the daily galaxy embraces the singuarity

February 8, 2010 — Leave a comment

Sometimes I think that Luke McKinney and Casey Kazan of the Daily Galaxy have never heard of fact-checking and will publish just about anything that has a shocking headline. They’ve done this before with their posts on research into aging, studies of how quickly natural selection can do its work, astrobiology and of course, a major pet peeve of mine, utopian transhumanism. Unable to resist the lure of yet another sensational article title, they decided to take a quote from retired computer scientist and sci-fi author Vernor Vinge and turn it into a tale of how in the coming Technological Singularity, artificial intelligence will leapfrog our brainpower in the next decade, and the changes this may entail to the world as we know it today. Oh boy, here we go again…

First of all, let’s give credit where credit is due. Vinge is generally noted for coming up with what we know today as the Singularity concept in a 1993 paper which touches on a very wide array of wild possibilities for cyborgs, computers, artificial intelligence and even the eventual obsolescence of humans. However, if you take a look, you’ll see that most of the predictions are severely lacking detail and range from AI takeovers, to humans who boost their brainpower and physical abilities by merging with machines, and everything in between. That kind of breadth could make a psychic’s cold reading seem very specific and detailed. Vinge hardly elaborated any of these ideas over the past decade and a half, so what we have are a lot of vague notions eagerly embraced by a number of transhumanists in a sort of quasi-New Age movement with computers and technology cast in the role of spirits and mystical woo which brings enlightenment and supernatural powers. This is why quite a few popular quotes about the Singularity speak of it with an almost religious reverence.

Over the past year, I’ve been sent plenty of papers on Singularitarian concepts of AI and mind uploading, quite a few of them very lengthy and passionate about their subjects. But, all of them begin with the idea that there’s a black box somewhere with whatever technology is required by the author(s) fully assembled, just waiting for someone to flip the switch. Then, the papers go on to describe how they’d like to manipulate this mystery box to create a friendly AI, a brilliant AI, an AI that gathers the wisdom of the web, and so on. If this isn’t putting your cart before the horse, I don’t know what is. Trying to customize a system that doesn’t yet exist is meaningless because how your system is put together will ultimately determine what you’ll be able to do with it and how. As we struggle to define intelligence in the first place, figure out the best first steps to an AI system, and try to come to a consensus on an objective definition of AI, there are hundreds of questions to answer, hundreds of papers to write and countless experiments to carry out before anyone can start pronouncing when and if we’ll be made obsolete by our computers with any shred of scientific validity.

And as long as I’m being an evil skeptic, I also want to take issue with the Galaxy’s description of Vinge as an AI pioneer and a legendary sci-fi writer. While he did a lot of speculating on the subject and talking about how amazing the Singularity will be, finding any peer-reviewed papers on artificial intelligence authored by him is a rather difficult task. By which I mean there don’t seem to be any. How exactly could one be a “world-renowned” artificial intelligence pioneer without a stack of papers on AI development in a high impact journal? Scientists are judged by their citations and their volume of published papers, so while he may be a renowned figure for the Singularitarians, calling him a pioneer based on a speculative, roaming paper in 1993 seems like a very, very long shot. But then again, it’s the kind of long shot that the Galaxy isn’t afraid to make. A similar dilemma applies to his sci-fi work. Yes, Vinge won prestigious awards for his writing and his idea of cyberspace was a major influence on several popular cult authors. But can we really affix the term “legendary” to his stories and novels? I’m pretty sure that’s a title bestowed only on sci-fi titans like Heinlein, Asimov and Bradbury.

Regardless of what we decide to make of Vinge’s objectively significant literary efforts, Singularitarian notions about a future ruled by computers that can outthink humans with half their processors in sleep mode, cyborg elites, or just exponentially growing technocracies, should be taken with a very heft grain of salt. Without hard details backed up with thorough experimentation and working AI prototypes showing some actual cognition in the lab, all their speculation is just that, futuristic speculation which assumes that somehow, by the power of the exponential curve, all the necessary technologies will suddenly click into place precisely as they envision it. Pardon me if I don’t hold my breath for that to happen anytime soon…