X-Message-Number: 29708 From: Date: Mon, 6 Aug 2007 12:42:49 EDT Subject: Re: SIAI Promotes Singularity, May Cause End of Humanity As W... Science fiction has pointed to the dangers of AI taking over for so long that nobody is going to give a singularity computer control of The Bomb, nor much other physical power, nor power to send messages to corrupt other computers either. They will be very careful for a long, long time. (Sample from 50 years ago, Arthur C. Clark I think: They turned on the ultra-computer and asked it the ultimate question: Is there a God? "Yes, NOW there is!" it said. They leaped to pull the plug, but a bolt of lightening from the clear blue sky struck them dead and fused the plug to the socket.) We know it's a danger. We have heard that many times. Will we ever get there? The Singularity means an intelligence so far above our own that we don't know what it can do. There are many ways in which a higher intelligence may arise. Human-computer combinations, smart AI's, or smarter people through gene selection and manipulation. We are already finding genes that affect intelligence. Selective breeding, creating a hundred embryos in vitro for each wanted child and then selecting the one with the best IQ genes, gene transfer and manipulation should allow us, in 20 years or so, to create generations 15 IQ points smarter than their parents. Going from 85 to 100 in some cases will improve society (fewer dangerous louts around). From 100 to 115 will give college graduates instead of store checkers. But from rare people with 160 IQ's to children with ultra-rare 175's will be the big change. Instead of a few of those per generation there will be thousands, and they are the type that makes the great discoveries, the Newton's and Feynman's. And they will study the problem of genes and IQ's... Meanwhile progress has hardly stalled in computers. Storage size and speed keeps increasing by factors of 100, data transfer rates will go up 40 times with new schemes that use light instead of wires-and-electrons, and chips do continue to improve -- if not quite as fast as before, perhaps it's because they were so fast already there was little point in making them any faster, at least until other bottlenecks had been fixed. And progress continues on quantum computing, which promises unimaginable speeds. AI can come in two ways -- we figure out how to write smart programs *or* we map the neurons of the brain and simulate their interactions. We can do that for sea slugs with 12 neurons now, I think. Obviously we have a ways to go to work our way up to trillions of neurons in the human brain, but there are sudden jumps in capacity to map things, as when we went to taking a week to map a gene to mapping the human genome in a couple of years, and now hoping to map it for $1000. For 20 years they've been predicting the Singularity in 40 years or so -- it does keep slipping further into the future -- but in fact we are closer now than 20 years ago. We know some IQ genes, can find genes much more easily, have much faster computers, have mapped a primitive brain. As with a perfected cryonics, we are not there yet but we keep making progress and there is every reason to believe we will get there some day. Alan ************************************** Get a sneak peek of the all-new AOL at http://discover.aol.com/memed/aolcom30tour Content-Type: text/html; charset="US-ASCII" [ AUTOMATICALLY SKIPPING HTML ENCODING! ] Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=29708