DARPA's Closer than ever to Man-hunting Capable Robot Cheetah

America. Victory through superior unexpectedness since December 26, 1776.

In seriousness, any hypothetical war animal robot would most likely be used as an assassination tool or an advanced scout, so nothing to worry about.

This is cool and all, but why the fuck would we need to prepare for war with our longtime ally India?

America's strategy for long-term survival is being prepared to fight everyone. Probably at the same time.

There's talk of DARPA working out the kinks on robots that can reproduce. So suddenly we have robotic cheetahs that can make more robotic cheetahs, that we could drop from orbit into major cities...

Even friends can prove to be rivals to your power. The US Military wants to be ready to take on anyone, anywhere. Besides, India kind of got put on notice after siding with the USSR during the Cold War.

Even if you could get a robot to replicate itself it would be on a much smaller scale. Although the idea of dropping tons of nanomachines onto a city to create a ton of robot cheetahs sounds infinitely more amusing then just bombing it.

See, this is how shit starts. First we look into robots, then into animal robots, then self producing animal robots, etc. and finally we get to the point were we are all having to wear our little badges to identify our selves as friendlies to the robots, and suddenly something goes wrong and we basically get the plot to SCREAMERS and that is an end of the world scenario we can all look foward to.

On the bright side, I'll never have to pay back my student loans.

@Raikage

You got it wrong. It's WHY NOT!?

Seriously, half the DARPA projects are like "I bet you can't make X" "Oh yeah, watch me!".

Still, I want to see details of the meeting where this was approved. I mean, someone had to read plans for Man-hunting Capable Robot Cheetah and said "This looks legit".

Aren't these guys the same guys responsible for the gay bomb?

Man-eating robot cheetas isn't that much sillier...

Hey, at least the gay bomb was something that's actually possible.

I also want to say DARPA's the flying humvee guys but that just might be regular RND.

You laugh at man-eating cheetah robots now, but as you can see, it's becoming more and more possible...

Flying humvee project was stupid as shit.

This thing looks pretty cool. We could use it to replace German Shepherds, send 'em into a room to clear ahead of us. I could see fitting the cheetah-bot with some sort of bladed arm or arms on the head and use it to impale the fuckers or slice them up or whatever. Then it makes shooting them easier.

I'd much rather send a Low AI-robot in over a dog

Or how about you just give it a soft, but firm, jaw for grabbing enemies and holding them in place until a capture team arrives? Robots are meant to reduce human misery, not compound it, and if you can safely confine someone with a robot without harming humans, why not?

Or just use a dog. Cheaper. Adorabler.

But inhumaner!

And since when has that been obstacle in anything?

Make them more humanoid.

Put skin on resulting Terminators.

Or better yet, drop these cheetahs into enemy cities, key in important figures in opponent military and civilian government. Use that nifty material that lets something blend in perfectly to an environment.

This is exactly what we need to utterly destroy Iran...and, umm, whoever else we need utterly destroyed.

Make them as cheap as possible, attach high explosives, and boom, ground-level Predator drones.

Friendly RCXD in your area.

Who was Germany's long-term ally, and biggest trading partner on the 21st of June, 1941? The Soviet Union. Who were they at war with on the 22nd of June, 1941? Have a cookie if you knew the answer to that without looking it up on google or wikipedia.

The point is, in terms of international relations, what is past is seldom prologue. Just because we have been enemies or allies to a country or bloc of countries for a long time does not mean that that situation is always going to be so.

So, I for one, welcome our new cheetah robot overlords, and pray that any needed hunting of their prey on the land of India is successful assuming any need does arise.

screw 3 laws of robotics, heres my 5.

1. No learning ais unless strictly necessary

2. DON'T LET THEM REPRODUCE

3. Ai MUST NOT have unrestricted internet access unless strictly necessary.

4. have a kill signal. if it need security, have it encrypted and changed often but HAVE ONE.

5. Don't make them better in all respects, and no fully autonomous, armed and self-maintaining robots. pick any 2, but not all 3.

You forgot the most important law, mate. Murphy's. Whatever can go wrong, will go wrong, and usually at the most inconvenient time for it to go wrong. That law trumps yours and Asimov's.

Strictly necessary learning for AI presupposes that anyone programming that learning AI isn't a complete psychopath and decides that learning everything is necessary. And since he or she is probably the only one that understands the tech and thinking behind what he or she is doing, no-one will be able to gainsay him or her in whatever is decided.

On the reproduction point, squick. Particularly if such a possibility exists, it will happen, regardless of what resources need to be consumed and whatever safeguards" are in place.

Lastly, what is to stop robots who cannot fix themselves being fixed by ones that can, and getting the capability transferred across in a software upgrade?

dont GIVE THEM that ability. if you give a predator drone the programming to fix itself, the best it could do is self-debug. It couldn't repair damage. A self repairing drone should be recon-only. it's be a pain in the ass to build anyway.

And yes, murphy's law applies. Mine was made to minimalise the risk. You have a point on the learning thing. maybe just stop that all together. But a "complete physcopath" hopefully wouln't have access to the means of building fully functional war-bots.

edited 6th Mar '12 6:33:26 PM by Joesolo

You know, all that assumes that the robots are incapable of a) being reasoned with to see other's points of view or b) treated bad enough that they want to stage a violent robot revolution. There's always the third option; sane discussion with what are apparently sentient beings. Also, people who build robots tend to take good care of them. Especially hardware like this.

True. It could just be sci-fi making me paranoid. I don't like the idea of true Ai though. As stated before, murphy's law. What can go wrong, will go wrong. What if they see how we treat each other as a threat?