Re:There's more to Singularity studies than Kurzweil
by philwelch on Tuesday August 24, 2010 @05:54PM
(#1631288)
The singularity (sorry, Singularity) has always struck me as religion in science-fiction garb: an end time is coming, eternal life is possible if you name a cryonics firm as the beneficiary of your life insurance policy, and we will soon enter either an unimaginable paradise or an unlivable hell where either humans will be totally extinct and replaced with robots or we will be immortal cyborgs, or maybe we'll "download our consciousness into computers", or something like that. There's remarkably little content, nor any real certainty, behind the hype.
It's perfectly reasonable to say that really interesting things are bound to happen when we fully understand the brain enough to build one ourselves. It's conjectural to say it'll be possible to build intelligences better than humans at the task of building intelligences (to significant degrees of recursion), and once you make that conjecture some sort of singularity is possible. (Not inevitable--there may indeed be unsurpassable barriers to any kind of intelligence. I'm talking about logical barriers like Gödel's incompleteness theorem, or the possibility that P != NP.) But there's no gain in "futurists" making empty predictions. I'd much rather hear from the people who actually try to build intelligences what they think is feasible, or the people who actually study neurology.
It's perfectly reasonable to say that really interesting things are bound to happen when we fully understand the brain enough to build one ourselves. It's conjectural to say it'll be possible to build intelligences better than humans at the task of building intelligences (to significant degrees of recursion), and once you make that conjecture some sort of singularity is possible. (Not inevitable--there may indeed be unsurpassable barriers to any kind of intelligence. I'm talking about logical barriers like Gödel's incompleteness theorem, or the possibility that P != NP.) But there's no gain in "futurists" making empty predictions. I'd much rather hear from the people who actually try to build intelligences what they think is feasible, or the people who actually study neurology.