5 ms·

Department of Defense outlines Skynet and Terminator development

51 points by ukdm 15y ago

danjme 15y ago
Cool, as long as they cast someone hot to play Sarah Connor
warech 15y ago
I'm surprised the DoD is spending as little as $250MM on this type of Big Data/AI research. Most of the technology outlined in Lemnois' memo is interesting and exciting, rather than intimidating and indicative of "Terminator development." I'm particularly excited about "reasoning and inference engines that can learn from experience with little training and within mission timeframes" and advances in "human-machine interfaces to enable seamless collaboration for operations and for training." Maybe a little less hyperbole in the post title next time?
- aoprisan 15y ago
  they're not only spending $250mil, they're spending an ADDITIONAL $250mil on this new project.
- georgemcfly 15y ago
  Agreed. In fact, none of it is indicative of terminator development. That's a totally different DoD initiative.
marshallp 15y ago
This is good news. An Artificial Intelligence arms race between the US and China would means the Singularity is on it's way.
- bh42222 15y ago
  May you live in interesting times.
  - tomjen3 15y ago
    Thank you.
- lukeschlather 15y ago
  Most singularities are probably not good things. (I would say one that starts as a war between the US and China would go South pretty quickly.)
- bermanoid 15y ago
  No. An AI arms race between the US and China means one of the many bad versions of the Singularity is on its way. You know, the ones where computers eventually hit self-improving territory and grow so powerful so fast that all the design mistakes inevitably made get magnified and the pathetic attempts at safety that were implemented prove to be completely pointless, leaving the computers to eventually wipe us out, whether by accident or on purpose. If we're lucky we'll get to lose an actual war against them that plays out on a human timescale, but chances are that won't even be an option, our loss will be so quick and decisive. That this is military research makes this outcome vastly more certain, as the systems would be weaponized from the beginning. However, there's hope. They mention "big data", which means they're unlikely to ever get truly general AI, let alone the self-improving kind - "big data" is what you do when your system is stupid and your data is cheap, and if that's your situation, you're never going to be build a system that's smart, because that requires extracting as much information from minimal input as possible. AGI is not an incremental improvement on machine learning methods, it's a whole different beast. As long as they're fundamentally misguided from the start, they won't be much of a threat. China, on the other hand, does have people working on the "right" parts of this problem, and that's a lot more concerning...
troels 15y ago
Personally I find the thought of humans killing humans much more scary than machines killing humans.
- groby_b 15y ago
  I hope you're aware that that's an entirely irrational thought. It really doesn't make a difference to a dead person who killed them.
- maratd 15y ago
  We are very very far away from having any kind of artificial intelligence on the level we currently exist. We don't even understand the mechanics behind our own intelligence. So if a machine is killing humans somewhere, it's because some human pressed the button. So I agree with you, I find the thought of humans killing other humans much more disturbing.
grannyg00se 15y ago
Let's not forget the two billion dollar internet spy center from the NSA http://www.wired.com/threatlevel/2012/03/ff_nsadatacenter/all/1 http://www.wired.com/threatlevel/2012/03/ff_nsadatacenter/al...
Craiggybear 15y ago
An AI developed by the military and without oversight would be a very bad thing. A very bad thing indeed. Ask about the motives. It wouldn't be an Asimov machine for starters

51 points by ukdm 15y ago

Cool, as long as they cast someone hot to play Sarah Connor

I'm surprised the DoD is spending as little as $250MM on this type of Big Data/AI research. Most of the technology outlined in Lemnois' memo is interesting and exciting, rather than intimidating and indicative of "Terminator development." I'm particularly excited about "reasoning and inference engines that can learn from experience with little training and within mission timeframes" and advances in "human-machine interfaces to enable seamless collaboration for operations and for training." Maybe a little less hyperbole in the post title next time?

aoprisan 15y ago
they're not only spending $250mil, they're spending an ADDITIONAL $250mil on this new project.
georgemcfly 15y ago
Agreed. In fact, none of it is indicative of terminator development. That's a totally different DoD initiative.

marshallp 15y ago

This is good news. An Artificial Intelligence arms race between the US and China would means the Singularity is on it's way.

bh42222 15y ago
May you live in interesting times.
- tomjen3 15y ago
  Thank you.
lukeschlather 15y ago
Most singularities are probably not good things. (I would say one that starts as a war between the US and China would go South pretty quickly.)
bermanoid 15y ago
No. An AI arms race between the US and China means one of the many bad versions of the Singularity is on its way. You know, the ones where computers eventually hit self-improving territory and grow so powerful so fast that all the design mistakes inevitably made get magnified and the pathetic attempts at safety that were implemented prove to be completely pointless, leaving the computers to eventually wipe us out, whether by accident or on purpose. If we're lucky we'll get to lose an actual war against them that plays out on a human timescale, but chances are that won't even be an option, our loss will be so quick and decisive. That this is military research makes this outcome vastly more certain, as the systems would be weaponized from the beginning. However, there's hope. They mention "big data", which means they're unlikely to ever get truly general AI, let alone the self-improving kind - "big data" is what you do when your system is stupid and your data is cheap, and if that's your situation, you're never going to be build a system that's smart, because that requires extracting as much information from minimal input as possible. AGI is not an incremental improvement on machine learning methods, it's a whole different beast. As long as they're fundamentally misguided from the start, they won't be much of a threat. China, on the other hand, does have people working on the "right" parts of this problem, and that's a lot more concerning...

troels 15y ago

Personally I find the thought of humans killing humans much more scary than machines killing humans.

groby_b 15y ago
I hope you're aware that that's an entirely irrational thought. It really doesn't make a difference to a dead person who killed them.
maratd 15y ago
We are very very far away from having any kind of artificial intelligence on the level we currently exist. We don't even understand the mechanics behind our own intelligence. So if a machine is killing humans somewhere, it's because some human pressed the button. So I agree with you, I find the thought of humans killing other humans much more disturbing.

grannyg00se 15y ago

Let's not forget the two billion dollar internet spy center from the NSA http://www.wired.com/threatlevel/2012/03/ff_nsadatacenter/all/1 http://www.wired.com/threatlevel/2012/03/ff_nsadatacenter/al...

Craiggybear 15y ago

An AI developed by the military and without oversight would be a very bad thing. A very bad thing indeed. Ask about the motives. It wouldn't be an Asimov machine for starters