Links to Keep Work at Bay

GrrlScientist has injected Encephalon 32 into the InterTubes. Among the very nice posts collected there is The Neurocritic’s take on that political brain study. And in the “Uh-oh” department, I note that the Neurocritic also mentions that Ray Kurzweil has a moving coming out next spring, entitled The Singularity Is Near (2008).

I get a weird vibe from the writings of Ray Kurzweil. (I emphasize that I’m talking about the writings here; I’ve only met the man once, at a speakers’ panel six years ago, which incidentally was moderated by Christopher Lydon — yes, the one from the Dresden Dolls song.) On the one hand, I think that all the evidence supports a materialist view of thought, consciousness and all that; like somebody once quipped, if the lungs breathe and the kidneys filter, then the brain minds. Furthermore, it seems to me that once you take that step, you have to consider the possibility that a mind could be implemented on a different material substrate. While it might be an insuperably difficult problem in practical terms, nothing in science as I understand it forbids “strong AI” in principle.

Yeah, yeah, we can argue about “Chinese rooms” some other time. My point today is that “Strong AI” is not nearly as kooky an idea as some of the claims Kurzweil has made, particularly in The Age of Spiritual Machines (1999). Machines which think and dream and emote are a possibility one must consider once one adopts a materialist stance on neuroscience, but telling when those machines will be built by “extrapolating” the increase of computer power into the 2020s is the futurist’s version of deducing the age of the Universe by adding up the ages in the Old Testament.

And, as PZ Myers pointed out long ago, Kurzweil’s attempt to anchor his “Law of Accelerating Returns” in the geological past is almost painfully silly. It depends upon conflating events in the past and using vague or inconsistent standards even for what constitutes a data point — and, of course, fitting straight lines on log-log plots is a risky business.

5 thoughts on “Links to Keep Work at Bay”

  1. When I first heard of the technological singularity, my initial reaction was something along the lines of, “Yeah, I guess I could see that.” The Napster/Kazaa/BitTorrent wars, the debate over net neutrality, etc., etc. ad nauseum show what happens when technological developments seriously tax the ability of a society to adapt to a changing landscape. It’s not a priori ridiculous to suggest that there could at some point be a discontinuity where that ability to adapt breaks down entirely.

    We’re only meat, after all, and slowly reproducing meat, limiting our rate of evolutionary response to the new environment created by a technological world.

    But predicting a) that the singularity, a particular kind of break-down, will absolutely definitely happen, and b) when this will happen, based on the kinds of selective projections he does, is foolhardy at best and assinine at worst.

  2. I guess I should actually clarify by saying that, if there is a genuine law of accelerating returns, it’s just as likely as any other possibility to assume that law will break down, and we’ll switch to a slower rate of technological development as we overtax the ability of our monkey brains to make sense of the deep physical voodoo that allows for technological breakthroughs.

  3. I’ve known for a while that sheer processing power is only one obstacle against strong AI. That’s the easy part. The tricky part, of course, is figuring out how to program the thing. Or at least programming the thing to be able to learn what it needs to know.

    I think the latter is something you can’t extrapolate so easily.

  4. I second Bronze Bog’s comments. The real difference in modern day computers and human brains is that the former is optimized for analytical computations while the latter is optimized for pattern-recognition, learning and induction. While we’ve made tremendous advances in research on such topics in computer science, we’re nowhere near the capability to produce something that operates on our level.

    As for the singularity, I’m skeptical for two reasons: 1.) the human race has already endured two dramatic, epochal periods where the structure of a society rapidly shifted to accommodate new technological advances (the agricultural revolution and the industrial revolution, to be specific). I don’t see any reason to assume that we won’t be able to completely assimilate a third epoch. And 2.), the phenomena on which the “inevitability” of the singularity are based (e.g., Moore’s law) are only temporally local (and largely artificial) economic phenomena instead of physical constants from which we can accurately predict the future.

  5. Joshua:

    The Napster/Kazaa/BitTorrent wars, the debate over net neutrality, etc., etc. ad nauseum show what happens when technological developments seriously tax the ability of a society to adapt to a changing landscape. It’s not a priori ridiculous to suggest that there could at some point be a discontinuity where that ability to adapt breaks down entirely.

    The people seemed to adapt to Napster/Kazaa/BitTorrent just fine! It’s the politicians who couldn’t cope. I could easily see new technological developments breaking our government wide open, but Yet Another Reason to shout les aristocrats à la lanterne is hardly a fundamental discontinuity in human nature, as history shows.

    We’re only meat, after all, and slowly reproducing meat, limiting our rate of evolutionary response to the new environment created by a technological world.

    But if that technology is so impressive that it’s gonna change everything, then maybe on the way it will give us — or some fraction of “us” — a better ability to synthesize and coordinate information? The technology of 2025 may be completely mind-blowing to a person from 1600, but I bet it’ll be well within the dreams of savvy teenagers in 2020.

    It just seems a little, well, insulting to paint everyone on the Net with the ossified, dunderheaded nature of the least flexible elements of society.

    With that caveat, I basically agree with all the comments above.

Comments are closed.