I grew up in the sixties. And a lot was written back then about how technological progress was constantly accelerating, and how the pace of social change would only increase. This led to predictions back in the day that we would soon be inundated by flying cars and moon bases and smellovision, that we’d suffer “future shock” from trying to assimilate the constant change, and that the next generation of kids would have an even huger generation gap with their own kids than they were then having with their parents. None of these predictions came true. No flying car for you!
But even now, I don’t think most people really understand just how far wrong they were. For instance... look around your house. How much is there that a middle class person of 1961 would not understand, at least on a basic level? This computer I’m typing on is damn near the only home appliance that they couldn’t quickly learn to use, doing something that is familiar to them in broad if not in detail. A person of 1961 could drive a 2011 car with just a few minutes of instruction. They could use the stereo and the microwave with similar ease, and the TV would be just a little harder. Compare that to how a person of 1911, coming from a house with no electricity, would respond to a midcentury electric stove or washing machine or vacuum cleaner — appliances which in many houses are essentially unchanged since 1961.
(One time when I was a teenager, I was grumbling about doing the dishes, and my dad told me to be grateful we had a dishwasher. “When I was your age, I had to wash all the dishes by hand.” I replied, “When I get to be your age, the dishwasher will wash the dishes.” Sorry, no. Nowadays I wash dishes by hand just like my dad did in the previous century, because for small loads, the dishwasher isn’t even worth using.)
Face it: the texture of life was far more radically transformed in the first half of the 20th century than in the second half. Both technically and socially, the rate of change in the first world has slowed. (Some forms of progress have even been sharply reversed, such as prosperity for the working class.)
Yet we still buy into the rhetoric of progress that builds and accelerates upward without limit. The basic fallacy is one that was often very explicitly stated back then (and sometimes still is now): that progress, and change of other kinds, is exponential. That is, that it accrues like compound interest, by growing at the same proportional rate for year after year, resulting in a curve like this:
This echoes the similar mistake made more than a century earlier by Thomas Malthus, who noted that growth of populations (whether of humans or mice or bacteria) like to follow an exponential growth curve, and predicted that humans would soon, by exponentially increasing their population, overrun their environmental carrying capacity and be subjected to catastrophic famine.
The thing is, exponential growth is, in the real world, fundamentally unsustainable. There is nothing in nature that can grow to infinity. There is always some factor which, after you grow, makes it more difficult to keep growing at that speed. So when something tries to grow exponentially, the end result is usually a curve that looks more like this:
That curve is called a sigmoid. It’s also called the logistic curve. It represents something that grows from a lower limit to an upper limit. The growth of world population is following a curve with very much that general shape. In developed countries it is already well into the flattening part in the upper right; in the developing world it’s more in the middle, but even there it may be past the point of fastest growth.
Technological progress follows a similar sort of curve. This is well understood for particular technologies. Tech that is called “mature” is understood to have reached a point where additional progress is slow and incremental. Piston-engined automobiles are an example of a mature technology. Progress from year to year is much more gradual than it was a couple of generations ago, whether you’re talking about suspension or aerodynamics or any other common aspect of the car. This is because the easy improvements have all been done already. For a really mature technology, like mechanical clocks, there is practically no sign at all of further progress being made.
It’s hard not to conclude that this same progression applies to technology as a whole. In a few hundred years, we will get to a point where it is practically unheard of for there to be a “revolutionary” invention. Of course, by then our lives will be unrecognizably transformed compared to today — we may, in fact, no longer be even remotely human. But I think we can still broadly foresee some of the limitations we’ll still live under. Technology is not magic; it can’t make whatever we imagine.
The next twenty years may be a time when “progress” seems like a bitter fairy tale. When prosperity founders but crises blossom, and lowered expectations — or even a struggle for bare survival — becomes the order of the day. But the twenty years after that...
There are still lots of areas of technology where we’re barely lifting off the ground as yet. Biotechnology and artificial intelligence are two obvious cases. They could transform our lives more radically than anything we’ve seen yet.
What brought this up for me was hearing about Vernor Vinge’s notion of the “technological singularity”. As briefly summarized, it’s basically an extension of the fallacious sixties idea of accelerating progress. In fact, much the same idea was floated back in the sixties by people like Arthur C. Clarke, before it got itself a fancy sesquipedalian name. The real originator, apparently, was Irving “Jack” Good, a former colleague of Alan Turing. His term was “intelligence explosion”.
The idea was that once any of our technology begins to intrude into the realms of extending our problem solving ability beyond the native human level, then progress will come so fast that it’s utterly impossible to predict what will come after. As Good and Clarke saw it, once we step over that line, those doing the stepping would be almost immediately pulled further along far beyond our present reach. Each step of becoming smarter — in ourselves or in our machines — immediately opens the path to become even smarter than that. Within just a few years, we might then have effectively attained a paradise, in which there is no longer anything we need to do, because our superintelligent machines take care of everything in the best possible way.
Clarke’s vision is highly optimistic, both in terms of social consequences — for instance, he assumes that those who create progress would naturally be concerned with the interests of the broad population, instead of some powerful minority — and in terms of how easy it would be to make further progress beyond the initial steps. I don’t belive it can work that way.
Good didn’t paint any picture of what society he envisioned after the “explosion”, but did say he considered this breakthrough to be essential to our long term survival. Didn’t say why.
That isn’t to say there can’t be a singularity of sorts. I just don’t believe it’ll be anywhere near as sudden and drastic as some assume. Progress from the mildly superhuman to the godlike will remain difficult and take time. But it’s still the case that we’ll be entering a realm where we can’t predict with any clarity what’s possible for us. Because it won’t just be our society that changes, but us.
The next hundred years may bring about truly staggering progress. But the hundred years after that? It may be that from that point on, nothing fundamental really changes.
[Update] It took some looking, but I found someone else who has made the same argument. This piece suggests the notion that maybe “the singularity” might correspond to the point of symmetry in the center of the sigmoid curve. Or at least to a region that includes that center point. But that would pretty much throw out everything that the coiners of that term wanted it to mean.