Scratch

sparks wrote:

You know researchers across the globe are attempting this, and have yet to truly suceed... we're not going to manage it any time soon    however, we can have a go 

sep, you have to remember that our brains work at least partly in 1s and 0s 

You are right, our brains probably work in 1s an 0s partly, at least. A while back ago, I thought, if computers can sense voltages, then why can't our brains? That got me thinking. Our brains could use input and output by different voltages, making binary more of a multinary. Ex. in binary 5 is 101 but to are brain it could be a voltage of, say, around 0.005. (I don't know anything about voltages, but I am just giving an example)

Last edited by XComputers (2012-06-13 19:44:12)

Creating an advanced AI that could 'think' and 'learn' would be like jumping in a volcano. I like volcanoes. I think I will make a majorly advanced AI starting right now. 

XComputers wrote:

sparks wrote:

ah, having a program that answers with a inbuilt set of replies is not artificially intelligent.

Sparks, you're a genius. Correct. An AI would only be programmed to have certain reactions...for a robot, heat and pain would make it try and go away, and the robot would learn if something happens a repeated amount of times then it might happen again. AIs are very advanced for Scratch. A ChatBot would not be a real AI, even if it learned. It would learn VERY slowly, in that of a robot could learn faster because it has 5 senses: Hearing, smelling, tasting, feeling (including pain), and sight. Let's say I poke a robot then slap it, then keep doing that several times. If programmed to 'learn' correctly, it would learn to back away when you poke it because it would 'know' that you were probably going to slap it after. Does this make any sense?

Grrrr necropost...

samtwheels wrote:

If we made true AI robots, it would be necessary either to limit it or to give it emotions. without morality, it could rewrite itself until it was able to do something like hacking the Pentagon.

Make it to where it can't rewrite itself without configuring it's circuitry. Make it to where the only thing it can change is it's memory. If it can't change it's own programming with ease, then it would probably not do anything evil. Let me ask you this: can we change our very own brain's memory and functions?

XComputers wrote:

SeptimusHeap wrote:

I developed a sort of 'advanced' learning AI in Panther, not one that uses human used phrases at any time, that uses phrases at the appropriate times.

http://img225.imageshack.us/img225/7681/scripts.gif

Can you post a download link for the panther project?

Nevermind. IDK How does it work, though? I slowly created it in panther by looking at the picture but if you type the same thing twice it just glitches and sits there. And when you type something, it will just ask you what you typed and keep doing that until you type the same thing again and it glitches. How does it work?