Posted on 07/12/2006 6:35:37 PM PDT by annie laurie
Ping
Skynet!
That might be of use to humans at some point.
That was the first thing that popped in my head too.
WOW! Computers ace alot of humans, most of whom belong to the Demonrat party. :o)
In four years, you'll have to dig deep, deep into the archives of the scientific literature to find out what happened to this project, and why it failed totally.
Oh, and the $24 mil? That'll be looking fine, in the form of wine cellars and indoor swimming pools for all the earnest BBN managers, located in the most upscale, most liberal suburbs of Beantown USA.
The Japanese sank about 10 billon into a "Fifth Generation" language project that was suppose to be the end of programming.
I remember that. Late '80's, right. That and their robot that could paint pictures from video images of live subjects.
All bogus. AI has been living on its promises for nearly 40 years.
(steely)
Skynet is not here yet. But this is the next step in that direction.
phreakinpsycho wrote:
If this works, maybe we can run them for the Senate or put them on the Supreme Court.
---Only if we get to program them first ;)
we're going to rely on people who can't use commonsense on a daily basis to train machines to use it in clutch situations? Good Luck! This should be a complete disaster!
I kind of figured that skynet wasnt here yet, afterall it is fiction.
Yeah, right.
If they managed to imbue a computer system with common sense, then they've already surpassed the human mind.
< }B^)
AI has some fundamental problems including:
-- In order to understand a little about something, you have to know a lot about everything
-- Computers have no good mechanism to integrate new information in its proper context the first time
-- Computers might be able to eventually recognize context, but it would involve a lot of failure (which people tend to put up with from other people, but not from computers)
-- And (last but not least), no one really understands how the human brain represents, stores, and retrieves information and creates usable knowlege.
Until you solve the above problems, AI will go nowhere fast. I always get a chuckle when so called "scholars" say that AI is just a matter of adding processing power. Its all about information organization and management.
|
how refreshing to hear a fact-based, thoroughly accurate technical discussion here. AI is a field that has struggled to little avail, making only incremental progress over the years. Not that people shouldn't keep trying...
I'm all in favor of darpa but the current head (Tony Tether) knows nothing about AI and up til now has been extremely reluctant to fund anything like this whatsoever--good judgement up to this point. Not sure what changed, though there's been all sorts of darpa programs for advanced AI development during his tenure that never even got funded in the last several years. Tether has a cold-war mindset and micromanages darpa; most program managers dislike him for it, but as far as AI goes, he's probably doing taxpayers a favor.
I agree w/Steely Tom: this project will go the way of so many that came before it--utter obscurity. Shame.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.