We Are All Going To DIE!!!!

HardOCP News

[H] News
Joined
Dec 31, 1969
Messages
0
DARPA is offering a $2M prize to anyone that can build Skynet.

The Defense Advanced Research Projects Agency (DARPA) intends to hold the Cyber Grand Challenge (CGC)—the first-ever tournament for fully automatic network defense systems. DARPA envisions teams creating automated systems that would compete against each other to evaluate software, test for vulnerabilities, generate security patches and apply them to protected computers on a network. To succeed, competitors must bridge the expert gap between security software and cutting-edge program analysis research. The winning team would receive a cash prize of $2 million.
 
Dying in a machine Armageddon would be a pretty bitching way to go out, at least compared to the more mundane alternatives.
 
This is a great idea.
11lh6r5.jpg
 
I sure hope Miles Dyson gets a little more for Skynet than $2mil. seriously, Govt: you are buying a working paradox. do you know the street value of that in the theoretical physics community?
 
stupid no edit button... its a Bootstrap Paradox for anyone looking in to that sort of thing :p
 
2M.. pffft.

Software that can generate its own security patches. Somehow I don't see that working too well. For that to work, the programmer(s) would already have to know every possible security issue, and if that was the case, there would be absolutely no reason for the software to generate security patches since it would already be perfectly secure.

The perfect "security patch" would be to turn itself off.
 
If anyone can develop that they better not giving it up for a measly $2m.

Yeah, definitely. They'd probably make a very intelligent, responsive system and sell it back to the US government for 1,000 times that amount.

Dying in a machine Armageddon would be a pretty bitching way to go out, at least compared to the more mundane alternatives.

Hmm, mundane alternatives?
  • Zombies
  • Aliens
  • Epidemic of a biological vector
  • Massive crust displacement
  • Natural disasters
  • Meteors
  • World War/nuclear war
  • Robots
  • Earth's core dying
  • Dinosaurs
Hmm, two of the above entails getting chewed at, three if you count aliens that would do that.

Earth's core dying and crust displacement sounds unlikely, but makes for a cheap thrill, over-the-top, poor movie with bad acting and bad science.

Natural disasters are boring, happens often anyway.

Biological vectors such as viruses and bacteria on an epidemic scale sounds beatable. So, that's out.

Meteors would be cool but death would be too instantaneous for a lot of us, while slow for everyone else away from the landing site.

Being enslaved and anal-probed by aliens to be the birthing vessels for their young--regardless if you are female or male... eh, no thanks.

Massive world war using nuclear weapons? Hmm, sounds exciting, coldly exciting.

Robots? Well, if they self-replicate, self-repair themselves, and are too hard to kill unless we shove a tactical nuke sent straight to their AI core, sounds like a lot more fun. Why? We get coldly and soundly beaten by something that we built ourselves as the robots laugh as Nelson would have laughed.
 
to evaluate software, test for vulnerabilities, generate security patches and apply them to protected computers on a network.

Sounds overly ambitious, Skynet started as a traffic monitoring and control system.
 
So a functional network admin team and software developer team in software for $2m dollars?

So basically you invent something cheaper than outsourced labor for a multi-billion dollar industry, and you sell it for $2 million?

Yeah, makes sense.
 
Nothing to fear, they'll get it wrong and it won't work. Looking forward to the congressional hearing over why it didn't kill us all and take over the world :p
 
"The targeting system was fully functional and passed all our field trials, but when we actually released our death machines on the masses we discovered that having to select a target from more then 20 possible targets overloaded the system, caused delays in processing and eventual crashes resulting in the death machines targeting each other."

Yea, it's been a real fuck up.
 
So a functional network admin team and software developer team in software for $2m dollars?

So basically you invent something cheaper than outsourced labor for a multi-billion dollar industry, and you sell it for $2 million?

Anyone with half a brain lets China bid for it to drive the price up.
 
I just played Tic-Tac-Toe with my 5 year old son yesterday. A strange game. The only way to win is not to play. Where's my money?
 
2M.. pffft.

Software that can generate its own security patches. Somehow I don't see that working too well. For that to work, the programmer(s) would already have to know every possible security issue, and if that was the case, there would be absolutely no reason for the software to generate security patches since it would already be perfectly secure.

The perfect "security patch" would be to turn itself off.

Actually possible using domain constraint rules. But only under limited conditions. It's the same A.I. that allows machines to learn. "This set of conditions if okay" "This set of conditions is not" where a domain range function is AllowToHappen = f(Condition1, ...Condition(x));
 
Actually possible using domain constraint rules. But only under limited conditions. It's the same A.I. that allows machines to learn. "This set of conditions if okay" "This set of conditions is not" where a domain range function is AllowToHappen = f(Condition1, ...Condition(x));

You could write it as a quickly parsing binary PAULA tree to look for commonly occurring lexicons. This is basic pattern skip matching.

The trick is identifying the attack vector. Each machine would have to realize when it was compromised then look for a common thread. This means the machine would fall outside a set of operating guidelines. (Response time, memory use, exception thrown, excessive disk writes, overwriting binary files, modification to system settings etc...) A common method used by security researchers is to compare a clean machine to a machine with a just infected machine to see what has changed.

Once the vector was identified it can be added to the domain constraint relationships.
 
Only $2mil to die in a nuclear holocaust? I'm charging at least $10mil.
 
I sure hope Miles Dyson gets a little more for Skynet than $2mil. seriously, Govt: you are buying a working paradox. do you know the street value of that in the theoretical physics community?

Why should he even care? :confused: With tons of dog owners buying overpriced vacuum cleaners, that tiny amount is pretty insignificant when he's counting the latest truckload of money to show up at his gold-plated front door. No one who owns a dog even realizes you can buy a vacuum made by Doctor Tupac Bissell for like 1/10th the price and then just take out a bag once every month instead of trying to empty some freaky wind tunnel canister thing full of icky carpet goo every other time you use it.
 
Back
Top