Applying Moore's law to every form of technological progress just doesn't work. Look at the internal combustion engine, space travel, or energy technology and try to fit an exponential curve to their datapoints since their introduction.
I agree, Kurzweil's analyses are just bunk, but they work reasonably well as a subjective "intuition pump". You can't logically claim anything from the "accelerating curves", except that technology does seem to be "accelerating" a lot. So "exponentially accelerating technological development" does not imply that there will be world-changing effects.
To me the whole "singularity" thing comes down to these two beliefs: a) AI is feasible. b) When true AI appears on the scene it will have a profound impact on humanity. If I ignore all the confusing and contradictory definitions of the word "singularity", then my claim is "appearance of AI leads to singularity" as opposed to Kurzweil's "accelerating change observed so far will lead to singularity".
Sure is. But look at how long it has taken us to reach a Type I class with fusion. Another technology that is eternally just 20 years away.
My main points of contention with Kurzweil are in his summary dismissal of fundamental problems in science, acting as if because Moore's law exists, nanobots running in our bloodstream are just a process node or two away.
He is also dismisses biological complexity as not being that inscrutable because our genome is only as long as microsoft word in bit length. This is just silly; the power in the code is it's expressiveness that explodes combinatorially in a given environment. You know probably far better than I do
that his statements about the ease with which we'll understand DNA, etc. to reverse engineer the brain is nowhere near as simple and linear a process as he makes it out to be.
I agree. Kurzweil makes for a nice fantasy though. I mean, it could happen...
I think we've already hit a lesser singularity and no one noticed except some crazy stoned guy that sent me spam emails: the singularity of information. I can find practically anything I want to with a few keystrokes. I can talk to anyone anywhere in the world with ease. It's pretty cool, even if it doesn't make me a post-human. Maybe human 1.1.
It does make for a nice fantasy...I got swept up in it for a while. After thinking about it for a while though, I started to have doubts, at least about his time frames.
You are right about the information singularity though, I still trip over how easy it is. I remember my earlier computing experiences back in the 80s with the commodore; the computer was a toy. Even back in 91 with my crash-prone
Mac classic; it was still a toy. Then in 96 people kept talking about this "Internet". I went to a PC at my school,
typed in a few words and answers came back. At that moment I thought Holy s---, it happened without me.
In a way, I've been in that state of shock ever since.
Applying Moore's law to every form of technological progress just doesn't work. Look at the internal combustion engine, space travel, or energy technology and try to fit an exponential curve to their datapoints since their introduction.