Welcome to the Second Life Forums Archive

These forums are CLOSED. Please visit the new forums HERE

The future of AI

Azelda Garcia
Azelda Garcia
Join date: 3 Nov 2003
Posts: 819
12-21-2004 01:57
> Develop a device that exactly mimics the behaviour of a single neuron, and at approximately the same scale
> Surgically replace a neuron in a human brain with this device
> Wait a while for it to be integrated into the brain's neural network
> Repeat the above two steps, until all the neurons are replaced

You do know that there are people who assert that if you replaced each neurone in your whole brain with perfect neuron replicas, one by one, like this, that you'd no longer be "intelligent"? Strange but true.

Azelda
_____________________
Cross Lament
Loose-brained Vixen
Join date: 20 Mar 2004
Posts: 1,115
12-21-2004 09:22
That is kinda silly. That's like saying a painting done with a synthetic paintbrush (as opposed to a natural hair one) is no longer art. :)

Something else I wonder about. Is it possible that we're too 'brain-centric' when we think about human intelligence? In a way, our 'brain' really extends throughout our whole body; it's called our nervous system. I imagine most people think of it in terms of phone lines communicating remote data to the processing centre... but a lot of pre-processing goes on all over the place; the retina are a good example. Heck, even our muscles can take action without communicating with the brain, when the need arises.

Maybe in our exploration of intelligence, we're missing a bigger picture. :)
_____________________
- Making everyone's day just a little more surreal -

Teeple Linden: "OK, where did the tentacled thing go while I was playing with my face?"
Rickard Roentgen
Renaissance Punk
Join date: 4 Apr 2004
Posts: 1,869
12-30-2004 11:54
who cares about AI, I want nanites! the ability to build anything I want from a personal design or via a template simply by programming teensy little robots and providing raw chemical materials.
_____________________
Rickard Roentgen
Renaissance Punk
Join date: 4 Apr 2004
Posts: 1,869
12-30-2004 11:55
From: Cross Lament
That is kinda silly. That's like saying a painting done with a synthetic paintbrush (as opposed to a natural hair one) is no longer art. :)

Something else I wonder about. Is it possible that we're too 'brain-centric' when we think about human intelligence? In a way, our 'brain' really extends throughout our whole body; it's called our nervous system. I imagine most people think of it in terms of phone lines communicating remote data to the processing centre... but a lot of pre-processing goes on all over the place; the retina are a good example. Heck, even our muscles can take action without communicating with the brain, when the need arises.

Maybe in our exploration of intelligence, we're missing a bigger picture. :)


mouse and keyboard as intrinsic components of an AI? interesting ;)
_____________________
Vendaia Shaftoe
A Two-Spirited Person
Join date: 30 Dec 2004
Posts: 7
It's already here...
12-30-2004 14:20
How did I get here? I heard this AM from my housemate that somebody had sold an island in VR on eBay. I came to work, Googled a couple of minutes, and found out about auctions of virtual property (2ndLife is the only VR game that promotes it), and signed up for trial membership after checking it out. I am disappointed that the system is down. That gives me pause. But I can't wait to enter the 2nd World.

As for the headline title above...

I have been involved with AI since 1980. I got an XBox for Christmas this year, along with Halo2. I read an online post advising one to allow an "AI" co-soldier to drive the Warthog. I didn't think about it much, until this morning. Within the game, this 'bot knows more than I do about the game world, and can outperform a human at a particular task. If I did not know that the 'bot was artificial, I would take it for an avatar backed by a meatbody. In other words, the Turing Test would be passed by this 'bot. Do you follow me?
Kurt Zidane
Just Human
Join date: 1 Apr 2004
Posts: 636
12-30-2004 14:32
Food for thought, in 2001 a leading researchers in the field of processor design found some interesting results in evolutionary processor design. He had found that his chips were taking advantage of flaw created in the processor during the construction. In short he could not take his evolved chips and import them into other 'identical' chips.

Also the neurons of the brain may look identical, but they do not function identically. And the underlying priceable that controls how each neuron will function differently is unknown. I've always though of each neuron as a small processor of sorts. Trying to mimic a network of thousands of processor design to do specific tasks with several larger more complex processors may ultimately be inefficient.

What are the general theory for artificial intelligence?
Emulate the human bring:
* millions of simple processors combined in properly
* by understanding how the brian works, and then emulating it through scientific principles.
* evolutionary design that ultimately creates self learning and awareness
* emulate basic parts of the brain, intentionally or unintentionally. And using the functions together.

What other ways are there?
My bet is on the millions of simplistic processor or under standing the brain. Evolutionary seems like it would be the simplest, but there doesn't seem to be allot of effort being put into improving reprogrammable chips.

I wouldn't be surprised if emulation through inventive thinking continues to make the most headway, but I think it'll ultimately be the long route.
1 2 3 4