New gaming AI chip in the works

[Helevitia]

Finally, someone is tackling this huge problem.

http://news.yahoo.com/s/zd/20050901/tc_zd/159343

[Nowhere]

AI?... :chuckles:

[borat]

nowhere - you seem to oppose anything that seems difficult to do to the point where you appear to link difficult with impossible. I admit AI is bloody hard, but im not going to insult these peoples efforts by laughing at what they are working on. just because something has never been done before does not mean it will not be done with a little blood sweat and hard work. for example - would you have been laughing at anybody saying flights to the moon were possible 100 years ago?

dont take offence at me, just find it hard to swallow when people mock some pretty interesting thoughts without any justification.

[Nowhere]

Huh?

Games have hardly any AI, period.

edit: and, please, even though this medium, text on a forum, is quite limited...at least try to read whole remarks I leave...and in a context also.

[Jon P. Inghram]

Your post looked more like you meant "AI in games? What a waste of time!" rather than "Game companies bother to put AI in games? Hahaha!" which is what I assume you really meant.

[Chucky Cheese]

you know what would be cool is to see this chip develope into "3D game accellerator" with a PCIe 1x bus interface and fiber ports for "SLI" operations...and its own memory subsytem of course.

its work could then be "bus transferred" to the video card...like a very advanced M3D

cc

[spadnos]

Or maybe someone should start using a GPU for simultaneous calculations of NPC actions.
"Action Shaders", if you like.
There's a lot of work being done on using the GPU for general purpose computing tasks. See http://www.gpgpu.org/

- Steve

[Nowhere]

Nah, I've meant more something like "uhmm...improve it so it'll change completelly machanics of games...THAN call it AI ". And one thing...it will change games a bit anyway, of course. But one big problem with this and approach of "AI chip": AI in games doesn't scale in the same way that graphics details do. To some extent that's also the problem of physics chip, but not to so big extent...
SImply put, changing Ai changes game significantly, leading to different level designs, different gameplay and so on... And while "physics chips" can somehow get out of this trouble by initially accerelating mostly unimportant stuff (falling snow, smoke, flowing water...), "AI chip" approach can have bigger problems here. And graduate upgrade from software solutions to hardware solutions, as was the case with graphics, won't be so easy (people could play essentially teh same game in software) - egg and chicken dillema for developers/users?
Hmmm...perhaps the best thing would be Nvidia/ATI/Creative buying these solutions and incorporating them in their future chips...users wouldn't notice the hassle, and once critical mass would come...

[Fluff]

Any thoughts on the issues with potential military applications for these components?

[Nowhere]

Ahhh...yes. After all, military wanted to use Emotion Engines from PS2 in ballistic missiles...

btw, three laws of robotics for military appliances:

1. A robot will not harm authorized Government personnel but will terminate intruders with extreme prejudice.
2. A robot will obey the orders of authorized personnel except where such orders conflict with the Third Law.
3. A robot will guard its own existence with lethal antipersonnel weaponry, because a robot is bloody expensive.

[Dr Mordrid]

The last friggin' thing I need is a smarter video game

Dr. Mordrid

[Fluff]

Nowhere,

(1) Serve the public trust, (2) protect the innocent, and (3) uphold the law (4) classified