What we don't know from these articles is why some or more of the AI computation can't be done in the GPU.
Because modern GPUs doesn't have a 1:1 mapping between their "cores" and general purpose logic units. Once you use a branch you effectively collapse your "cores" down to the general purpose logic associated with those "cores." (I think my gtx 660ti has like ~1400 cores and... 8 general purpose logic units?) For graphics you essentially are just doing vector/matrix math calculations with no branching so you can use all of the "cores" in parallel. This is not to mention that there are costs in streaming data to the GPU you would need to engineer around to get this to work.
All my knowledge comes from CUDA land, so maybe not 100% accurate here, but the general principle is probably correct.
TL;DR - GPUs are not general purpose chips like x86 CPUs, and there are major caveats when using them.
I looked out my window, and there was a 747 150-200 feet from me. I should clearly see the pilots, and it was being tailed by a f-whatever. You feel your building being shook by that shit and I'll see if you don't go running for the door as well.
You sir don't know how women work. They want someone to listen to their problems and agree with them. They do not want someone to fix their problems.
-A Male
Take these computers and try to teach them something.
Example:
User: So, let us say 1=A.
Computer: OK
User: Now, let us assume 2=B
Computer: Alright
User: What does 12122111 equate to?...
Now, I realize this would be easy to preprogram a response to, but if you made up something to teach it each time, and the computer actually learned it, then it would be true "intelligence." This is the only way to make the turing test legitimate.
Or maybe this is a different test altogether?
As with all fantasy the rules aren't concrete physical ones, but instead the rules of the "human experience." AKA: If the characters aren't realistic it doesn't work, and tha tis the only bound on fantasy.
So... The easy way to be a famous programmer is to be a woman?
Re:Is it ok to keep kids off the internet these da
on
Good Email For Kids?
·
· Score: 2, Insightful
The kid is five. You need to monitor a kid of 5 like 24/7 anyways.
And even if he/she even sees something "wrong" they won't understand.
I generally agree with you, but what you are saying about the internet not causing any damage is more relevant to a 10 year old.
Porn is porn, they will find a way to view it if they want to.
Algorithms, eh? My job is to design them, but before I came to this job I never worried about it in a professional setting. (school was another thing, since CS is a lot of algorithm analysis as a degree)
Seriously, you will find most sorts and searches are already programmed for you, and are optimized. And if you work in a larger framework there will probably be an optimized sort and search already implemented for you for that specific data set.
Consensus is that people do not lie on surveys. Or the fraction who do lie is so small that it is statistically insignificant.
Everyone who has taken Stats 101 knows that!
Sending those messages is of course illegal.
The article specifically says that the law covered all spam, fraudulent or not, and that the courts refused to tailor it specifically to fraud. Thus it was thrown out.
I don't blame them at all. I wish, however, that they would create laws to deal with malicious spamming, at least to some degree.
I never liked the slippery slope arguement. It operates on induction, but it is false induction, because there is no basis for the other "dominos" to fall.
The basic fact is this. The videos shows you ways to kill other humans, in general. Not self defense, suicidal methods for killing and maiming. I am not saying this information should be banned. I just think that the information shouldn't be so accesible that a 8 year old can just click a link and see the best way to bomb something.
Yeah... because comparing videos of people "hacking" to videos on how to kill people (through suicide bombing or slitting throats) is really a fair comparison.
They don't own all of one stock. Any investment bank has enough cash to effect the price of a stock in seconds if they want. They didn't really lose money, mind you. (Some people shorted, and some people bought when it was low) The real tragedy is that a company's stock value fluxed down due to inaccurate news and then didn't recover to where it was before inaccurate news was posted. This causes the stock to inaccurately reflect the actual value of the company.
I wonder though. Is this really a solid
investment? Google has struck gold mines with the whole adword and adsense, but giving broadband to countries with little to no GDP? It is a very kind move, but it strikes me as irresponsible. Of course, even if it does pan out in the end wouldn't they have seen a larger return investing in a BRIC country?
It is really a question of: Why take the greater risk?
What kind of power supply does this require? (The pump and such has to draw a bit of power?)
Does the pump run all the time? Or just when it is needed?
How much more power does this draw then a normal card?
Having done small amounts of research myself I know how easy it is to make up things (or make extended guesses) when looking in to theories/hypothosis. However, in this case, the potential damage done (medical research) is massive; people could have died. The issue of wasting money is kind of mute in comparison to damage caused by the information spread by this researcher. Sure, I can understand the first initial reaction of: "That is a lot of money!" But, in reality, this isn't that much money for a research grant. It really is disturbing that someone would falsify medical research in order to secure grant funds. The physist claiming a theory about gravity and proving himself wrong is one thing, the doctor providing medication on a fake basis is another.
This sort of litigation is the cause of so many injustices. While we waste our time smacking down these stupid cases all sorts of money/time/attention is drawn away from more important issues. I mean, there are real and serious things to be debated and decided by those who work for and with our justice department, but we spend all our time arguing about something that doesn't particularily make sense at all. Since when does google HAVE to do anything? People have a choice in using their system or not; they are not a monopoly. (Isn't MSN the most used search engine still?)
Well, if google is running a "beta" ISP somewhere I can quickly imagine them becoming a real ISP. I mean, look at google mail. It is still in "beta" but it functions 100 times more usefully than most of the webmail services I have used, and is quickly becoming popular. Just because google claims something is beta doesn't mean it isn't actually working, is what I am trying to say. I am also not sure of the needs that google has with fiber to begin with. I work at a small private college (5k students) and we use thousands of miles of fiber. I would imagine google would need a little more than us. How much did they actually buy?
Sure, distraction is bad when you are driving. However, this study seems pretty strange. I mean, first off, 40 people isn't a large sample size. Also, the types of people who are driving and using the cell phone matter. If you have some bimbo girl riding about at 100 mph talking to her friends (even faster then she drives) about her new purse she is going to be a little more likely to crash then oh... a 40 year old saying goodnight to his/her spouse on a business trip.
Futhermore, cause and effect are not clear here. I think that the factors that go in to the "cause" of an accident are a little more complex then: "She was talking on her cell!" The difference between this and drunk driving is the ability to measure the impairment. When you are drunk your level of impairment can be scientifically quantified. Your vision has decreased by X, and you r reaction time by Y, ect... When you talk on the phone this isn't the case, because some people are more easily distracted then others. In reality, if you are talking with a hands free set then you are just as impaired if you are talking to a person sitting next to you. So, by that logic, if you are going to make hands free sets illegal you should make talking in the car illegal.
Slashdot Mirror
What we don't know from these articles is why some or more of the AI computation can't be done in the GPU.
Because modern GPUs doesn't have a 1:1 mapping between their "cores" and general purpose logic units. Once you use a branch you effectively collapse your "cores" down to the general purpose logic associated with those "cores." (I think my gtx 660ti has like ~1400 cores and ... 8 general purpose logic units?) For graphics you essentially are just doing vector/matrix math calculations with no branching so you can use all of the "cores" in parallel. This is not to mention that there are costs in streaming data to the GPU you would need to engineer around to get this to work.
All my knowledge comes from CUDA land, so maybe not 100% accurate here, but the general principle is probably correct.
TL;DR - GPUs are not general purpose chips like x86 CPUs, and there are major caveats when using them.
I looked out my window, and there was a 747 150-200 feet from me. I should clearly see the pilots, and it was being tailed by a f-whatever. You feel your building being shook by that shit and I'll see if you don't go running for the door as well.
You sir don't know how women work. They want someone to listen to their problems and agree with them. They do not want someone to fix their problems. -A Male
Someone needs to play professional football, or flip my burgers if it doesn't work out for em.
Take these computers and try to teach them something. Example: User: So, let us say 1=A. Computer: OK User: Now, let us assume 2=B Computer: Alright User: What does 12122111 equate to? ...
Now, I realize this would be easy to preprogram a response to, but if you made up something to teach it each time, and the computer actually learned it, then it would be true "intelligence." This is the only way to make the turing test legitimate.
Or maybe this is a different test altogether?
As with all fantasy the rules aren't concrete physical ones, but instead the rules of the "human experience." AKA: If the characters aren't realistic it doesn't work, and tha tis the only bound on fantasy.
So... The easy way to be a famous programmer is to be a woman?
The kid is five. You need to monitor a kid of 5 like 24/7 anyways. And even if he/she even sees something "wrong" they won't understand. I generally agree with you, but what you are saying about the internet not causing any damage is more relevant to a 10 year old. Porn is porn, they will find a way to view it if they want to.
He has that Billy O charisma working for him. In the sense they both conjure up images of red butted furry mammals.
Algorithms, eh? My job is to design them, but before I came to this job I never worried about it in a professional setting. (school was another thing, since CS is a lot of algorithm analysis as a degree) Seriously, you will find most sorts and searches are already programmed for you, and are optimized. And if you work in a larger framework there will probably be an optimized sort and search already implemented for you for that specific data set.
Consensus is that people do not lie on surveys. Or the fraction who do lie is so small that it is statistically insignificant. Everyone who has taken Stats 101 knows that!
Sending those messages is of course illegal. The article specifically says that the law covered all spam, fraudulent or not, and that the courts refused to tailor it specifically to fraud. Thus it was thrown out. I don't blame them at all. I wish, however, that they would create laws to deal with malicious spamming, at least to some degree.
I never liked the slippery slope arguement. It operates on induction, but it is false induction, because there is no basis for the other "dominos" to fall. The basic fact is this. The videos shows you ways to kill other humans, in general. Not self defense, suicidal methods for killing and maiming. I am not saying this information should be banned. I just think that the information shouldn't be so accesible that a 8 year old can just click a link and see the best way to bomb something.
Yeah... because comparing videos of people "hacking" to videos on how to kill people (through suicide bombing or slitting throats) is really a fair comparison.
I don't think inaccurate information moves prices closer to their actual value any way you spin it.
They don't own all of one stock. Any investment bank has enough cash to effect the price of a stock in seconds if they want. They didn't really lose money, mind you. (Some people shorted, and some people bought when it was low) The real tragedy is that a company's stock value fluxed down due to inaccurate news and then didn't recover to where it was before inaccurate news was posted. This causes the stock to inaccurately reflect the actual value of the company.
I wonder though. Is this really a solid investment? Google has struck gold mines with the whole adword and adsense, but giving broadband to countries with little to no GDP? It is a very kind move, but it strikes me as irresponsible. Of course, even if it does pan out in the end wouldn't they have seen a larger return investing in a BRIC country? It is really a question of: Why take the greater risk?
What kind of power supply does this require? (The pump and such has to draw a bit of power?) Does the pump run all the time? Or just when it is needed? How much more power does this draw then a normal card?
Having done small amounts of research myself I know how easy it is to make up things (or make extended guesses) when looking in to theories/hypothosis. However, in this case, the potential damage done (medical research) is massive; people could have died. The issue of wasting money is kind of mute in comparison to damage caused by the information spread by this researcher. Sure, I can understand the first initial reaction of: "That is a lot of money!" But, in reality, this isn't that much money for a research grant. It really is disturbing that someone would falsify medical research in order to secure grant funds. The physist claiming a theory about gravity and proving himself wrong is one thing, the doctor providing medication on a fake basis is another.
This sort of litigation is the cause of so many injustices. While we waste our time smacking down these stupid cases all sorts of money/time/attention is drawn away from more important issues. I mean, there are real and serious things to be debated and decided by those who work for and with our justice department, but we spend all our time arguing about something that doesn't particularily make sense at all. Since when does google HAVE to do anything? People have a choice in using their system or not; they are not a monopoly. (Isn't MSN the most used search engine still?)
Well, if google is running a "beta" ISP somewhere I can quickly imagine them becoming a real ISP. I mean, look at google mail. It is still in "beta" but it functions 100 times more usefully than most of the webmail services I have used, and is quickly becoming popular. Just because google claims something is beta doesn't mean it isn't actually working, is what I am trying to say. I am also not sure of the needs that google has with fiber to begin with. I work at a small private college (5k students) and we use thousands of miles of fiber. I would imagine google would need a little more than us. How much did they actually buy?
Sure, distraction is bad when you are driving. However, this study seems pretty strange. I mean, first off, 40 people isn't a large sample size. Also, the types of people who are driving and using the cell phone matter. If you have some bimbo girl riding about at 100 mph talking to her friends (even faster then she drives) about her new purse she is going to be a little more likely to crash then oh... a 40 year old saying goodnight to his/her spouse on a business trip. Futhermore, cause and effect are not clear here. I think that the factors that go in to the "cause" of an accident are a little more complex then: "She was talking on her cell!" The difference between this and drunk driving is the ability to measure the impairment. When you are drunk your level of impairment can be scientifically quantified. Your vision has decreased by X, and you r reaction time by Y, ect... When you talk on the phone this isn't the case, because some people are more easily distracted then others. In reality, if you are talking with a hands free set then you are just as impaired if you are talking to a person sitting next to you. So, by that logic, if you are going to make hands free sets illegal you should make talking in the car illegal.