Move a cursor across the screen just by thinking about it?

calculon00 said:
I hope it's a Linux PC so there will be several versions of skynet that are incompatible with each other and so the whole thing becomes too fragmented to compete with humanity.
Click to expand...

:rofl:

The thing is machines ai'nt better than the people that make 'em!

I can just imagine a super-intelligent machine assembling a huge horde of mechanical monsters right under humanity's nose...then as the time comes for them to fight they all get stuck trying to leave through ONE door and end up walking back and forth for long periods of time...or just start killing each other.

Anyone who's played RTS games knows what Im talking about.

Sincerely,
The Vault Dweller
 
Yup, I know whatcha mean.

What really annoys me is when you have a strong defensive position in a RTS (numerous towers, castles, etc etc) as well as ranged units and the enemy keeps sending a stream of attackers in, but only one or 2 at a time, therefore getting massacred, when a mass attack would overwhelm you.
 
PsychoSniper said:
What really annoys me is when you have a strong defensive position in a RTS (numerous towers, castles, etc etc) as well as ranged units and the enemy keeps sending a stream of attackers in, but only one or 2 at a time, therefore getting massacred, when a mass attack would overwhelm you.
Click to expand...
Yes, because a super intelligent computer would obviously use the AI from a computer game.
To assume that it would be programmed in the same way, with the same faults, as existing computers is pretty stupid.
 
Do you know how long it would take to code an AI that improves its own programming and actually evolves as a whole?

Not to mention the ability to build sensors so it can actually increase the amount and kinds of input it can receive.

I don't think we'll ever have such a brain built by anyone. Maybe some kind of foundation or subdivision of a megacorporation could afford the waste of time and money, but the direct benefits of such a development would be nil.

Using an organical brain would be a logical first step. Neural interfaces can easily find everyday applications. Some geek has already got himself an interface to a robotic hand (a subcutaneous microchip connected to his arm's nerves sends the impulses to a chip controlling the prosthesis (well, if you can call it that, as it's not actually attached to his body and not a replacement either)) and we already have experiments with virtual keyboards being controlled "by thought" IIRC.

The only thing required is to remove the objections people have against getting electronical "implants". In a few more years we will likely be able to develop a robotic prosthesis that works nearly as efficiently as the real limb.

If we'd understand how exactly the neurons are set up we could also develop hearing aids that don't require existing organs or brain cells (unlike most do these days) because we could compute the input the way the neurons would and then send the computed signals to the existing cells.

Completely replacing organic brains with computer chips (or electronic devices) would be an unethical and mostly useless step tho. It's pointless to give a complex tool a full-fledged self-improving and auto-adjusting nervous system that can evolve beyond the tool's purpose. Nobody needs a TV that can develop theories on its meaning of life. If you want to go that way, go and read Asimov's I, Robot.
 
In a recent issue of National Geographic they covered a sponsored race through an African desert. The catch is the contestants were all robot's with AI brains built into vehicles. Some were just buggies or simple toy cars. One was an actual humvee. Although the race was only ten miles...none made it. The best one got three miles than ran into a fence. Most of the machines sat still at the start or moved forward a little before becoming confused by a small rock, ditch, or plant. So as you can see thinking machines are not easy to make. Too many "if, then" statements I guess.

Sincerely,
The Vault Dweller
 
Y'know, the problem is both the programming AND the input. You need to decide what kind of input is needed and what kind of sensors could provide it. Then you need to know how the program should react in certain scenarios.

It seems to me that none of the machines was actually capable of finding the destination or calculate an alternate path if it ran into obstacles.

I don't think they were insane enough just to run it through if-then-else logic exclusively, tho, because otherwise they really are a waste of perfectly useful organic material.
 
You must log in or register to reply here.
Back
Top