Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Yanlin]

You want my prediction?

2050+

Not earlier. Until then, we're going to be fighting an uphill battle to create an AI able to think and learn.

But we must keep it limited. Asimov FTW.

[Jakkarra]

speaking of Aasimov, he died of AIDS from a blood transfusion, didnt he?

[Yanlin]

Irrelevant is irrelevant.

[Tigershark13]

there is a major difference, machines are created by man, if we feel it necessary we can remove the AI parts that make it 'unethical' can't really do that to a living creature short of slaughtering it

[Jakkarra]

well, i guess if we lobotomised an animal, removing the parts that make it feel pain/be sentient then we could say that...

[Tigershark13]

hmm... I sppose however thats much more messy and complex with much more room for error than removing some lines of code from a machine

[Jakkarra]

meh.

[Ampersand]

Beyond all of this there already exist neural networks that are perfectly capable of novel and inventive behavior, which I've posted about before. There's a fairly impressive video out there somewhere, but I can't find it. The basic idea is two artificial neural networks, one outputting data with small a moderate degree of variation between them, and a second trained to judge the output by it's quality. It takes the best of these outputs and feeds them back into the first.

This system has actually invented products, most notably one of the smallest high powered radio antennas ever invented, meant to be used on light weight microsatellites.

Also, earlier, I didn't mean to imply that supercomputers are simulating whole brains perfectly, I meant to say it was being worked on.

[woose1]

I've already won this argument.

Q: Ohreally?
A: Yareally!
Q: If we destroy a thinking and learning AI, is it the same as slaughtering an animal?
A: 1st point: We made it. 2nd point: A computer doesn't roll around on the ground, screaming in its death throes unless we program it to.
Q: Are we ever likely to see an AI that learns and thinks?
A: Probably. There are numerous advances made in that field, demonstrating AI's that learn from their mistakes, being implemented into games/the military, etc. So I can probably safely say that the human race will live to see one.
Q: Are we ever likely to see an AI that perfectly simulates a huma-
A: NO! There are some things that just can't be programmed. Sure, if mathematically simulate every part of the autonomy, from the digestive system to the billions upon billions of individual nuerons, what would we have? A tool. A tool that is no different than your screwdriver or wrench at home. If we kicked it, there wouldn't be a consciousness telling itself, "Why did the mean man kick me?". It would run through a series of inputs and outputs that we programmed.
Besides, judging from the sheer number of weapons on this planet that sane people are too afraid to use, it's only a matter of time before a madman comes into power and decides it's a viable solution. Which, as you may have noticed, is happening right now.

And, to religious people: There is a major difference between a tool created by man and a creature created by god.

[cowofdoom78963]

My prediction is: soon.

Cause you know theres gonna be some genius who spent his whole life trying to build a robot, then does.

[Ampersand]

If I were to slowly and individually replace every neuron in your brain with one made out of silicon, at which point do you cease being a human and start being a machine?

[Yanlin]

At the point where you open my head.

[Jakkarra]

i like kitties.

[cowofdoom78963]

If I were to slowly and individually replace every neuron in your brain with one made out of silicon, at which point do you cease being a human and start being a machine?

You wouldent.

Its an injury. If someone loses a leg and gets a synthetic one their still human. If you replaced the neurons you removed with silicon ones, then you would be doing the same thing.

The only way to change if your a human is if you change your genetics.

[Neonivek]

Wow, how many times has Peter given his "interactivity" speech?

Hey he is all about interactivity.

We forgive him because he is almost the only guy left trying to make games like these. He can be as cocky as he wants.

Could be worse though, For Spore Will Wright not only kept saying the same things for all his demos but he also made the same mistakes. How many times did he accidently cause a interstellar conflict?