So was there too much irony in my post? In your terms: bad editting means the readers lose, which could (eventually) drive them away from Slashdot, by which nobody would gain anything.
What is the last sentence doing there: "Or will it mean film studios can stop..."? It's clear from the preceding text that that (i.e., copy while travelling, not copy afterwards) is one of the potential uses. So it's completely redundant. At the same time, the implicature of this particular phrase suggests Something Bad: Big Companies are trying to stop You from your Right To Download, or something akin, implying that these "researchers" have hidden agendas and are enemies of open source, Linux, Ruby, Apache and probably of world peace. That's of course complete and utter nonsense, so the last sentence should have been cut out by the editor. Why didn't that happen? And what's the link to www.absolutegadget.com doing there? Who gains by putting this link on the/. front page?
And (I forgot to add) they measured the predictive capacity of a part of the data on a test set that was kept separate. So the training procedure didn't see all the data and identified unknown data, although the 71% accuracy is not great (compared to the 50% base rate).
But you know they are heavily shielded, don't you? On the outside, there are magnets with reversed poles, so the field drops quickly. When there wouldn't be any shielding, there would be a 0.01T field at up to 30m. And that's really, really strong. Strong enough to attract light iron objects. Did you ever see the security videos?
No, the presentation of the numbers and thus the action was delayed (between 3 and 10 seconds), and if they did it properly (the article is skimpy on details) they only used the BOLD response before the numbers were shown.
They did differentiate between the two operations subjects were planning to perform. Not with great accuracy, but they did. Check the article for yourself: it's in Current Biology, in a section with papers accepted but not yet published. Haynes is the first author.
I suppose you've never been near an MRI scanner. I can assure you that not shielding them off makes it practically impossible to be within a radius of 10 to 20 meters. You would even faint if you were walking too fast in their neighbourhood (the quick change in magnetic field will make a current run through your brain).
But there are also other ways of reading brain signals...
By the way, the "prediction" accuracy was 71% at best (after choosing the optimal site for this particular experiment), whereas a random selection will have a 50% accuracy. So that's not really a great leap.
Are they typical neuroscientists? Perhaps in your country the train engineers are running MRI scanners, but in my country -- a bit more traditional, perhaps -- they just build trains. Mon Dieu, que idiotez, diese Amerikaner!
I guess you've seen a BCI (Brain-Computer Interface) based on EEG (Electro-Encephalogram). You attach a few electrodes to the skull or put a cap or band with electrodes around the head and that allows you to measure some of the electrical activity in the brain, but only of the outmost layer (the cortex). That signal is very, very noisy.
For a BCI to do what you want, you have to train. It takes many hours to get an accurate decoding even of a few simple actions (up, down, left, right, enter, that's just five actions). The idea is that you learn to use a part of your brain to provide signals that are strong enough to be distinguished from the electrical chaos that makes up the rest of your thinking. It's almost like training a muscle. So in that sense it is artificial brain activity.
Although I'm not completely unsympathetic to the idea of a third world underdog vs a giant pharma-tech company, the truth is that the virus isn't Indonesia's Intellectual Property. Did the country conceive it? If so, they should be punished for releasing it. As a matter of fact, the outbreak didn't start in Indonesia. If I remember properly, the first reports (of people dying) came from Cambodia or Vietnam, albeit from a different strand, and the Indonesian strand has been found in other countries as well.
The real point here is that everybody resents big companies making money by holding our health hostage. But that's how life is at this moment. The only way we can change that is by using our votes and our wallets.
You're a troll. These people have their own funding and Germany/Europe has more. The Max Planck Gesellschaft has set up institutes likes this to avoid this kind of problem. Just check the facts before submitting. And Siemens and Philips build quite nice scanners, thank you.
No, your precious DHS dollars are safe. Go and fund your own research now.
Ok, I work as a post-doc in the field and actually know the work of Haynes. They are not predicting someone's actions. Their fMRI data can distinguish between their subjects' state of mind after the fact. There are several fundamental differences between this experimental set-up and real action prediction. One of them is that fMRI doesn't yield a reliable signal until 6 seconds after the decision has been made. Another one is that in this experiment the action was carried out, i.e. it was not a hidden intention. In this experiment, subjects had to hold on to their decision during a variable time; i.e., they had to wait for a signal before taking the action, but they had to perform it. So in reality, the experiment looked at the process of holding on to a certain intention, and that intention was rather artificial. And it still cannot be done without knowing the outcome of the action, i.e., a large number of samples has to be taken with the subject's cooperation before any "prediction" can be made. So I would conclude that, interesting as the outcome may be, the article is highly exaggerated.
You are not contradicting him: you are repeating his point...
In this thread, someone wrote: "You're forgetting the minor detail that your computer is also *doing* about 200 times more than it was when running Word Perfect 5." That's not true. It's not doing 200 times as much. I had a Macintosh SE at 8 MHz running Word 5.1a and it had about the same functionality as Word now has. Auto-formatting and spell-checking do not take 2GHz (I've actually written part of a MS Word spell check plugin, I know what I'm talking about), and the other things are done off-line.
I also have the feeling that programmers get away with badly written, top-heavy software because modern systems are so fast. A lot of the performance is wasted in abstraction layers that in the end offer no possibility for re-use nor improve reliability. If you've ever tried to interface from a CORBA/.NET/whatever-MS-calls-it DLL to some higher application that uses Visual Basic for an API, you understand me...
A good high-level language and a library is not enough to build an adequate mail client or anything that comes even close to an office suite. Even more skills are needed to integrate everything (including system configuration) in a user-friendly manner. I also know a few great examples of single person development, but that's not enough to swing the balance. Everything, from network interfacing and power management to browsing and printing needs to be painless and user-friendly.
I am a sort of Apple fanboy and I do write applications in Cocoa, but user interface design doesn't stop nor start with placing a few buttons and text fields in a window. Example: I'm sure that MS Word's dialog for editting paragraph properties couldn't have been layed out better, but the whole experience of going through the menu and one or two buttons just to open a non-modal dialog that contains dozens of options, some of which quite obscure, is not user-friendly. UI design starts with the interaction; design comes later.
And about Perl we heavily disagree: I think it's one of the worst languages in existence...
It's impossible that the company has total control of what comes up for every query. There are simply too many queries and too many pages. Anybody who wants to control that needs hundreds of billions of control knobs. Google would need a lot of employees to twist and turn them.
BTW, the suggested approach was tried by AskJeeves and failed. They needed too many editors to edit page ranks per keyword and combinations. And they covered not even 1% of the pages Google covers.
He didn't leave it out. It may surprise you, but these 7 points were actually in the article. It surprised me, because I thought the article would mention things like "The subject line of your e-mail just says Hello Mr/Mrs X". But no, the article mentions these 7 points that indeed are highly indicative of a spammer. The title was clearly wrongly chosen. But your 8th point could certainly be added...
Well, I can assure you my GP doesn't work like that, and many Spanish GPs still work with paper files. Not to mention that in The Netherlands the GP's work has been heavily politicized (prescription of medicines has been severly limited), and that Spain has two health care systems working along-side each other. And sometimes ticking boxes is not enough.
But what exactly is the great advantage of such a system (this particular one)?
You're just a total egotistic w*nker, aren't you? If you paid a little bit more, you could educate these poor people and their children in decent schools instead of forcing them to accept very badly paid jobs under conditions which I don't find fit for my dog, just so you can get your coffee cheap at WalMart, or send them to institutions that are grimmer than prisons, yet have a level of violence which the average Iraqi would consider excessive. Your way of thinking has brought your country to a fiscal and moral bankrupcy.
How in the hell is a Spanish physician going to understand my Dutch GP's notes? And such a system has so little potential use and so many ways of ending up on http://www.dailywtf.com/, that the mind boggles at the thought of hundreds of millions of being wasted on another prestigious EU project.
This is just a wild guess, but it smells very French to me.
"By use of the word 'nit-picking" I am rather obviously referring to statements that affect nothing." In that case, I was not nit-picking, and you are just playing with words, as the rest of your comments seem to be doing. I'll repeat it again: in logic a conclusion is only valid when all premises are valid and the derivation is valid (check Aristotle, someone whom I think you hold in high regard; there are more modern views on logic, but they only add stipulations). That's not a statement that affects nothing, since my objections invalidate the premise that a very small probability of event X requires an infinite number of trails before X happens.
The qualification gibberish refers to comments such as "even a prime". There is nothing special about prime numbers that would make them a better candidate for empirical probabilities.
"... I withdraw from discussion..." That's plainly cowardly. You can't win, so you leave.
This discussion is not about winning, it's about an incorrectness in your objection to a certain statement. If you want to assert that the premise of non-existence of God disqualifies it, fine, that's a matter of belief. But other aspects are a matter of ratio and can be discussed.
Anyway, you attribute the statement to Hawking, so why should you feel offended when I point out an error in that statement, one that you seem to resent?
Nit-picking is of course completely valid in logic statements. If one of the premises is false, the conclusion is not necessarily true.
Furthermore, you're talking gibberish. Why should 0 (obviously impossible), 1 or infinite be the only choice? We might very well be in the 349584778478478th universe. And even then, infinite is not well defined: there are multiple infinites, so which one is supposed to make most sense to the philosopher/mathematician?
Furthermore, if the probability of life in a random universe were zero, then an infinite number of universes would not be enough to bring life.
So, your original statement is nonsense.
BTW, what's the relation with a Chinese professor having a go a SHA-1?
The probability is very small in a random universe, not any one you pick. And it still only implies a finite number of universes. And the correct spelling is "astronomically", which however means extremely large. You probably meant "infinitesimally"
That is 1 for school masterism, 0 for responding without thinking.
That's not significant at all (p = 0.2734375) and it certainly doesn't have enough power. Of course I would take surgery, since estimates favor survival, but I would not take it as a guarantee to survive. Testing should continue. Furthermore, the example you give is unnatural, since there would have been only 8 people in the world with this "physical ailment", apparently over a long time span (otherwise, you wouldn't be able to judge survival), which makes me think you just made it up.
An about the 2 people example: I'm sure it generalizes over the judgement capabilities of these 2 people, but that doesn't generalize over the entire population. And women would not dress better on the off-chance of meeting one of these two; evolution would have found a better allocation of resources...
Slashdot Mirror
So was there too much irony in my post? In your terms: bad editting means the readers lose, which could (eventually) drive them away from Slashdot, by which nobody would gain anything.
What is the last sentence doing there: "Or will it mean film studios can stop ..."? It's clear from the preceding text that that (i.e., copy while travelling, not copy afterwards) is one of the potential uses. So it's completely redundant. At the same time, the implicature of this particular phrase suggests Something Bad: Big Companies are trying to stop You from your Right To Download, or something akin, implying that these "researchers" have hidden agendas and are enemies of open source, Linux, Ruby, Apache and probably of world peace. That's of course complete and utter nonsense, so the last sentence should have been cut out by the editor. Why didn't that happen? And what's the link to www.absolutegadget.com doing there? Who gains by putting this link on the /. front page?
And (I forgot to add) they measured the predictive capacity of a part of the data on a test set that was kept separate. So the training procedure didn't see all the data and identified unknown data, although the 71% accuracy is not great (compared to the 50% base rate).
But you know they are heavily shielded, don't you? On the outside, there are magnets with reversed poles, so the field drops quickly. When there wouldn't be any shielding, there would be a 0.01T field at up to 30m. And that's really, really strong. Strong enough to attract light iron objects. Did you ever see the security videos?
No, the presentation of the numbers and thus the action was delayed (between 3 and 10 seconds), and if they did it properly (the article is skimpy on details) they only used the BOLD response before the numbers were shown.
They did differentiate between the two operations subjects were planning to perform. Not with great accuracy, but they did. Check the article for yourself: it's in Current Biology, in a section with papers accepted but not yet published. Haynes is the first author.
I suppose you've never been near an MRI scanner. I can assure you that not shielding them off makes it practically impossible to be within a radius of 10 to 20 meters. You would even faint if you were walking too fast in their neighbourhood (the quick change in magnetic field will make a current run through your brain).
But there are also other ways of reading brain signals...
By the way, the "prediction" accuracy was 71% at best (after choosing the optimal site for this particular experiment), whereas a random selection will have a 50% accuracy. So that's not really a great leap.
Are they typical neuroscientists? Perhaps in your country the train engineers are running MRI scanners, but in my country -- a bit more traditional, perhaps -- they just build trains. Mon Dieu, que idiotez, diese Amerikaner!
Nope, but you're not far off.
I guess you've seen a BCI (Brain-Computer Interface) based on EEG (Electro-Encephalogram). You attach a few electrodes to the skull or put a cap or band with electrodes around the head and that allows you to measure some of the electrical activity in the brain, but only of the outmost layer (the cortex). That signal is very, very noisy.
For a BCI to do what you want, you have to train. It takes many hours to get an accurate decoding even of a few simple actions (up, down, left, right, enter, that's just five actions). The idea is that you learn to use a part of your brain to provide signals that are strong enough to be distinguished from the electrical chaos that makes up the rest of your thinking. It's almost like training a muscle. So in that sense it is artificial brain activity.
Although I'm not completely unsympathetic to the idea of a third world underdog vs a giant pharma-tech company, the truth is that the virus isn't Indonesia's Intellectual Property. Did the country conceive it? If so, they should be punished for releasing it. As a matter of fact, the outbreak didn't start in Indonesia. If I remember properly, the first reports (of people dying) came from Cambodia or Vietnam, albeit from a different strand, and the Indonesian strand has been found in other countries as well.
The real point here is that everybody resents big companies making money by holding our health hostage. But that's how life is at this moment. The only way we can change that is by using our votes and our wallets.
You're a troll. These people have their own funding and Germany/Europe has more. The Max Planck Gesellschaft has set up institutes likes this to avoid this kind of problem. Just check the facts before submitting. And Siemens and Philips build quite nice scanners, thank you.
No, your precious DHS dollars are safe. Go and fund your own research now.
Ok, I work as a post-doc in the field and actually know the work of Haynes. They are not predicting someone's actions. Their fMRI data can distinguish between their subjects' state of mind after the fact. There are several fundamental differences between this experimental set-up and real action prediction. One of them is that fMRI doesn't yield a reliable signal until 6 seconds after the decision has been made. Another one is that in this experiment the action was carried out, i.e. it was not a hidden intention. In this experiment, subjects had to hold on to their decision during a variable time; i.e., they had to wait for a signal before taking the action, but they had to perform it. So in reality, the experiment looked at the process of holding on to a certain intention, and that intention was rather artificial. And it still cannot be done without knowing the outcome of the action, i.e., a large number of samples has to be taken with the subject's cooperation before any "prediction" can be made. So I would conclude that, interesting as the outcome may be, the article is highly exaggerated.
You are not contradicting him: you are repeating his point...
In this thread, someone wrote: "You're forgetting the minor detail that your computer is also *doing* about 200 times more than it was when running Word Perfect 5." That's not true. It's not doing 200 times as much. I had a Macintosh SE at 8 MHz running Word 5.1a and it had about the same functionality as Word now has. Auto-formatting and spell-checking do not take 2GHz (I've actually written part of a MS Word spell check plugin, I know what I'm talking about), and the other things are done off-line.
I also have the feeling that programmers get away with badly written, top-heavy software because modern systems are so fast. A lot of the performance is wasted in abstraction layers that in the end offer no possibility for re-use nor improve reliability. If you've ever tried to interface from a CORBA/.NET/whatever-MS-calls-it DLL to some higher application that uses Visual Basic for an API, you understand me...
A good high-level language and a library is not enough to build an adequate mail client or anything that comes even close to an office suite. Even more skills are needed to integrate everything (including system configuration) in a user-friendly manner. I also know a few great examples of single person development, but that's not enough to swing the balance. Everything, from network interfacing and power management to browsing and printing needs to be painless and user-friendly.
I am a sort of Apple fanboy and I do write applications in Cocoa, but user interface design doesn't stop nor start with placing a few buttons and text fields in a window. Example: I'm sure that MS Word's dialog for editting paragraph properties couldn't have been layed out better, but the whole experience of going through the menu and one or two buttons just to open a non-modal dialog that contains dozens of options, some of which quite obscure, is not user-friendly. UI design starts with the interaction; design comes later.
And about Perl we heavily disagree: I think it's one of the worst languages in existence...
It's impossible that the company has total control of what comes up for every query. There are simply too many queries and too many pages. Anybody who wants to control that needs hundreds of billions of control knobs. Google would need a lot of employees to twist and turn them.
BTW, the suggested approach was tried by AskJeeves and failed. They needed too many editors to edit page ranks per keyword and combinations. And they covered not even 1% of the pages Google covers.
He didn't leave it out. It may surprise you, but these 7 points were actually in the article. It surprised me, because I thought the article would mention things like "The subject line of your e-mail just says Hello Mr/Mrs X". But no, the article mentions these 7 points that indeed are highly indicative of a spammer. The title was clearly wrongly chosen. But your 8th point could certainly be added...
Well, I can assure you my GP doesn't work like that, and many Spanish GPs still work with paper files. Not to mention that in The Netherlands the GP's work has been heavily politicized (prescription of medicines has been severly limited), and that Spain has two health care systems working along-side each other. And sometimes ticking boxes is not enough.
But what exactly is the great advantage of such a system (this particular one)?
You're just a total egotistic w*nker, aren't you? If you paid a little bit more, you could educate these poor people and their children in decent schools instead of forcing them to accept very badly paid jobs under conditions which I don't find fit for my dog, just so you can get your coffee cheap at WalMart, or send them to institutions that are grimmer than prisons, yet have a level of violence which the average Iraqi would consider excessive. Your way of thinking has brought your country to a fiscal and moral bankrupcy.
How in the hell is a Spanish physician going to understand my Dutch GP's notes? And such a system has so little potential use and so many ways of ending up on http://www.dailywtf.com/, that the mind boggles at the thought of hundreds of millions of being wasted on another prestigious EU project.
This is just a wild guess, but it smells very French to me.
"By use of the word 'nit-picking" I am rather obviously referring to statements that affect nothing." In that case, I was not nit-picking, and you are just playing with words, as the rest of your comments seem to be doing. I'll repeat it again: in logic a conclusion is only valid when all premises are valid and the derivation is valid (check Aristotle, someone whom I think you hold in high regard; there are more modern views on logic, but they only add stipulations). That's not a statement that affects nothing, since my objections invalidate the premise that a very small probability of event X requires an infinite number of trails before X happens.
..." That's plainly cowardly. You can't win, so you leave.
The qualification gibberish refers to comments such as "even a prime". There is nothing special about prime numbers that would make them a better candidate for empirical probabilities.
"... I withdraw from discussion
This discussion is not about winning, it's about an incorrectness in your objection to a certain statement. If you want to assert that the premise of non-existence of God disqualifies it, fine, that's a matter of belief. But other aspects are a matter of ratio and can be discussed.
Anyway, you attribute the statement to Hawking, so why should you feel offended when I point out an error in that statement, one that you seem to resent?
Nit-picking is of course completely valid in logic statements. If one of the premises is false, the conclusion is not necessarily true.
Furthermore, you're talking gibberish. Why should 0 (obviously impossible), 1 or infinite be the only choice? We might very well be in the 349584778478478th universe. And even then, infinite is not well defined: there are multiple infinites, so which one is supposed to make most sense to the philosopher/mathematician?
Furthermore, if the probability of life in a random universe were zero, then an infinite number of universes would not be enough to bring life.
So, your original statement is nonsense.
BTW, what's the relation with a Chinese professor having a go a SHA-1?
The probability is very small in a random universe, not any one you pick. And it still only implies a finite number of universes. And the correct spelling is "astronomically", which however means extremely large. You probably meant "infinitesimally"
That is 1 for school masterism, 0 for responding without thinking.
That's not significant at all (p = 0.2734375) and it certainly doesn't have enough power. Of course I would take surgery, since estimates favor survival, but I would not take it as a guarantee to survive. Testing should continue. Furthermore, the example you give is unnatural, since there would have been only 8 people in the world with this "physical ailment", apparently over a long time span (otherwise, you wouldn't be able to judge survival), which makes me think you just made it up.
An about the 2 people example: I'm sure it generalizes over the judgement capabilities of these 2 people, but that doesn't generalize over the entire population. And women would not dress better on the off-chance of meeting one of these two; evolution would have found a better allocation of resources...