« Economics: the freaky science? | Main | Knocking on the door again »
December 27, 2005
Kurzweil's Singular Sensation
Last night C-Span 2 showed a talk by Ray Kurzweil, speaking recently at the Council on Foreign Relations. The subject was his new book, The Singularity is Near. Regular readers will recall my column on the subject from last year.
I haven't read the book and probably won't get around to it -- but I must say, Kurzweil makes a pretty good case for the logarithmic advance of technology, which, it would seem, must necessarily lead to the Singularity at some point, exactly when depending on where we are on the curve. He mentioned advances in medicine, energy research, telecommunications, and numerous other fields, showing graphs that illustrated everything on logarithmic scales. He didn't talk much about the possible implications in terms of immortality, although he touched briefly on the subject of replacing the brain with a chip -- I think he estimated that computers will have superhuman intelligence by about 2030, and we'll have the ability to do a brain scan and replace a human brain with a chip by 2045 (at which time I'll be 82, if I'm still around). "Keep yourself alive the old-fashioned way for awhile longer," he says, and immortality will be possible. He argues that the brain is really not as complex as implied by the raw number of nerve cells -- the data required to describe it could be stored on one CD. In terms of mapping the brain, he points to the logarithmic progress that happened with the human genome mapping project as an example -- nothing seemed to be happening for several years, then all of a sudden, they got the whole genome sequenced in seemingly no time.
I'm not sure what to make of it all. The part that's interesting -- not just about the Singularity, but about technological progress as a whole -- is that no one individual, company, government, etc. is running the thing, and individual projects that flop have no effect on the overall trend.
If you're interested, here's a transcript of Kurzweil's talk at the Council on Foreign Relations.
My take on it is that there's no sense worrying, because (a) there's nothing that can -- or, maybe, should -- be done to stop it, and (b) a lot of the miraculous improvements in technology we've seen have been great to use but have not fundamentally changed human nature. I guess the proof of the pudding will be when (if) it does become possible to replace the brain with a chip -- will human nature be fundamentally changed at that point?
Just to follow up on a point I discussed in last year's column about the Singularity, it might be interesting to see what happens in the area of chessplaying software -- an area in which we have already achieved "superintelligence" of a sort, in the sense that even the strongest human chess players can no longer beat the strongest programs. Will developers "hit the wall" and fail to achieve further advances, now that their programs are smarter than they are? Or will we eventually -- and most likely sooner rather than later -- find that computers have solved the ultimate question posed by chess: Can white force a win, or can black force a draw? Look ahead, say, five years, and will chess cease to be a fascinating game, since computers will have it all figured out? Or will software have gone as far as it can go, once humans can no longer understand what the software is doing (and hence not be able to develop it any further)?
In any case, I maintain a healthy skepticism about the Singularity. I'll believe it when I see it. Still, Kurzweil does make a compelling argument; one that's hard to dismiss.
Posted by Urbie at December 27, 2005 07:20 AM