The singularity is not in our past

I think this person is absolutely right about the significance of the industrial revolution but also misses the point about the technological singularity.

There can be no doubt that the industrial revolution represented a profound break in human history. Adding to what Mr. Shalizi states in his post, we can also include massive movements of population, unprecedented increases in worker productivity and accumulation of wealth, an almost 50% increase in human life expectancy in as little as a century, a radical transformation of the cost and therefore the meaning of war, to say nothing of all the impressive and interesting intellectual advances. Of course the seeds of the industrial revolution can be traced back to the middle ages, even back to Roman times. But there can be no doubt that a large, qualitative shift took place in the 19th century which left virtually no aspect of human existence untouched.

Now, the singularity can mean a lot of things, depending on the author you read. Vernor Vinge's early paper on the subject deals mostly with superhuman intelligence, as did I.J. Good's. Jürgen Schmidhuber places emphasis on the acceleration of paradigm shifts in computing, starting back in the 17th century and going to the present. Raymond Kurzweil thinks there will hardly be a single area of life that will not be revolutionized by the singularity.

In no case, though, is the singularity simply about "unpredictable growth of technology", a "profoundly dis-orienting transformation in the life of humanity", "distributed systems of information-processing", or any of the other things Mr. Shalizi correctly states belong to the present and past of humanity. If that's what is meant by a singularity, then Mr. Shalizi is being conservative. The singularity really took place some time in the early 16th century. Schmidhuber makes this claim somewhere, but I can't find it on his byzantine web page. Basically he argues that, since historians think the European arrival in the New World and the Protestant Reformation were the two most important events of the last 1,000 years, this proves an acceleration of events that should have culminated some time around the middle of the 16th century.

But if you think about it, that's again somewhat conservative. Why not bring it back to the agrarian revolution 10,000 years ago? After all, that resulted in civilization: population increase, stratification, science, religion, art, and abstract thought. Isn't everything else a mere footnote to that?

Of course a technological singularity will disrupt history, and of course its effects are unpredictable. And when we look at the past, of course we tend to downplay the disruption and impose necessity, even though, if you lived back then, you would have found events to be neither predictable nor smooth. But I don't think that's the point. Or at least it's not the only point.

What's unique about the singularity is that it represents an exponential increase in the ratio between non-biological and biological intelligence. Merely pointing out that the rate of change in history is increasing doesn't make one a singularitarian. After all, it's easy to point out that it's been increasing for a long time. Merely pointing out that some day non-biological, superhuman intelligence will replace biological intelligence isn't distinctive, either. What makes a person singularitarian is that they believe this transformation will take place on the order of decades, not centuries. On this interpretation, it makes little difference whether you think a machine will pass the Turning Test in 2020, 2030, 2040, or 2080. Most people, if they give the issue thought at all, think it won't happen until 2100 or 2200 or never.

The same goes for the application of information technology to other fields like medicine. It's easy to grasp that life expectancy (and quality of life) have increased drastically in the last 200 years for everyone on earth. And if you read U.S. Department of Health and Human Services papers, you know the government is preparing for a population where people routinely live beyond 100 years of age. What most people do not think about—and what singularitarians do think about—is the inevitability of living indefinitely and what this means for human civilization and for the meaning of an individual life. Since a singularitarian believes this will be achievable in decades rather than centuries, it's a practical problem, not a mere theoretical problem.

So the question isn't whether there have been paradigm shifts which have thrown everything up in the air and have changed everything humans took for granted. There have been many such shifts. (And before humans, there were many such "game-changing" shifts with regard to the rules of life generally.) It's more a question of how close together those paradigm shifts are occurring now. It's a question of the time-scale you're talking about. Are the changes depicted in the wildest science fiction going to occur by the year 2500, or are they going to occur by 2050 or 2100? The scale of 10² is too large to make a difference to an individual human life (as we know it so far). But the 10¹ scale, well, that's a different story. That's why people who read about the singularity get so excited about it. That believe they'll live to see it and benefit from it.

Jared Diamond

I've been reading Guns, Germs and Steel. It's thought-provoking. The thesis that geography is what gave Eurasia the "head start" that allowed it to develop the tools to conquer the rest of the world is compelling. Diamond thinks on the largest scale of human history imaginable. He'd accept what Weber has to say about the role of Protestantism and capitalism while at the same time complaining that that sort of account is not an explanation, let alone an ultimate explanation for why Europe overtook the world. (Why did capitalism develop there? Why did Protestantism come into existence there?) As someone who has been deeply influenced by Aristotle and Marx, I relate to that search for ultimate causes.

There's also something compelling in the argument that energy production and use drives history on the largest scale. Diamond's argument is that the agrarian revolution in the Fertile Crescent 11 ka (and its subsequent spread along that line of latitude) is what ultimately gave the advantage to European civilization. But of course food has everything to do with energy. Food is converted into energy in human digestion, and it takes energy to produce food. The society that revolutionizes food production is going to set loose the innate human ability to revolutionize every other aspect of the society. The Arab agrarian revolution launches the scientific revolution in the 11th century CE. The British agrarian revolution in the 18th century launches the industrial revolution in the 19th. The Green Revolution in the 20th century is still transforming life in Asia in ways we have yet to fully witness. Yet all of these are revolutions in how food is produced, i.e., in how much energy and how many hours/person it takes to make food. Diamond's book is in many ways an updated version of Leslie White's technological theory of history.

Diamond's approach might succumb to the same accusations of technological determinism White's did. Already by the 60s, Gerhard Lenski came up with what is (at least to my mind) a more reasonable, probabilistic account of the relationship between technological development and macrosociological change. For example, hunting is predominantly a male activity in every kind of society studied by anthropologists, and polyandry is practiced in less than 1 percent of societies studied. Things like food production specifically and technology more generally probably matter very little in determining either of those sociological traits. Yet many phenomena—specialization in metal working, two or more levels of government above the community level, patrilineal clans, bride price or bride service required, and many more—do show strong correlations with the kind of subsistence technology in those societies. And since these sociological characteristics come into being after these subsistence technologies are adopted, it's a safe guess that technological advance is their cause. Broadly, population, language, social structure, and ideology appear to have technology as their chief determinant, according to Lenski.

Still, that's a great deal. And what's of most interest to me, as someone who thinks a lot about the future, is that, as technological change accelerates, the "macro" starts to collapse into the "micro". What I mean is that, in my opinion, folks like Lenski, Diamond, and White (and even to some extent Lewis Morgan, Friedrich Engels, and Karl Marx) have correctly identified the general cause of historical change in its broadest features and over its largest time scales. So while such a theory might not help explain why Germany and its allies lost World War I, it could tell us why (on Diamond's account) Pizarro took Atahualpa prisoner and why Atahualpa didn't take Charles I prisoner. Or why medieval society turned into modern industrial society.

But as time goes on and the pace of change increases, the sorts of changes the historian sees happening over the course of hundreds of years now can occur on the order of decades. It's one thing to note that we can fit twice as many features on a microprocessor every year or two. Okay, so there might be some radical implications in that for AI, and in the meantime we get cooler looking video games and cell phones. But the fact that the same law of accelerating returns is at the same time working on energy production alters those implications radically. It means that the sort of radical social change we associate with the neolithic revolution or with the industrial revolution, which unfolded over the course of thousands or hundreds of years, could be compressed into decades or years. Those quantum leaps in the way we produce food and/or power work set the creative and intellectual stage to bring about further changes in energy production, creating a positive feedback loop. So too should we expect the next revolution in energy to follow closely on the heels of the first.

To my mind, the inevitability of this shift isn't the interesting part. I'm not 100% confident we'll witness a technological singularity in which the non-biological portion of our intelligence widely outstrips the biological portion, but I am quite confident this revolution in energy production will take place. Or I should say I'd be astounded if it didn't happen in the next 20 years (barring, you know, plague or being invaded by aliens or something).

The interesting part is: What do you do with the information? The technological change is inevitable. People still have to make decisions about what to do with the technology and what to use it for. Our system of social organization—the way we're fundamentally related to one another—still has to cope with this change somehow.

This is why I proselytize a little bit about accelerating change. Not because I like toys (I can barely support a data plan for a smartphone right now) but because no one seems to realize it's coming and what's going to happen when it does. People aren't prepared for it, but they need to be. Especially our political leaders.