Author: KarinsDad
Date: 19:43:58 03/28/00
Go up one level in this thread
On March 28, 2000 at 21:33:34, John Coffey wrote: >OK a little different take on this: > >Computer intelligence will never be exactly like human intelligence unless you >actually do a simulation >of the human brain (which would be a terrible waste of computer power.) So what >you would have are called Expert Systems. These are intelligent programs for >specific tasks. We already >have Expert Systems doing all sorts of work. There are programs out there >already with the knowledge >of a thousand people making decisions. (For example the stock market.) > >Ten years ago I worked for a company that developed Expert Systems. > >Now one problem with Expert Systems is that they are GIGO. Garbage in and >Garbage out. Means some >human has to tell the computer how to think. There is some learning that >takes place, but that usually >involves some sort of programmer bias as well. > >Which brings me to the point of my original post: Computers can already do a >great many things better >than humans. We already have super-intelligence in specific areas. The sky >has not fallen. I view computer intelligence as an extension of our own >intelligence. We program the computers how to think and how to make decisions. > They also provide us with information so that we can better make decisions. >Computers >are just better tools and I suspect that they always will be. > >John Coffey I agree that the sky has not fallen. If it does, it is probably many many decades away. However (and obviously, there is always a however): CPUs are currently limited (to an extent) to two dimensional space. The reason for this is the heat buildup of the chips. The third dimension is used to shunt away the heat. What happens when someone is able to create extremely low heat chips such that you could either create a three dimensional chip, or a series of two dimensional chips with small air passages between them where cool air could be pushed through them to disipate what little heat there is? Well, for one thing, you could create larger cpus. You could also increase their capacity with on chip multiprocessor functions. You could have an entire cpu the size of a basketball. How many circuits is on that? What about a cpu the size of a room? Or a house? How many orders of magnitude increase would that result in? Now, there are quite a few people around the world that are currently experimenting with artificial intelligence. And some of them are actively pursuing avenues of self learning. Are you telling us that these people will automatically fail in their efforts to create self learning programs that are NOT expert systems? Are you saying that it is impossible for a computer to ever be self learning and self aware? Are you saying that even if this is not impossible, someone will step in and prevent it from happening if it does happen? If you are not saying that, then I guess the possibility of cognizant, aware, conscious, computer systems are possible. And when man invents the hardware with 10^10 times more power than today's system (approximately 50 years, give or take), such a computer system could be born. Nothing to worry about today. Granted. But that does not mean that we should not think about it and sit around with our heads buried in the sand. KarinsDad :)
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.