khalid
Unelected Mod
- 14,071
- 6,775
Also, I like you sourcing your knowledge of AI on robot movies. Good job.
Bill Gates Says AI Will Be As Dangerous as Nukes
- Thread starter Shonuff
- Start date
1.) The concept of AI is the fact that you will get it to that point and it will grow by itself.
2.) Movies, books, and other media have always been a good example of the concepts we understand.
khalid
Unelected Mod
- 14,071
- 6,775
Science as portrayed in robot movies has absolutely nothing to do with real research in AI or actual concepts in AI. If you look at some hard sci-fi books, you might get somewhere depending on the authors. However, you cited "robot movies". Sorry, but Terminator and Transformers are not real science and have nothing to do with real science or research in AI.
khalid
Unelected Mod
- 14,071
- 6,775
Trying to get away from the Robot movies comment huh? Yes, the movie "I, Robot" was fucking horrible and probably made by a moron. The science, what there is in the various short stories of I,Robot, are horribly outdated as it came out in the 1950s, the very beginning of computer science as a science.
As for Philip K Dick, I love his books but you are mistaken if you think they cover seriously the concepts, limitations and difficulties in current AI research (or even things that are currently contemplating). They are enjoyable though.
Shonuff
Mr. Poopybutthole
- 5,538
- 790
Bill Gates disagrees with you. It's people like you who will be the first to welcome our robotic overlords, on bended knee.
Tuco
I got Tuco'd!
- 45,505
- 73,597
Let me just skip all the smoke a_skeleton_03 unconsciously throws up when he gets into retarded-argument mode and jump to his initial point:
Do you still believe this, a_skeleton_03? If so you're wrong and I'll prove it to you in 3 lines of joke.
The concept of learning is actually really interesting. There's an algorithm running at Carnegie Mellon for the past 5 years that is crawling the web, reading articles, learning new concepts (functions), then re-reading its database to see if it now has a better understanding of what it read before:Read the Web :: Carnegie Mellon University
Obviously it makes mistakes, but it's actually pretty accurate in what it believes are facts. e.g. it learned recently that "hairdryer is often found in guest_rooms" -- which is pretty cool given that (a) nobody told it that hairdryers or guest rooms exist, and (b) that nobody taught it the function "is often found in."
One issue with Asimov's three laws is that he assumes we can just program such abstract concepts. But if true AI learns much like a human (except much faster), then it'd have to learn morality and ethics on its own. An AI raised on the internet... that's pretty damn scary.
Obviously it makes mistakes, but it's actually pretty accurate in what it believes are facts. e.g. it learned recently that "hairdryer is often found in guest_rooms" -- which is pretty cool given that (a) nobody told it that hairdryers or guest rooms exist, and (b) that nobody taught it the function "is often found in."
One issue with Asimov's three laws is that he assumes we can just program such abstract concepts. But if true AI learns much like a human (except much faster), then it'd have to learn morality and ethics on its own. An AI raised on the internet... that's pretty damn scary.
Shonuff
Mr. Poopybutthole
- 5,538
- 790
How long before Bowser branches out from the Mushroom Kingdom, to ours? At some point, he's going to discover that there's a world with real princesses. And everyone knows that real princess pussy is better than digitized cartoon pussy. See how it works? It's not too many steps in AI. He's been stealing Princesses for decades. All it takes for an emergent Bowser is to merge with Skynet, and the next thing you know, the skies are covered with flying-robotic-Koopas-with-death-rays-for-eyes.
Watch Super Mario Gain Artificial Intelligence - Dice News
Watch Super Mario Gain Artificial Intelligence - Dice News
Mist
Eeyore Enthusiast
- 30,488
- 22,399
Yeah I throw up smoke and I am in retarded argument mode because I have an opinion on AI? One that Bill Gates also has? I guess we are both just knuckle dragging, mouth breathing, cum guzzling morons, Bill and I.
Why don't you impress us with your joke since that is really why you posted. Impress us.
Tuco
I got Tuco'd!
- 45,505
- 73,597
a_skeleton_03's modus operandi:
Step 1: Act incredulous that someone would state I make retarded and distracting arguments.
Step 2: Equate such a statement with calling Bill Gates a knuckle dragging, mouth breathing, cum guzzling moron.
Step 3: Ignore the question posed that would force me to stick to my original point instead of drifting off to a different one.
Anyway, I'll accept your latest execution of a_skeleton_03ness with acknowledgement that maintain one can never argue with an AI.
Tuco is not in the sudoers file. This incident will be reported to tarrant.
Step 1: I didn't make a retarded or distracting argument, quote it for proof.
Step 2: You made a statement that my idea was stupid because you have a web comic which was retarded and and distracting. My opinion is the same as Bill Gates, therefore if you state that my opinion is wrong you are saying his opinion is wrong. I filled it with hyperbole because that is fun and colorful.
Step 3: I change nothing about my opinion and will now go into why your web comic is flawed and a joke. You just needed an excuse to post a web comic that came to mind so I gave it to you.
Artificial Intelligence has to be at a certain point that it is autonomous or it isn't intelligent. When it gets to that point you may have built a "fail safe" into it where you have root access and you have to run the chance that it has taken that and done away with it. Put it in a protected space, encrypt it, make it part of the hardware with a switch you say? Do you truly have AI then? The concept of AI is that it has to be able to operate independent of user control or it's just a software simulation aka Deep Blue.
To the original point that negotiating with an AI would be easier than to negotiate with a religious terrorist I will repeat and go into detail. An AI looks at a situation logically and has a lot to do with how it is programmed, what information it has access to, and how well it can process new information. It follows a certain set of logic within parameters that science fiction writers have theorized about like Asimov. If it has made up it's mind you aren't going to change it with a passionate speech about how you feel. You will only be able to present it with facts that it may or may not need to verify and then make a decision based on this new information. A religious person however needs not pay attention to facts at all. They can obviously be persuaded by a mere discussion or a passionate speech. Refute this with more then a web comic.
Step 2: You made a statement that my idea was stupid because you have a web comic which was retarded and and distracting. My opinion is the same as Bill Gates, therefore if you state that my opinion is wrong you are saying his opinion is wrong. I filled it with hyperbole because that is fun and colorful.
Step 3: I change nothing about my opinion and will now go into why your web comic is flawed and a joke. You just needed an excuse to post a web comic that came to mind so I gave it to you.
Artificial Intelligence has to be at a certain point that it is autonomous or it isn't intelligent. When it gets to that point you may have built a "fail safe" into it where you have root access and you have to run the chance that it has taken that and done away with it. Put it in a protected space, encrypt it, make it part of the hardware with a switch you say? Do you truly have AI then? The concept of AI is that it has to be able to operate independent of user control or it's just a software simulation aka Deep Blue.
To the original point that negotiating with an AI would be easier than to negotiate with a religious terrorist I will repeat and go into detail. An AI looks at a situation logically and has a lot to do with how it is programmed, what information it has access to, and how well it can process new information. It follows a certain set of logic within parameters that science fiction writers have theorized about like Asimov. If it has made up it's mind you aren't going to change it with a passionate speech about how you feel. You will only be able to present it with facts that it may or may not need to verify and then make a decision based on this new information. A religious person however needs not pay attention to facts at all. They can obviously be persuaded by a mere discussion or a passionate speech. Refute this with more then a web comic.
That's a quote from I Robot, but according to Nick Bostrom it is much worse then that. In order for us to never come to harm, they will never let us exist at all
- 6,029
- 5,915
Replace "ignore" with "process" and I'm with you on this.
Outright ignoring emotional pain is not an appropriate response unless you're talking about an intelligence that will never actually work like ours. If you strip the emotion out of the process of thought, you're inviting decisions that will be made without taking the "humanity" of a course of action into account. Enter V.I.K.I, who will gladly sacrifice a bunch of people and the freedoms of the rest in order to maintain her purely logical vision of the status quo.