I'll backtrack and have a bit more fun with Ray Kurzweil's The Singularity Is Near. As you may recall, Kurzweil's basic thesis is that within a few decades we will (a) develop super intelligent machines and (b) this will produce an inflection in human development, and as a result massive intelligence will shortly thereafter spread throughout the entire universe ("shortly" compared to the age of life on earth, that is).
Drum goes on to worry that the experience will make us lose the will to live.
I think Kurzweil puts too much emphasis on raw computing power and too little on efficient design. The brain is a kludge but it has been evolving for hundreds of millions of years.
Consider a comparison of the basic progam of this pc and the person typing on it. That is compare the pre-installed instructions on my hard disk with my genome. My genome is 3,000,000,000 base pairs of DNA which equals 750 megabytes and would (almost) fit on a CD.
I have a rescue disk for the computer on which I type. It is not a CD it is a DVD since a CD is no where near big enough. The processing power of a pc might or might not be equivilant to that of an ant, but the basic instructions are vastly more complicated than a human genome.
Somehow all of that software mainly seems to get in the way.
Why does Kurzweil think we will make more efficient use of computers processing power than we have made of their memory ?
By the way, the human genome is not efficient. There is very little selective pressure to save on DNA (we have a few grams in our bodies) and the human genome is almost certainly about 10 times the size of an set of instructions to make a person.
I'm not afraid of the singularity. I'm sure microsoft will find some way to keep computers dumb.
No comments:
Post a Comment