"The Singularity"

A number of my friends have gotten wrapped up in a movement dubbed "The Singularity." We now have a "Singularity Institute", a NASA-sponsored "Singularity University", and so on as leading futurist organizations. I tend to let my disagreements on these kinds of esoteric futurist visions slide, but I've been thinking about my agreements and disagreements with this movement for long enough that I shall now share them.

One of the basic ideas behind this movement, that computers can help make themselves smarter, and that growth that for a time looks exponential, or even super-exponential in some dimensions and may end up much faster than today may result, is by no means off the wall. Indeed computers have been helping improve future versions of themselves at least since the first compiler and circuit design software was invented. But "the Singularity" itself is an incoherent and, as the capitalization suggests, basically a religious idea. As well as a nifty concept for marketing AI research to investors who like very high risk and reward bets.

The "for a time" bit is crucial. There is as Feynman said "plenty of room at the bottom" but it is by no means infinite given actually demonstrated physics. That means all growth curves that look exponential or more in the short run turn over and become S-curves or similar in the long run, unless we discover physics that we do not now know, as information and data processing under physics as we know it are limited by the number of particles we have access to, and that in turn can only increase in the long run by at most a cubic polynomial (and probably much less than that, since space is mostly empty).

Rodney Brooks thus calls the Singularity "a period" rather than a single point in time, but if so then why call it a singularity?

As for "the Singularity" as a point past which we cannot predict, the stock market is by this definition an ongoing, rolling singularity, as are most aspects of the weather, and many quantum events, and many other aspects of our world and society. And futurists are notoriously bad at predicting the future anyway, so just what is supposed to be novel about an unpredictable future?

The Singularitarian notion of an all-encompassing or "general" intelligence flies in the face of how our modern economy, with its extreme specialization, works. We have been implementing human intelligence in computers little bits and pieces at a time, and this has been going on for centuries. First arithmetic (first with mechanical calculators), then bitwise Boolean logic (from the early parts of the 20th century with vacuum tubes), then accounting formulae and linear algebra (big mainframes of the 1950s and 60s), typesetting (Xerox PARC, Apple, Adobe, etc.), etc. etc. have each gone through their own periods of exponential and even super-exponential growth. But it's these particular operations, not intelligence in general, that exhibits such growth.

At the start of the 20th century, doing arithmetic in one's head was one of the main signs of intelligence. Today machines do quadrillions of additions and subtractions for each one done in a human brain, and this rarely bothers or even occurs to us. And the same extreme division of labor that gives us modern technology also means that AI has and will take the form of these hyper-idiot, hyper-savant, and hyper-specialized machine capabilities. Even if there was such a thing as a "general intelligence" the specialized machines would soundly beat it in the marketplace. It would be very far from a close contest.

Another way to look at the limits of this hypothetical general AI is to look at the limits of machine learning. I've worked extensively with evolutionary algorithms and other machine learning techniques. These are very promising but are also extremely limited without accurate and complete simulations of an environment in which to learn. So for example in evolutionary techniques the "fitness function" involves, critically, a simulation of electric circuits (if evolving electric circuits), of some mechanical physics (if evolving simple mechanical devices or discovering mechanical laws), and so on.

These techniques only can learn things about the real world to the extent such simulations accurately simulate the real world, but except for extremely simple situations (e.g. rediscovering the formulae for Kepler's laws based on orbital data, which a modern computer with the appropriate learning algorithm can now do in seconds) the simulations are usually very woefully incomplete, rendering the results usually useless. For example John Koza after about 20 years of working on genetic programming has discovered about that many useful inventions with it, largely involving easily simulable aspects of electronic circults. And "meta GP", genetic programming that is supposed to evolve its own GP-implementing code, is useless because we can't simulate future runs of GP without actually running them. So these evolutionary techniques, and other machine learning techniques, are often interesting and useful, but the severely limited ability of computers to simulate most real-world phenomena means that no runaway is in store, just potentially much more incremental improvements which will be much greater in simulable arenas and much smaller in others, and will slowly improve as the accuracy and completeness of our simulations slowly improves.

The other problem with tapping into computer intelligence -- and there is indeed after a century of computers quite a bit of very useful but very alien intelligence there to tap into -- is the problem of getting information from human minds to computers and vice versa. Despite all the sensory inputs we can attach to computers these days, and vast stores of human knowledge like Wikipedia that one can feed to them, almost all such data is to a computer nearly meaningless. Think Helen Keller but with most of her sense of touch removed on top of all her other tragedies. Similarly humans have an extremely tough time deciphering the output of most software unless it is extremely massaged. We humans have huge informational bottlenecks between each other, but these hardly compare to the bottlenecks between ourselves and the hyperidiot/hypersavant aliens in our midst, our computers. As a result he vast majority of programmer time is spent working on user interfaces and munging data rather than on the internal workings of programs.

Nor does the human mind, as flexible as it is, exhibit much in the way of some universal general intelligence. Many machines and many other creatures are capable of sensory, information-processing, and output feats that the human mind is quite incapable of. So even if we in some sense had a full understanding of the human mind (and it is information theoretically impossible for one human mind to fully understand even one other human mind), or could somehow faithfully "upload" a human mind to a computer (another entirely conjectural operation, which may require infeasible simulations of chemistry), we would still not have "general" intelligence, again if such a thing even exists.

That's not to say that many of the wide variety of techniques that go under the rubric "AI" are not or will not be highly useful, and may even lead to accelerated economic growth as computers help make themselves smarter. But these will turn into S-curves as they approach physical limits and the idea that this growth or these many and varied intelligences are in any nontrivial way "singular" is very wrong.

Making a toaster the hard way

(h/t Andrew Chamberlain)

For more on the economics at work here, see my essay Polynesians vs. Adam Smith.

Some conjectures and facts regarding the Malthusian isocline and the industrial revolution

Reader Matt comments on the Malthusian isocline here,, observing that I placed 17th century Britain a bit "higher" (presumably meaning more advanced, i.e. more "northeast") on the graph than 17th century China.

My placement in this case of one absolutely more advanced than the other is highly conjectural. It's worse than comparing apples to oranges. I had to put them somewhere on the graph and I used my best judgment. Britain was semi-pastoral (yet still stationary) while China had a much more productive climate for grain agriculture (much wetter in the summer and drier in the winter than Britain) and thus could support a far higher population density per hectare. But whether the Chinese population density per hectare adjusted for these natural advantages (presumably leaving technological and institutional differences, what I call denstiy per natural global hectare), which is what I'm trying to graph on my x-axis, was still much higher is far more conjectural. If we draw the x-axis as just population density per hectare (rather than "natural hectare", i.e. adjusted for natural advantages) the Chinese point is far higher, not just somewhat higher as I drew it, and thus China's isocline was clearly more advanced, but I argue due to natural advantages rather than technology or institutions. In terms of technology and institutions, it's worse than apples to oranges not only because of the radical difference between their basic agricultural strategy (labor-intensive grain-and-bean vs. semi-pastoral and heavy use of draft animals) but each had a number of advanced agricultural techniques the other rarely or never used.

See here and here for the background to these theories and the graph in question, as well as a very interesting film of street scenes that allow us to compare the Chinese and British street transportation in the early 20th century (largely before the IC engine replaced horses and rickshaws).

Note that in contrast to the above graph which is schematic and partially conjectural, the following, showing the relentless advanced of the isocline in England from the Late Middle Ages, is based on actual statistics of population and real agricultural labor wage income based on a food-dominated commodity basket:



This differs from the "noisy until 1800" graph of Clark because I've graphed 80-year averages rather than decadal averages to smooth out the noise of pests and short-term climate variations on the quality of harvests which dominated the position of the isocline before the 19th century. The points are from the data but the slope of the isoclines I drew threw them are still conjectural.

Much less conjectural than the relative levels of European and Asian 17th century isoclines and probably more important for the issue of how Britain escaped from the Malthusian trap is that the British laborer and peasant per capita nutrition (roughly corresponding to real income in an era where food dominated the peasant and laborer budget) was much higher than the Chinese. Britain had more and higher quality protein in the diet (partly from more dairy, which has genetic causes, but mainly from more meat, the British having a stationary yet semi-pastoral agriculture). China, having far fewer draft animals per capita, used laborers for arduous and repetitive tasks the Brits considered fit only for draft animals.

Part of Britain's better marginal standard of living can be put to disease, especially the Black Death, but by the 18th century I'd agree that Britain was as healthy as China. What we see after the Black Death is that the Brits greatly increased their meat-eating and use of draft animals, and, what is more important and seemingly unprecedented, were largely able to keep up this intensive use of livestock after their population recovered and even boomed beyond pre-Plague levels, whereas in (AFAIK all) prior agricultural eras and areas increasing human populations in recovery from plagues replaced draft animal with human labor and moved the diet towards grains and away from meat: sliding "souteast" down a static isocline. Also frequent periods of war and excess taxation had often led to the loss of draft animals and their replacement by human labor, in these cases with a regressing (moving "southwest") isocline. In both cases the capital accumulated in draft animals and meat livestock was destroyed. As Steve Sailer correctly points out, in Great Britain there has not been a major famine since before the Black Death while in China they were common until very recently. After the Black Death British agriculture went through a very long period of capital accumulation, in the form of soil conditioners (especially lime), commercial seeds and breeding, engineered water meadows that replaced phosphates and other scarce nutrients, drainage (important in the wet British winters), crops and rotations that fixed more nitrogen (at first growing beans for fodder, then replacing beans with the even better nitrogen fixer clover), and many other improvements besides the tremendous accumulation of livestock. Indeed "improvement" was the watchword of British agriculture as soon as works about it started being published.

So the Chinese peasant and laborer worked harder on a worse diet and lived much closer to starvation, whereas in hard times British peasants could go back to eating the foods that after the Black Death they considered animal fodder (beans, oats, etc.) whereas the analogous soybean was a protein staple of the Chinese peasant as much as it was animal fodder. Again this doesn't tell us whether China's isocline adjusted for its natural advantages was more or less advanced, it tells us that Britain was operating with a higher marginal and mean per capita standard of living and lower population density relative to its natural advantages than China, i.e. further "northwest" on its isocline.

Besides the Columbian Exchange, and even more important, there may have been an even greater but far less heralded exchange going on between Europe and East Asia in the centuries after Europe established regular oceanic contact between the two. Not so much in species, but in techniques and institutions, aided by cheap paper, the printing press, and the resulting expansion of literacy. As this contact increases we see both the East Asian and Western European isoclines advancing rapidly. But the British isocline was advancing faster than Asia's (regardless of which was more advanced absolutely), and probably more importantly in Britain and at least early on the Netherlands were largely maintaining the higher peasant/laborer standard of living despite population growth, whereas in East Asia until the 20th century the isocline started at an already low per-capita income yet the advance was still directly entirely rightward towards higher population growth rather than in catching up to British marginal standards of living. So Britain kept its big lead in per-capita standard of living and farm labor productivity, and thus presumably in the proportion of surplus labor that could be put to work on non-agricultural tasks. And since Britain did not need as large an army as a Continental power, nor as high taxes to support it, more of this labor could go into industry.

Another major factor in industrialization was urbanization, that is the proportion of this surplus labor that could be relocated where other industry or industrial resources were available rather than to where the food was grown. Urbanization (and presumably the general proportion of non-agricultural population) during the 17th-19th centuries was growing rapidly in both Western Europe, especially Britain, and Tokugawa Japan but less so in China, a phenomenon I have yet to see explained (draft animals can't explain the Japanese case as they resembled China in being relatively bereft of them). As far as agriculture and urbanization goes, contrary to myth Tokugawa Japan saw great progress, almost as much as Britain during the same period, albeit much less technological progress in industry than Britain. Urbanization depended most on water transport, i.e. the ability to transport grain to remote regions (the same thing that made ancient Rome a very large city, in her case grain imports across the Mediterranean from Sicily and Egypt). Without good water transport labor beyond that needed for agriculture had to stay in the rural areas near where the grain was grown. Japan, Britain, and the Netherlands had good water transport with many farmers near navigable water due to geography and engineering to extend the navigability of rivers and especially in the Netherlands to build canals which served a triple role of defense barrier, water control for reclamation of formerly submerged areas, and transport. Britain and the Netherlands also had an advantage in that their preferred and available source of protein, beef, could be transported over long distances over land on the hoof, whereas the pigs, fowl, and soybeans of the Chinese required water transport. Urbanization and industrialization also required transport for fuel (wood and coal: forests near navigable water were soon denuded and forests far from navigable water were generally useless. The Brits could use far more forest area per capita because their draft animals could transport more timber and fuel farther. Both the Brits and Chinese had ample and easy-to-mine coal but only in Britain were there mines within ox- or horse-transport distances of navigable water combined with a plentiful supply of these draft animals, which explains why Britain was possibly mining more coal than the rest of the world combined by the 17th century. Horse gin powered pumps also drained the British coal mines before the dawn of the Savery and Newcomen steam engines).

China had for many centuries also had good water transport due to the Grand Canal and its tributaries, but perhaps because of its artificial and inland nature it was far more bureaucratically controlled and subject to excess taxation and other institutional problems than the coastal transport in Japan and Britain. Japan, with its long and thin coastline, probably had the greatest amount of farmland next to coastline controlled by peoples speaking a single language and thus able to trade food with low transaction costs. But Britain's advantage in draft animals more than made up for having a fatter island, as their horses probably made four or more times as much farmland accessible to the navigable water than the mostly human-powered Japanese transport, and its horses also increased the efficiency of river navigations and (eventually) canals. After the railroad, and even moreso the IC engine, reached Japan it quickly caught up to and leapfrogged ahead of Britain (something Clark's theory can't explain).