The floppy disks were fine. I hate AOL because they had the same kind of user base as Bebo or the majority of Myspace today. For me, that makes them tainted for all eternity.
it's good that the orbit isn't sable, but i'm starting to think there should be some international law regulating space junk. i'd hate to see earth's orbit becoming so cluttered that sending anything up there just ends up generating more junk due to high speed collisions. that could really be the end of space exploration, at least until we make a http://en.wikipedia.org/wiki/Gravity_gun and pull all the debris down from the surface.
As always, you can't take any of these "papers" seriously. They're never even close to mathematically rigorous, and always designed merely to attract media attention.
what, you wanted Gordon Freeman to be a black, 6 year old girl? i'm not sure how many people would enjoy battling mutants and enemy soldiers with a little girl as their character...
of course it should have. the way i see it, things like this happen because people still don't realize that data security should be treated the same way as good old physical security. if the president's bodyguards say duck, the president ducks without asking any questions. people like the secret service are respected, and they have authoritah. until everyone gets it through their thick skulls that what they do in their computer is just as real as what they do in the physical world, stupid things like these will keep happening.
ah yes, common sense. the holy grail of AI. the equivalent of a poly-time solution of NP problems. the one thing that's easiest for humans to do, and hardest to code. but the reason why is quite simple - "common sense" merely means having an extremely wide understanding of the world. you can't have "common sense" in, say, natural language processing, unless you have "common sense" about human emotions, their daily lives, environment, current events, popular culture, bla, bla, bla. in other words, the one problem in AI that, if solved, would open the doors for the solution of the majority of other AI problems that to this day remain Sci-Fi is "common sense".
A completely rational and unemotional overlord would see nothing wrong with killing people
pardon my french, but that's bullshit. you're assuming that "rationality" implies a certain ultimate purpose, like economic growth, or survival of the species, or conquering space, or whatever. this is what being rational and unemotional means:
given a set of goals, you take the steps most likely to produce the best outcome.
if one of the goals of your unemotional overlord was to maximize average human lifespan, or maximize average human lifespan while keeping the standard deviation within certain bounds, or minimizing human suffering (with some way of calculating suffering numerically), or anything like this, then the overlord wouldn't just massacre everyone in Detroit because they are costing a lot of money without giving anything back or whatever. (disclaimer: that was just a fictional example, i'm not actually saying that Detroit is useless. though it probably is...).
the way i see it, emotions are just the brain's way of telling itself what to do. i came up with an example a while ago to explain this:
say i'm hungry, and there's an apple tree next to me. the problem is, all the apples are really high up the tree. i have incentive to try to retrieve one - hunger. that incentive is represented by my desire to get the apple. (i know calling desire an emotion is a little tricky, since it doesn't seem to have any measurable physiological effects in most cases, but just go with it for now). so, i want to get that apple. but there's a risk that i will fall, which is represented by fear. my brain was an economist, it'd try to calculate the probability that i'll fall, and do a risk-benefit analysis. and that's essentially what it's doing - making me feel fear which is supposed to be proportional to the risk involved, and desire proportional to how much i need the apple. if the tree is really high, and i'm not that hungry, i probably won't take the risk, and vice versa.
an AI that makes decisions of any kind, i.e. choices between alternative courses of action, needs to make those choices in a similar manner. in that sense, it will have the equivalent of "emotions". in fact, a simple http://en.wikipedia.org/wiki/Decision_tree is the equivalent of emotions. what really matters is, as you said, motivators. thus, assuming that for now the "motivation" (or tasks, or whatever) that are given to AI don't involve self-preservation, or anything else that might lead it to harm others, we'll be OK (as long as the AI doesn't have the option to somehow change its own purpose, suddenly deciding that one of its purposes is to, i don't know, take over the world).
however, there are many applications for AI where self-preservation is important. most obviously, in the military. say i make a drone with complex AI, and tell it to go bomb some dudes in Iraq. as long as it's in battle, it has to protect itself, otherwise it'd be kind of useless. then i make a newer model, and decide to decommission the old one. sure, it should be programmed in a way that it follows orders from whoever made it. but what if a programmer screwed up, and forgot to make orders from superiors override the self-preservation "task"? you might have a problem on your hands in that case.
the point is, complex AI can cause problems even if it's not "self-aware" (whatever the heck that means... i personally think it doesn't mean anything), and wasn't designed to replicate human behavior.
just tell him to learn Lisp. one of two things will happen:
1) his brain will melt. problem solved.
2) he'll see the matrix and ascend to a higher plane of existence. With no more motivation left to get a job, he won't be any competition to you. problem solved.
Slashdot Mirror
yeah, well, i was talking about the 90's AOL, with their email and chat and all. (web 1.0 social networking, i guess)
The floppy disks were fine. I hate AOL because they had the same kind of user base as Bebo or the majority of Myspace today. For me, that makes them tainted for all eternity.
so what exactly are you proposing, banning people from becoming CEO's if they have a certain brain structure abnormality?
minority report, anyone?
no the true random number is 9
why go to war, if you can power your car by shitting in the gas tank? all you need is a couple of kids, and you'll never run out of fuel.
oh. and here i thought it was a james bond-esque RC car, with missiles, and spikes, and the whole nine yards.
it's good that the orbit isn't sable, but i'm starting to think there should be some international law regulating space junk. i'd hate to see earth's orbit becoming so cluttered that sending anything up there just ends up generating more junk due to high speed collisions. that could really be the end of space exploration, at least until we make a http://en.wikipedia.org/wiki/Gravity_gun and pull all the debris down from the surface.
if there has ever been a self-fulfilling prophecy, this is it.
i don't get it.
freakiest. game. ever.
loved it.
As always, you can't take any of these "papers" seriously. They're never even close to mathematically rigorous, and always designed merely to attract media attention.
there. fixed that for you.
what, you wanted Gordon Freeman to be a black, 6 year old girl? i'm not sure how many people would enjoy battling mutants and enemy soldiers with a little girl as their character...
I'm allergic to subliminal messages. I can't watch Fox News without going into anaphylactic shock.
I'm allergic to allergies.
Incidentally, I'm related to Godel. And Escher. But not Bach.
(not really).
Ewww
that's right. i don't get music from big labels. i get it from isohunt.
of course it should have. the way i see it, things like this happen because people still don't realize that data security should be treated the same way as good old physical security. if the president's bodyguards say duck, the president ducks without asking any questions. people like the secret service are respected, and they have authoritah. until everyone gets it through their thick skulls that what they do in their computer is just as real as what they do in the physical world, stupid things like these will keep happening.
ooh yes, mod parent up!
i'm sure many people felt the same when King Priam decided to accept the Trojan Horse.
ah yes, common sense. the holy grail of AI. the equivalent of a poly-time solution of NP problems. the one thing that's easiest for humans to do, and hardest to code. but the reason why is quite simple - "common sense" merely means having an extremely wide understanding of the world. you can't have "common sense" in, say, natural language processing, unless you have "common sense" about human emotions, their daily lives, environment, current events, popular culture, bla, bla, bla. in other words, the one problem in AI that, if solved, would open the doors for the solution of the majority of other AI problems that to this day remain Sci-Fi is "common sense".
A completely rational and unemotional overlord would see nothing wrong with killing people
pardon my french, but that's bullshit. you're assuming that "rationality" implies a certain ultimate purpose, like economic growth, or survival of the species, or conquering space, or whatever. this is what being rational and unemotional means:
given a set of goals, you take the steps most likely to produce the best outcome.
if one of the goals of your unemotional overlord was to maximize average human lifespan, or maximize average human lifespan while keeping the standard deviation within certain bounds, or minimizing human suffering (with some way of calculating suffering numerically), or anything like this, then the overlord wouldn't just massacre everyone in Detroit because they are costing a lot of money without giving anything back or whatever. (disclaimer: that was just a fictional example, i'm not actually saying that Detroit is useless. though it probably is...).
the way i see it, emotions are just the brain's way of telling itself what to do. i came up with an example a while ago to explain this:
say i'm hungry, and there's an apple tree next to me. the problem is, all the apples are really high up the tree. i have incentive to try to retrieve one - hunger. that incentive is represented by my desire to get the apple. (i know calling desire an emotion is a little tricky, since it doesn't seem to have any measurable physiological effects in most cases, but just go with it for now). so, i want to get that apple. but there's a risk that i will fall, which is represented by fear. my brain was an economist, it'd try to calculate the probability that i'll fall, and do a risk-benefit analysis. and that's essentially what it's doing - making me feel fear which is supposed to be proportional to the risk involved, and desire proportional to how much i need the apple. if the tree is really high, and i'm not that hungry, i probably won't take the risk, and vice versa.
an AI that makes decisions of any kind, i.e. choices between alternative courses of action, needs to make those choices in a similar manner. in that sense, it will have the equivalent of "emotions". in fact, a simple http://en.wikipedia.org/wiki/Decision_tree is the equivalent of emotions. what really matters is, as you said, motivators. thus, assuming that for now the "motivation" (or tasks, or whatever) that are given to AI don't involve self-preservation, or anything else that might lead it to harm others, we'll be OK (as long as the AI doesn't have the option to somehow change its own purpose, suddenly deciding that one of its purposes is to, i don't know, take over the world).
however, there are many applications for AI where self-preservation is important. most obviously, in the military. say i make a drone with complex AI, and tell it to go bomb some dudes in Iraq. as long as it's in battle, it has to protect itself, otherwise it'd be kind of useless. then i make a newer model, and decide to decommission the old one. sure, it should be programmed in a way that it follows orders from whoever made it. but what if a programmer screwed up, and forgot to make orders from superiors override the self-preservation "task"? you might have a problem on your hands in that case.
the point is, complex AI can cause problems even if it's not "self-aware" (whatever the heck that means... i personally think it doesn't mean anything), and wasn't designed to replicate human behavior.
just tell him to learn Lisp. one of two things will happen:
1) his brain will melt. problem solved.
2) he'll see the matrix and ascend to a higher plane of existence. With no more motivation left to get a job, he won't be any competition to you. problem solved.
heck, read a bunch of Kant while you're at it.