You may have heard of it: the idea that as science and technology advance, the rate at which technological progress increases will itself increase. In other words, as computers get cheaper, Internet connections speed up, research methods improve, and better engineering technologies are developed, we will be in a better position to further reduce the cost of computers , increase connection speeds, generate research at an even higher volume, and improve production rates. This gives rise to an exponential increase which eventually happens so quickly that predictions about what life-changing invention will come along even as soon as next week will be impossible; life itself will have changed beyond imagining. This hypothetical event is described as the technological singularity. The most outspoken proponents of the idea are probably Ray Kurtzweil and Vernor Vinge, an inventor and a science fiction author.
And the idea is easily recognizable as science fiction; a what if proposition about technology and the future. On the other hand, the reasoning behind the singularity is a straightforward extrapolation of trends that have been in effect for the length of human history. In the last 100 years we have seen more technological revolutions than in the 1000 years before that, which in turn have seen more game-changing technologies arise than the 10000 years before that. All we have to do is keep drawing the curve and we can see that before long things will be going too fast to keep up with. Kurtzweil puts that date at around 2045, so… some of you may want to reconsider your life insurance policies.
Or maybe not. Because I can think of three very good reasons to doubt why things will work out this way; in this post I’ll concentrate on just one.
Reason 1: Problems with the graphical approach
The graphical approach is a good one, in that it gives more information than we would have without it. It isn’t a good one in the sense that it always returns accurate results. The graphical approach works better when dealing with a simple or sell-understood system, in areas where higher-order terms are known to have a negligible impact on the outcome. Unfortunately the world and its technological output isn’t a simple system; rather, it has the characteristics of a chaotic one, where subtle changes in initial states can have drastic implications for later states. And not only can slight changes in initial states change the outcome, but often the hidden existence of higher-order terms means that the rules change as a system enters a new regime.
But wait, there’s more: The graph we’re using may be wrong to begin with. Technophiles and futurists may not be terribly interested in history, so it’s easy to see why they might think of progress as slow but steadily increasing in speed. In fact, it hasn’t been as slow as is generally believed. For instance, we can consider:
Ancient Egypt. The Egyptians had key-locks, water clocks, cement, and glass jars. They were also able to manufacture the colossal pyramids to a level of precision that has been described as uncanny, indicating a technological sophistication that cannot currently be gauged.
Ancient China. I mentioned gunpowder above. But the Chinese also had flame throwers, automatic crossbows, and deadly chemical warfare (in the form of bellows, mustard smoke, and lime) before the time of Christ. Less exciting, but arguably much more life-changing inventions, included paper and the compass. Likewise, the Han dynasty’s public education and standardized examinations for selecting civil servants shouldn’t be ignored.
Classical Antiquity. Public education was also a common feature of Greek city-states. Although they are well known for the invention of simple machines, archaeologists have also recovered the antikythera mechanism, an analog computer capable of predicting astronomical phenomena before the time of Christ. The Romans also had high quality cement, aqueducts capable of transporting water hundreds of miles, and possibly even some form of germ theory. There is also good reason to believe that they had effective herbal birth control.
Prehistoric Europe. That Stonehenge served astrological functions has long been known, although the methods for its construction remain mysterious. Even more mysterious is the evidence of brain surgery from healed skulls, evidently with a high survival rate.
What is most telling about these innovations is that many of them disappeared, sometimes for thousands of years, before rediscovery. Europeans during the high middle ages did not have the technological prowess to build pyramids to Egyptian standards, nor elegant astronomical devices like the Antikytheria mechanism, nor paper, or standardized examinations, nor germ theory, public education, nor even concrete for a Roman-style aqueduct. Given the spotty archaeological record, it is likely that:
- We don’t know everything the ancients knew, and,
- The possibility exists that we may not even have rediscovered everything that has been forgotten. (This isn’t mere speculation; if the ancients really did have effective birth control, the recent development of the contraceptive pill represents an example of a lost technology that was only very recently rediscovered.)
So progress to this point has not been at all smooth. The fall of Egyptian, Mayan, Babylonian, and Classical civilizations, each time with an attendant loss of information (such as the ruined Library of Alexandria) points to a halting, jagged, increase in knowledge, sometimes characterized by rapid improvements and other times decline and stagnation.
It would be foolish to debate the fact that we are currently on a technological upswing. Yet some commentators (e.g. Charles Murray) have argued that for the last half-century the rate of increase is no longer increasing, but declining. Even considering the graph at the top of this post, which focuses on Industrial and modern technology at the expense of many earlier discoveries, it is hard to miss the widening time gap between discoveries over the past fifty years. When the Atlantic compiled their list of the 50 Greatest Breakthroughs Since the Wheel, sixteen were introduced from 1900 to the 1970’s, but no technologies of the past 40 years made their list. I chose mobile technology to represent the most recent advance, but its overall impact may not be very high, particularly when compared with inventions such as writing or the transistor.
So the graphical approach, even insofar as it can be relied on to provide an accurate forecast, doesn’t indicate a clear, exponential trend, rising with increasing steepness. Instead what we see is a bumpy graph that rises by fits and starts, with many seemingly “new” technologies having their origins in antiquity, and a slight leveling off over the past forty years.
This is enough for now; next time I’ll look at the issue from another perspective, considering the feasibility of making the discoveries we’d need to see for a singularity to occur.