Pickett have been questioned, but they have posted rebuttals of the points raised by their critics.

One of the most urgent questions now is how to reduce the impact of economic activity on the environment. The viable solution, according to Wilkinson and Pickett, is a steady-state economy which, instead of growing, adapts to changing demands. Such an economy will have its best chances in an equal society where consumerism is obsoleted in favour of community building.

The astute reader will have found that this list of urgent reading does not add up to ten titles. The reason is that the reviewer has yet to read four of those important books.

No comments:

Post a Comment

I'm not home right now, but please leave a message and I'll get back to you in the next few years.