1,000,000 AI @ 1000fps one day

[Fidcal]

The future: an FM with a city with a million AI each with their own processor and 1,000fps? Why not? Just a few years away I guess. We can wait taffer ... we can wait...

 

http://www.physorg.com/news193896845.html

[aidakeeley]

"code breaking"

 

This is going to wreak havoc; hopefully in a good way.

 

Hopefully there are enough smart people to utilize this to bring down the messed up system and not continue to foment the current mess instead.

 

The nonsense use of the word "miracle" is not a good sign as far as counting on "smart people" (physorg? physorg dot com?)... but this kind of glitch is just a glitch that can be corrected.

Edited May 24, 2010 by aidakeeley

[Fidcal]

I can never understand how brute code breaking can work if a system insists on only a small number of failed tries then 5 minutes wait before trying again. To test billions of combinations would just take years even if the hackers code could test a billion a second he wouldn't be able to. But clearly there must be a lot of 'secure' systems that do not limit the attempts. In any event, by then we'll have quantum entanglement encryption which is 'uncrackable'. Probably <ahem>

[aidakeeley]

Logon access/authentication systems, yes, but encrypted documents under current schemes? Many. And assumed safe. Not safe anymore (or soon).

 

That's what I'm talking about. If someone somewhere has amassed a giant store of documents that the "owner" of didn't worry about because they were "strongly encrypted"... they will opened now.

 

And, yes, documents can be re-encrypted with new encryption methods, but as imagined above the formerly thought safe data is not so safe anymore (again: soon).

[aidakeeley]

Anyway, Fid,

 

1000 fps?

 

Might be overkill

[demagogue]

I like reading about quantum computing. I read a technical article talking about the state of the art just yesterday (not understanding much, of course, but liking it). There are a lot of teams trying a lot of different approaches, though apparently breakthroughs in one approach helps other approaches, so even though there's some competition and rivalry, everybody wins. But it's also clear about the major problems still to be solved ... Some things are engineering problems, but there are still fundamental unsolved questions in the physics, how to do something even theoretically.

 

As for 1000 AI at 1000 FPS, I'd be a lot more interested in 1 AI having the cognitive power approaching 1 human brain (or even a small mammal brain) at something approaching a normal timeframe. An AI with the ability to make its own decisions would be worth its weight in q-bits.

[ungoliant]

As for 1000 AI at 1000 FPS, I'd be a lot more interested in 1 AI having the cognitive power approaching 1 human brain (or even a small mammal brain) at something approaching a normal timeframe. An AI with the ability to make its own decisions would be worth its weight in q-bits.

 

I think they tried to do that in Oblivion, and all the AI went completely apeshit and started stealing from and killing eachother constantly. Probably looked like a massive riot all over the main city.

[aidakeeley]

I think they tried to do that in Oblivion, and all the AI went completely apeshit and started stealing from and killing eachother constantly. Probably looked like a massive riot all over the main city.

 

You bastard, are you making a cynical comment on man's "nature"?

 

I will fight you on this!... peacefully.

[stumpy]

still reads as 2 dimensionally as startrek

[7upMan]

I think they tried to do that in Oblivion, and all the AI went completely apeshit and started stealing from and killing eachother constantly.  Probably looked like a massive riot all over the main city.

 

Yeah, I remember how excited I was when I first heard of and watched the promo videos of Radiant AI, as it was/is called. Man, how awesome it would habe been... but the final product was horribly dumbed down. This was to a large part because, say, drug addicts with all their money spent went to their local drug dealer and killed him for his dope. Instead of finding ways to prevent that (say, by raising levels on when to do what), Radiant AI was severely limited in its abilities. A real pity... 

[demagogue]

Not exactly the sort of thing I was talking about. I meant more like actual cognitive simulation, rather than the tricks that gaming AI uses. I mean, we're talking about *quantum* computers here! You have to enlarge your ambitions a bit when you're dreaming like that.

 

But speaking of the state-of-art of gaming AI as it is now -- now that you mention it -- I think there's a lot of untapped potential there as well. I wish a studio would invest a lot in sophisticated AI because it's lagging behind other systems and can probably add a lot more to a game for the same buck. And have you guys noticed how classic games have explored AI-gameplay a lot more than they do now (at least that's my impression)? I can think of a few classic games where the gameplay/goal of the entire game was constructive interaction with the AI, and how often do you see that today? Seems strange. There's so much more potential now but so much less exploration of how to use it.

[ungoliant]

well, the point of AI in most games these days is to either talk to them and get some dialogue, or point and shoot at them. No sense wasting valuable time developing that further, right? Who on earth would be interested in things like immersion and suspension of disbelief by creating believable AI? I just want to kill shit, damnit! If you make them smarter, that just makes my job harder, and I am all about ezmode!

[stumpy]

this would be like that 100000ghz cpu chip someone in england developed in 1994 and was then told they wouldn't be funded because no-one wanted computer systems running that fast. A chip like that could probably easily make that work. They were prediting that by the turn of the centuary people be using there chips in parallel computers, where they would have 50 of these chips working together.