• Log in or sign up for an account here: https://www.sythe.org/login

IBM progresses towards real AI - Opinions?

Axilant

Forum Addict
Joined
Feb 13, 2011
Sythe Gold
0
Messages
358
Location
Sixburgh
http://www.youtube.com/embed/JvdjYPLp1Uo

Sorry about not embedding, my youtube tags fail to work.

I personally find it thrilling that we've come this far, but I'm a bit apprehensive of creating real "artificial intelligence". iRobot anybody?

Discuss.

EDIT: Please watch before you post.
 
Wouldn't a real artificial intelligence realize that its way superior to human beings and would then start creating an army of robots that'll take over the world, enter The Matrix.
 
Like I said, it's not real AI yet... but as of right now it can interpret the human language as good if not better than your average human, and can come up with a response faster than the human brain can think. Also, even if it was real AI, it has no physical form other than computer parts, and it's not connected to the internet. Therefore it would have no way to control any other movable object.

It also can't do anything without user input.
 
It can control mechanical parts, such as the ones used to create cars.
They could easily do any number of tasks, including build better functioning robots.

skynet is here!!

Like I said, it's not connected to any sort of network other than itself, therefore no access to any other type of machine. Let's say it was connected to the internet, and didn't require user input... It could do whatever the fuck it wants.
 
Watson can't understand irony, simile, metaphor or sarcasm. It's still a mere machine; its build cannot even be compared to a human brain.

EDIT:

It is cool but i dont support IBM

did you know that they sold their services to the Nazi's

EVERY SINGLE PERSON IN A CONCENTRATION CAMP WAS RECORDED USING IBM'S RECORD KEEPING SYSTEM.

look it up.
they actively helped kill millions of people sent to these camps.

What the fuck does this have to do with anything?
 
How is this progressing toward "real" AI? It seems to me like it uses the same style of computing we've seen for years, just that it's bigger and faster.
 
Wouldn't a real artificial intelligence realize that its way superior to human beings and would then start creating an army of robots that'll take over the world, enter The Matrix.
By Matrix, I think you mean Terminator.. OH GOD SKYNET IS COMING RUN FOR YOUR LIVES!
 
the problem with real ai is it cant learn until it can learn from mistakes i will not consider it because everything is put on a memory card by another human so all it is doing it searching its huge database and coming up with the answer that best fits
 
I think it is funny how people judge how good AI is based on how well it emulates human behavior.

it would be funny if it, instead, surpassed us and never even tried to emulate humans as we are so completely illogical some times.


also, I laughed when one dude got an answer wrong, Watson buzzed in second and answered with the same wrong answer.

The other dude was like WTF
 
Watson can't understand irony, simile, metaphor or sarcasm. It's still a mere machine; its build cannot even be compared to a human brain.

That's where you're wrong. Watson understands irony, similes, metaphors, and sarcasm. That's why it's the most effective AI computer built to date. It can interpret, AND come up with a response to the English language faster than any English speaking person can.
 
the problem with real ai is it cant learn until it can learn from mistakes i will not consider it because everything is put on a memory card by another human so all it is doing it searching its huge database and coming up with the answer that best fits

Watson can learn from it's mistakes. If it messes up on a question because the question was actually sarcastic, it realizes that, and attempts to gain more information and create it's own algorithms to scan and solve problems of the same nature faster and more efficiently.
 
How is this progressing toward "real" AI? It seems to me like it uses the same style of computing we've seen for years, just that it's bigger and faster.

False. Although the components are relatively the same, it's not just searching a database. It interprets the human language the same way the brain would. It continues to get better at understanding how emotion and intonation can change a words definition and how it's being used. It does this by itself.
 
Watson can learn from it's mistakes. If it messes up on a question because the question was actually sarcastic, it realizes that, and attempts to gain more information and create it's own algorithms to scan and solve problems of the same nature faster and more efficiently.
I wouldn't say it generates algorithms on the fly, it's really more just an adjustment of how much weight the different algorithms/data used will have during the answer filtering and decision process.

False. Although the components are relatively the same, it's not just searching a database. It interprets the human language the same way the brain would. It continues to get better at understanding how emotion and intonation can change a words definition and how it's being used. It does this by itself.
I think it's more accurate to say that it interprets the human language the same way WE SUSPECT the brain would. Since we still don't really understand the brain all that well, that explains why this is still primarily a search + join related data + weigh answers solution as opposed to however our brain actually manages to THINK.

I am still highly impressed with what they've come up with and I think it will definitely have its applications in several big data crunching areas.

I just hope we don't have to wait another 14 years for their next big challenge like this.:)
 
when they get to the stage of Irobot movie ill start getting creeped out, till then they can sit there an AI the shit outa my new car or paint my house or do whatever assign them to do, when they start back chatting just never upgrade them again
 
Runescape Gamblingrunescape gambling
Back
Top