We are the last generation to know death
A while back (and I keep meaning to talk about it), I finished the excellent book The Age of Spiritual Machines by Ray Kurzweil, an excellent book that I would recommend to anyone.
I have only two problems with the book: the first is that Kurzweil sets up an ontology I disagree with by stating that human intelligence is the apex of natural evolution. Because human intelligence is more intelligent than the natural process that evolved it, he continues, it is only logical to assume that the silicate intelligences that we create will be smarter than us. Since I try not to think hierarchically and I fundamentally disagree with both the notions that evolution has an apex and that human intelligence is something to be valued above almost all other things, I cannot subscribe to this notion.
The second is that he says a machine intelligence realized with a silicate neural structure (neural networks, et. al.) will necessarily be smarter than humans because their neurons will be able to fire a few orders of magnitude faster. Biological neurons can fire, at most, about 300 times per second while silicate neurons could fire many millions of times per second. To quote one of my favorite books, Wyrm by Mark Fabi, by how many times would one have to speed up a dog’s brain until it was smarter than a human? 100 times? 1,000,000 times? A dog that could think a million times faster than a normal dog would merely take 1/1,000,000th the time to decide whether it wanted to sniff your crotch or not.
All in all, though, the book is fantabulous. It touches upon artificial intelligence, human and post-human extensions through technology, nanotechnology, bio-engineering… along with some good ol’ fashioned futurist predictions.
One of which I agree with: we are the last generation to know death.
Given that, according to Moore’s Law, we will have a computer in 2020 that has the computational abilities of the human brain and it will cost $1,000, I think it’s safe to say that we will, at least, be able to non-intrusively scan our neural architecture into a computer and be able to simulate the execution of that architecture within a computer. Yes, I mean downloading our personalities into a computer.
What we should begin asking ourselves now, besides, “Where did I put all those existential philosophy books from Philosophy 101?”, is how these people (for they will be people) will relate to us, and how we to them. What will it be like to be sentient in a machine? What will they be able to sense? What senses will they, indeed, grant themselves? What will our reactions be and how can we start improving them now?
So, congrats, we’re probably going to be immortal. Now if we could just figure out that pesky angst thing, we’d be set for post-life.
September 9th, 2004 at 5:22 pm
The most important application for this type of artificial intelligence will be in the area of interpersonal prosthetics.
Girlfriend: “What are you thinking?”
Boyfriend: “Hold on, lemme put my computer on the phone.”
And it proceeds to rattle off a monolouge of insecurity-alleviating flattery she was really asking for. Boyfriend can then do whatever his Real Self wants to do. That would probably involve a VR porn version of the girlfriend, just to make it nice and symmetric.