Thursday, June 24, 2010

Luddites Beware

I just finished The Singularity Is Near by Ray Kurzweil. It is an absolutely fantastic dissection of that rapid growth of technology, specifically the greatly overlooked fact that information technologies are advancing not at a linear rate, but an exponential one. Kurzweil is a fantastic writer, and he approaches the material with the brain of a scientist, giving pages and pages of evidence and explanation (about 300 or so), before diving headfirst into the detailed ramifications. In the interest of grabbing your attention, and because I usually write like a novelist as opposed to a scientist, I am going to start with the ramifications and work backwards.

What is the future going to look like? Well, it depends on how far ahead you want to go, obviously, but what most people don't often understand is how much if going to change in an incredibly short span of time. As the book explains in great detail, we often look at the growth of technology as a linear curve when predicting the future. In other words, when approached with the task of imagining what the world will be like a hundred years from now, one will often think of the world one hundred years before and hypothesis, based on this, what is next. In reality, however, we are at the base of a curve in exponential growth in technology, so when projecting our expectations for the next hundred years, we might be better served to imagine billions of years of progress. Based on his calculations, Ray Kurzweil believes that we will have cured death by this time, and that technology will have progressed so far that we will essentially have filled the universe with sentient matter and energy in the form of nanobots representing our collectively reverse engineered brains, in what he describes as The Singularity. But then, that's jumping ahead a little to far isn't it?

You are probably already reassuring yourself that this somewhat terrifying concept is the lazily concocted ravings of a mad scientist. Who is this guy anyway?

Ray Kurzweil is a scientist, inventor, author and futurist, among other things. Kurzweil one of five members on the United States Army Science Advisory Board, he has received 19 honorary doctorates in various fields and won more scientific recognitions than I could even list here. He invented many of the technologies that we use every day, including scanners, speech recognition software and the first synthesizers of classical instruments. Forbes magazine ranked him as the #8 entrepreneur in the United States and described him as the “rightful heir to Thomas Edison.” Bill Gates describes him as "the best person I know at predicting the future of artificial intelligence." Suffice to say, the guy knows what he's talking about.

Because the book is 652 pages of, what some might consider, fairly dense material, attempting to summarize the concept of The Singularity in my little blog here is not going to be all that insightful compared to the actual book, which I highly recommend, but here goes.

The best way to start talking about The Singularity is to explain Moore's Law. Essentially, this is a way of describing the very apparent, long-term observation that the amount of computer hardware transistors we are capable of placing on a circuit is doubling approximately every year. This is a form of exponential growth which, by it's nature, will start slow, gradually curve, and then take off into the sky at an almost infinitely straight line. Note that the graph below is not actually linear, but displayed in approximate powers of 10.

But transistors are not the only thing growing at this rate, essential all computational technology is, while simultaneously retaining efficent costs. At the rate we have been going, the calculations of a computer worth 1000$ will surpass the intelligence of all the humans on earth by around 2040.

Or I could just let Ray explain it:

There are plenty of arguments to take into account. As my younger brother put it to me about a month ago, "People get really stoked about exponential growth because they remember learning about it in math class. They see that line going off into nowhere, but what they don't realize is that exponential growth almost never continues like that in real life." This is absolutely true. A metaphor that people often use for this is the story of a group of rabbits that find a new habitat in the form of a canyon, a small population enters and begins breeding exponentially (10 rabbits, then 20, 40, 80, 160, 320, 640, 1280, 2560, and so on). Eventually, this explosion will have to slow because there is not longer enough food to support the rabbits in this habitat, and the pattern of growth will level off into what is called an S curve.

The Singularity does has an S curve, based on the energy and resources that we have available to us, and most of the book is spent calculating what that is. Amazingly however, through the development of new technology such as quantum computing, or reversible computing, Kurzweil estimates the top of that S curve to be a "virtually unlimited limit". Specifically, he estimates the "capacity of matter and energy in our solar system to support computation to be at 10 to the 70 power calculations per second."

There are many other criqutes of Kurzweil's predictions, and he fortunately includes a "Response To Critics" section at the end to go over most of these. Additionally, they are already explained pretty well in the primary text.

If this book is to be believed, the future that we are about to experience is going to me an amazing and beautiful place. It's a text that's definitely worth checking out, or you could just wait for one of the many documentaries coming out this year.


Max said...

This is rad. I need to read that book.

Although I have one question about something. You were talking about how everything that grows exponentially has an S-curve, and that due to more advanced and efficient technology, the top of the S-curve for the exponential growth of technology will be a "virtually unlimited limit." I think eventually we can reach that, but before we do, unless there is an almost unimaginative growth in technology within a very short period of time, or people change drastically, won't there be a bump in the graph where our earthly resources have run out?

If there is that unimaginative growth before resources run out, then I can see us bypassing the problem of depleting resources. But if it never reaches that growth before depletion, the only other way I can see us reaching the proposed S-curve is if people change, which is possible, but people can be stubborn. However I'm sure the situation could dictate the necessity for change (aka if we run out of resources, we have no choice but to change).

In this case, I guess we must rely on greener forms of harnessing energy. But that seems like it could be slower, and reaching the level of exponential growth we had prior to resource depletion would take longer. However, I guess that is a major reason why are trying to find greener energy currently. So maybe we can bypass this problem as well.

Anyway, I hope that made some sense, it probably wasn't as scientific phrased as it could have been. But yeah, that's the only problem I see for us reaching that top S-curve. It could just be a lack of understanding on my part since I haven't read that book, so if you have a response to this, that would be cool. Either way, this is real interesting stuff.

The Moth said...

Good question. The exponential advances are seen in size and energy efficiency as well. Especially things like reversible computing, which is essentially almost energy-less. The resources needed are therefore rapidly declining as technology progresses.