It was programmed to win. If you want it to "understand the nature of its errors" that will be an additional million but they'd do it for you if it were important enough.
Actually, I expect it already is programmed to learn from its errors. It just may not be able to do that in real time while playing the game.
From what I recall reading, Watson's AI was trained on years of historic Jeopardy questions. While one possibility is that it merely stored all that as data, more likely is that it would use the questions to attempt an answer, check the real answer, and compare where the real answer appeared in its own list of possibilities, and adjust its weightings accordingly (learning from error), and then go on to the next question. If it were to have had time in the real games to do this, the nature of its errors would likely have been changed, and the two humans would likely have only made about half of what they actually did.
Either way, this is it learning from its errors - whether that's "understanding" or not really depends on whether you want to ascribe intelligence to a machine, not on what it was really doing.
This would be a core functionality for Watson in a number of ways. First off, it probably made it easier to input large amounts of data for its base knowledge of the world and comprehension of natural language. Second, IBM doesn't do this type of thing for fun. It's hard research that targets specific real-world problems. If you watched the clips on the first two days, you'd know they were targeting the health sector with this technology. Watson's successor, built off the same research, will be simply useless if it doesn't learn from its mistakes.
IBM has already spent the "additional million". In fact, I expect it's far more than that - $1m probably covers only about 5 of these developers for one year, when you take into account base salary, bonuses, benefits, and infrastructure like office space, phones, management, and physical site security, and 5 person-years of time is probably only a fraction of what this feature cost. It is important enough.
Open source architecture? I don't think so. A library of ideas is one thing, but having a general contractor build something based on the drawings of someone who is not necessarily a licensed architect/engineer, who may not even be in the country of manufacture, is just asking for liability problems. Seriously. Unless you're talking about sheds or other inconsequential buildings, I don't see this passing legal muster, either.
You still need to hire the architect and/or civil engineer to sign off on the drawings. And they'll need to crunch all the numbers. And random patches are not going to be accepted, period.
I like the idea of open-source hardware. I'm just not seeing a practical side yet.
Under these exigent circumstances, it may be possible to confess via a two-way communication device (though I'm not sure an iPhone, or any cellular phone, will work from space). But that doesn't make it an "iPhone" - just a "phone". It's still not an app.
Yes, under truly dire circumstances, confessing without a priest may be allowed. But it's still not an app.
George Burns smoked like a chimney and lived to 100. Does that disprove smoking as a risk factor for cancer and shortened life span?
Population density definitely is a factor. Homogeneity of culture probably is a bigger factor. Culture itself may be an even larger factor. Phase of the moon probably is a fairly minor factor, though I've read some curious claims that it, too, may be a factor (which allegedly explains the werewolf/full-moon mythology). Not sure how much I accept the latter one, but doubtfulness is not the same thing as nonacceptance of a possibility prior to actual proof.
(Ok, these guys are fighting against power lines at 60Hz, not cellular bands - but their obvious cluelessness on simple science and statistics makes them prime candidates for opposing such "harmful" solutions.)
Assuming you're right, which probably is stretching things, you're still wrong.
Assuming that we're comparing "32 transactions per second" to "60 thousand simultaneous transactions (over a span of time)", let's just see what that means. If we assume all 60-thousand transactions are spread out over the minimum amount of time to match the 32 transactions per second, that's 1875 seconds to handle a single transaction from each user. That's over a half hour.
I think a reasonable assumption is that 30 minutes just to get through everyone once might be a bit long, and thus we need to handle more than 32 transactions per second. I suspect that 3 minutes might be too long, so it's reasonable to me that one needs at least an order of magnitude better performance than 32 transactions per second. However, 3 seconds might be reasonable turn around, so that's two orders of magnitude better performance.
See? It's not hard. You just need to take what's in front of you, and apply it with a bit of reason, and comparisons, in this case at least, are valid.
The problem is often that they don't realise it's "broke (sic)". That is often the issue. When you bring in a physical item in more pieces than it's supposed to be, that's easy to tell that it's broken. When it's a stream of digital pulses, not so easy, unless your target is technologically aware, and not always even then. I still have problems convincing otherwise smart people to use placeholders in their SQL instead of concatenation.
You missed part of the equation. If he gets n options at a strike price ("current" price at time of grant, or sometimes less) of $m, and, at time of excersise, the shares are worth $s, then if s m, then the total "pay" is actually $n*(s-m). If the price per share has risen $10, then the holder of options gets $n*10. In good times, this can amount to a lot. Not-so-good times, not so much.
The theory is that, as CEO, he is being hired to directly impact the share price, since he's hired by the Board of Directors, and the BoD is hired (in theory) by the shareholders. Thus, to properly motivate the CEO, they tie his income directly to the delta of the stock price. Unfortunately, as pointed out above, if the CEO does poorly, he never ends up risking everything the way the company is, so it's a little lop-sided. On the other hand, these guys often live extravagantly, which leaves the $1 salary as seriously draining on their net worth. That's not very compelling to me, though. Also, as always, focus on short-term stock prices is generally bad for companies, even if (temporarily) good for shareholders. Then again, options usually have some sort of vesting period whereby the holder cannot access them for a minimum amount of time after the grant. If that's 2+ years out, that can mitigate short-term concerns. I honestly don't know what Jobs, Ellison, Palmissano, et al, get as riders on their options.
No. Those ones don't pretend to be science (sci-fi) and deliver something completely different. They're more honest about what they offer: the ability to drool over someone who is more talented than you. Or laugh at those who are pathetic but think they're talented. They're still fake, but at least they approach plausibility.
Which is close to my problem with Mythbusters. Here's a show that continues to throw around the word "science" but generally fail to deliver it. Oh, sometimes they run tests multiple times, averaging their results (where's the error bars!), but not often. And they throw around "busted" a bit too often - it's hard to prove a negative. I don't think they get all the way there. A better result would be "unlikely" most of the time. On the other hand, explosions are cool, and so if you take it as science-related entertainment, it's fine.
Which brings me back to Fringe - if you want to enjoy it as wacked entertainment, go right ahead. It's when it's called "sci-fi" that some may question it. Then again, if CSI (any of them) tried to call themselves sci-fi, they'd get the same treatment. (Instead, they call themselves a drama, which is about right, though too many people think it's real science instead of made-up crap.)
I do not like the precedent it sets for having a government agency (especially non-elected officials) run roughshod over those they regulate unilaterally. Much of the developed world has adverserial justice for a reason: to ensure both sides get heard, allowing an arbiter (judge, justice, jury, or lay arbiter) to decide which side has merit/credibility/the law on their side in the dispute.
Nor would I be in favour of the regulatees running roughshod over the regulators, as seems to be the case in some vaguely-unnamed-here industries. The law, as written by elected officials, requires regulation, and thus should be followed - if you don't like it, have it declared unconstitutional/illegal or lobby the electorate to vote for a new government.
Of course, Verizon doesn't like it, and is suing to have the arbiter (likely a judge in this case) declare it illegal (overstepping legal authority). While I hope the judge disagrees with them, I wouldn't want the FCC to become immune to challenge just because of mafia-like intimidation.
You're calling this a false comparison because the title/summary are vague, and you have an expectation of comparing modern to modern. But that's far from the only way of comparison. Often you'll see products compare model-to-model. "We've managed to squeeze 13% more mpg out of this year's vehicle" or "we've added feature X to the new version" or "we've fixed 100 more bugs this release than we did last release."
Sounds to me like the article (which, of course, I've not read, and don't plan on) is comparing the next-generation of the OLPC to the previous-generation of the OLPC. That is a fair comparison. They're not trying to get into a design debate with geeks like us. They're telling everyone that their design is different, and one of the benefits they got was a faster system.
You're right that it's not as big of a deal as the title sounds when considering the choices they could have made staying on x86. But they had other goals, met (most of) them and still managed to come out ahead of the previous offering in raw speed. As a package, it sounds like a pretty decent thing, though I don't think any of it was OLPC's design or hard work. They just took components, stuck 'em together, and measured what they ended up with. The amazing thing to me is how fast ARM is coming along speed-wise without giving up on the power consumption. Which, of course, has no bearing on OLPC, and doesn't help them sell product. So if they want to sell stuff, they gotta make it sound like it's their uniqueness in the market that is making progress rather than just being another computer store selling other companies' parts simply smooshed together.
If he had hacked in to his neighbour's Yahoo! account and posted those details everywhere, his neighbour might not even find out because there would likely be no interested partisans wishing to harass him or his family. However, when you post private details of a famous person, the amount of interest goes up, and the harm follows the interest. A lot of people disagree with Palin, and, in any group of some size, there are wackos who take it too far, and that's what happened here. Wackos got hold of the information and started harassing Palin and/or her family.
It's about the same reason why a DUI gets a lesser punishment than a DUI-causing-bodily harm which gets a lesser punishment than a DUI-causing-death. It's simple: the harm is different, thus the punishment is different.
Should Palin be required to better secure her email? Of course. Her handlers are somewhat to blame here - much like Obama, I don't expect either of them to be IT experts, but I expect that they have some on staff. However, being large targets, with sufficient pools of wackos opposed to them, it's merely a matter of time before web-based email gets hacked for either one of them, regardless of the strength of their passwords or reminder questions/answers. In Obama's case, his accounts merely need to survive his one-or-two terms in office, then interest will wane. Palin's weak security just allowed it to happen faster.
Non-issue. The government is not going to be able to convince a judge that seizing money held by a lawyer (officer of the court) is legal without some sort of evidence that the lawyer is in on the illegal activity in some manner more than merely being their lawyer. And attempting to "covertly" seize the money is not going to fly far, either, when the lawyer's day job is talking to judges.
That's very interesting. Did anyone understand the poor bloke? Did the ones who may have understood actually admit they could understand such an uninformed outburst of English?
Perhaps you're so deluded by your fan-boism that you don't realise what the/. crowd generally find distasteful about the iPhone and iPad. It's not that it's from Apple (Macs are generally still well-liked around here, partially based on the BSD kernel - though I have no use for Macs myself). It's the lock-in. In general,/. hates lock-in from any vendor. Heck, I remember even ChromeOS getting the hate-on for being so locked down, and that's from one of our favourite companies (Google) based on one of our favourite OS's (Linux). It's not Apple. It's lock-in.
If I could flip a switch to say, "I'm an advanced user, get out of my way," on either the iPhone or the iPad, I'm sure it wouldn't be nearly so bad. Some way to get arbitrary applications from arbitrary locations would solve pretty much the whole/. issue with these devices. That there isn't even a secondary app-store that is recognised by the devices in the meantime also grates on the collective consensus. If there was at least competition in stores, that would mitigate, though not eliminate, concerns. Yes, newbs would have to look in multiple places to find what they want, but since nearly everything goes through a search engine anyway, that's probably not such a big deal.
In the meantime, the disregard your attitude gets on/. is simply for your lack of stand for freedom. Instead of saying "I am happy with the device," people may be expecting you to say "I may be happy with the device, but it's too locked down for general use, so I'm going to vote against lock-in and not buy it to discourage vendors from attempting such restrictions on freedom." Maybe that's a bit harsh of a requirement, though I understand where it comes from. (Largely, a mob-mentality on some issues. Oddly,/. seems to have multiple mobs on issues, so there is likely a mob that supports you, too.)
I'm guessing here (translation: pulling it out of my ass), but I expect that while most Americans get that there is a concept of freedom which America is supposed to stand for, few realise how it affects them in day-to-day life. There was NOT a massive rejection of the Repubs after the Patriot Act. There was NOT a massive rejection of the Dems after Obamacare or the TSA's groping powers (it was pretty slight so far, and I'm betting on term #2 for Obama regardless). Why should something so minor as a phone to garner a bigger reaction?
It's not opposites attract. It's the plebs eating cake.
Slashdot Mirror
It was programmed to win. If you want it to "understand the nature of its errors" that will be an additional million but they'd do it for you if it were important enough.
Actually, I expect it already is programmed to learn from its errors. It just may not be able to do that in real time while playing the game.
From what I recall reading, Watson's AI was trained on years of historic Jeopardy questions. While one possibility is that it merely stored all that as data, more likely is that it would use the questions to attempt an answer, check the real answer, and compare where the real answer appeared in its own list of possibilities, and adjust its weightings accordingly (learning from error), and then go on to the next question. If it were to have had time in the real games to do this, the nature of its errors would likely have been changed, and the two humans would likely have only made about half of what they actually did.
Either way, this is it learning from its errors - whether that's "understanding" or not really depends on whether you want to ascribe intelligence to a machine, not on what it was really doing.
This would be a core functionality for Watson in a number of ways. First off, it probably made it easier to input large amounts of data for its base knowledge of the world and comprehension of natural language. Second, IBM doesn't do this type of thing for fun. It's hard research that targets specific real-world problems. If you watched the clips on the first two days, you'd know they were targeting the health sector with this technology. Watson's successor, built off the same research, will be simply useless if it doesn't learn from its mistakes.
IBM has already spent the "additional million". In fact, I expect it's far more than that - $1m probably covers only about 5 of these developers for one year, when you take into account base salary, bonuses, benefits, and infrastructure like office space, phones, management, and physical site security, and 5 person-years of time is probably only a fraction of what this feature cost. It is important enough.
That's only if you want to pretend to play the game. Otherwise it's H0H 0H0, and you live at the North Pole.
Open source architecture? I don't think so. A library of ideas is one thing, but having a general contractor build something based on the drawings of someone who is not necessarily a licensed architect/engineer, who may not even be in the country of manufacture, is just asking for liability problems. Seriously. Unless you're talking about sheds or other inconsequential buildings, I don't see this passing legal muster, either.
You still need to hire the architect and/or civil engineer to sign off on the drawings. And they'll need to crunch all the numbers. And random patches are not going to be accepted, period.
I like the idea of open-source hardware. I'm just not seeing a practical side yet.
Awww... you're supposed to claim it's a six syllable word, and then sit back and watch 'em try to find the missing syllable. :-)
Under these exigent circumstances, it may be possible to confess via a two-way communication device (though I'm not sure an iPhone, or any cellular phone, will work from space). But that doesn't make it an "iPhone" - just a "phone". It's still not an app.
Yes, under truly dire circumstances, confessing without a priest may be allowed. But it's still not an app.
"Factor" != "sole actor"
George Burns smoked like a chimney and lived to 100. Does that disprove smoking as a risk factor for cancer and shortened life span?
Population density definitely is a factor. Homogeneity of culture probably is a bigger factor. Culture itself may be an even larger factor. Phase of the moon probably is a fairly minor factor, though I've read some curious claims that it, too, may be a factor (which allegedly explains the werewolf/full-moon mythology). Not sure how much I accept the latter one, but doubtfulness is not the same thing as nonacceptance of a possibility prior to actual proof.
Also, it's legal and doesn't put the staff at risk.
I have some bozos living around here who might not agree.
(Ok, these guys are fighting against power lines at 60Hz, not cellular bands - but their obvious cluelessness on simple science and statistics makes them prime candidates for opposing such "harmful" solutions.)
Assuming you're right, which probably is stretching things, you're still wrong.
Assuming that we're comparing "32 transactions per second" to "60 thousand simultaneous transactions (over a span of time)", let's just see what that means. If we assume all 60-thousand transactions are spread out over the minimum amount of time to match the 32 transactions per second, that's 1875 seconds to handle a single transaction from each user. That's over a half hour.
I think a reasonable assumption is that 30 minutes just to get through everyone once might be a bit long, and thus we need to handle more than 32 transactions per second. I suspect that 3 minutes might be too long, so it's reasonable to me that one needs at least an order of magnitude better performance than 32 transactions per second. However, 3 seconds might be reasonable turn around, so that's two orders of magnitude better performance.
See? It's not hard. You just need to take what's in front of you, and apply it with a bit of reason, and comparisons, in this case at least, are valid.
That's not a good reason to use telnet. That's a good reason not to use Godaddy.
(Using dreamhost.com here, and I use ssh and rsync-over-ssh to do all of that... I wonder if sshfs would work, I imagine it would.)
"Don't fix what ain't broke!"
The problem is often that they don't realise it's "broke (sic)". That is often the issue. When you bring in a physical item in more pieces than it's supposed to be, that's easy to tell that it's broken. When it's a stream of digital pulses, not so easy, unless your target is technologically aware, and not always even then. I still have problems convincing otherwise smart people to use placeholders in their SQL instead of concatenation.
None of these guys would be transparent enough to just stand up and give a speech where they say what they really think.
Fixed that for you. I promise, if I ever was elected to office, I'd be 100% transparent about my nefarious plans.
I really wish you'd stop beating about the bush and tell us how you really feel.
You missed part of the equation. If he gets n options at a strike price ("current" price at time of grant, or sometimes less) of $m, and, at time of excersise, the shares are worth $s, then if s m, then the total "pay" is actually $n*(s-m). If the price per share has risen $10, then the holder of options gets $n*10. In good times, this can amount to a lot. Not-so-good times, not so much.
The theory is that, as CEO, he is being hired to directly impact the share price, since he's hired by the Board of Directors, and the BoD is hired (in theory) by the shareholders. Thus, to properly motivate the CEO, they tie his income directly to the delta of the stock price. Unfortunately, as pointed out above, if the CEO does poorly, he never ends up risking everything the way the company is, so it's a little lop-sided. On the other hand, these guys often live extravagantly, which leaves the $1 salary as seriously draining on their net worth. That's not very compelling to me, though. Also, as always, focus on short-term stock prices is generally bad for companies, even if (temporarily) good for shareholders. Then again, options usually have some sort of vesting period whereby the holder cannot access them for a minimum amount of time after the grant. If that's 2+ years out, that can mitigate short-term concerns. I honestly don't know what Jobs, Ellison, Palmissano, et al, get as riders on their options.
You and everyone else in Denver.
No. Those ones don't pretend to be science (sci-fi) and deliver something completely different. They're more honest about what they offer: the ability to drool over someone who is more talented than you. Or laugh at those who are pathetic but think they're talented. They're still fake, but at least they approach plausibility.
Which is close to my problem with Mythbusters. Here's a show that continues to throw around the word "science" but generally fail to deliver it. Oh, sometimes they run tests multiple times, averaging their results (where's the error bars!), but not often. And they throw around "busted" a bit too often - it's hard to prove a negative. I don't think they get all the way there. A better result would be "unlikely" most of the time. On the other hand, explosions are cool, and so if you take it as science-related entertainment, it's fine.
Which brings me back to Fringe - if you want to enjoy it as wacked entertainment, go right ahead. It's when it's called "sci-fi" that some may question it. Then again, if CSI (any of them) tried to call themselves sci-fi, they'd get the same treatment. (Instead, they call themselves a drama, which is about right, though too many people think it's real science instead of made-up crap.)
I like the sweet justice of your proposal.
I do not like the precedent it sets for having a government agency (especially non-elected officials) run roughshod over those they regulate unilaterally. Much of the developed world has adverserial justice for a reason: to ensure both sides get heard, allowing an arbiter (judge, justice, jury, or lay arbiter) to decide which side has merit/credibility/the law on their side in the dispute.
Nor would I be in favour of the regulatees running roughshod over the regulators, as seems to be the case in some vaguely-unnamed-here industries. The law, as written by elected officials, requires regulation, and thus should be followed - if you don't like it, have it declared unconstitutional/illegal or lobby the electorate to vote for a new government.
Of course, Verizon doesn't like it, and is suing to have the arbiter (likely a judge in this case) declare it illegal (overstepping legal authority). While I hope the judge disagrees with them, I wouldn't want the FCC to become immune to challenge just because of mafia-like intimidation.
That's why I run Windows XP in a VirtualBox session, so that my Linux box, too, can share in the richness that is malware.
You would, too, if you had "more bugs than a bait shop".
Sorry, I've been rewatching the original Tron. I couldn't help it.
You're calling this a false comparison because the title/summary are vague, and you have an expectation of comparing modern to modern. But that's far from the only way of comparison. Often you'll see products compare model-to-model. "We've managed to squeeze 13% more mpg out of this year's vehicle" or "we've added feature X to the new version" or "we've fixed 100 more bugs this release than we did last release."
Sounds to me like the article (which, of course, I've not read, and don't plan on) is comparing the next-generation of the OLPC to the previous-generation of the OLPC. That is a fair comparison. They're not trying to get into a design debate with geeks like us. They're telling everyone that their design is different, and one of the benefits they got was a faster system.
You're right that it's not as big of a deal as the title sounds when considering the choices they could have made staying on x86. But they had other goals, met (most of) them and still managed to come out ahead of the previous offering in raw speed. As a package, it sounds like a pretty decent thing, though I don't think any of it was OLPC's design or hard work. They just took components, stuck 'em together, and measured what they ended up with. The amazing thing to me is how fast ARM is coming along speed-wise without giving up on the power consumption. Which, of course, has no bearing on OLPC, and doesn't help them sell product. So if they want to sell stuff, they gotta make it sound like it's their uniqueness in the market that is making progress rather than just being another computer store selling other companies' parts simply smooshed together.
Legally speaking, 16 and 17 are children. When talking about the law, we have to use legal definitions. Full stop.
If he had hacked in to his neighbour's Yahoo! account and posted those details everywhere, his neighbour might not even find out because there would likely be no interested partisans wishing to harass him or his family. However, when you post private details of a famous person, the amount of interest goes up, and the harm follows the interest. A lot of people disagree with Palin, and, in any group of some size, there are wackos who take it too far, and that's what happened here. Wackos got hold of the information and started harassing Palin and/or her family.
It's about the same reason why a DUI gets a lesser punishment than a DUI-causing-bodily harm which gets a lesser punishment than a DUI-causing-death. It's simple: the harm is different, thus the punishment is different.
Should Palin be required to better secure her email? Of course. Her handlers are somewhat to blame here - much like Obama, I don't expect either of them to be IT experts, but I expect that they have some on staff. However, being large targets, with sufficient pools of wackos opposed to them, it's merely a matter of time before web-based email gets hacked for either one of them, regardless of the strength of their passwords or reminder questions/answers. In Obama's case, his accounts merely need to survive his one-or-two terms in office, then interest will wane. Palin's weak security just allowed it to happen faster.
Non-issue. The government is not going to be able to convince a judge that seizing money held by a lawyer (officer of the court) is legal without some sort of evidence that the lawyer is in on the illegal activity in some manner more than merely being their lawyer. And attempting to "covertly" seize the money is not going to fly far, either, when the lawyer's day job is talking to judges.
That's very interesting. Did anyone understand the poor bloke? Did the ones who may have understood actually admit they could understand such an uninformed outburst of English?
Perhaps you're so deluded by your fan-boism that you don't realise what the /. crowd generally find distasteful about the iPhone and iPad. It's not that it's from Apple (Macs are generally still well-liked around here, partially based on the BSD kernel - though I have no use for Macs myself). It's the lock-in. In general, /. hates lock-in from any vendor. Heck, I remember even ChromeOS getting the hate-on for being so locked down, and that's from one of our favourite companies (Google) based on one of our favourite OS's (Linux). It's not Apple. It's lock-in.
If I could flip a switch to say, "I'm an advanced user, get out of my way," on either the iPhone or the iPad, I'm sure it wouldn't be nearly so bad. Some way to get arbitrary applications from arbitrary locations would solve pretty much the whole /. issue with these devices. That there isn't even a secondary app-store that is recognised by the devices in the meantime also grates on the collective consensus. If there was at least competition in stores, that would mitigate, though not eliminate, concerns. Yes, newbs would have to look in multiple places to find what they want, but since nearly everything goes through a search engine anyway, that's probably not such a big deal.
In the meantime, the disregard your attitude gets on /. is simply for your lack of stand for freedom. Instead of saying "I am happy with the device," people may be expecting you to say "I may be happy with the device, but it's too locked down for general use, so I'm going to vote against lock-in and not buy it to discourage vendors from attempting such restrictions on freedom." Maybe that's a bit harsh of a requirement, though I understand where it comes from. (Largely, a mob-mentality on some issues. Oddly, /. seems to have multiple mobs on issues, so there is likely a mob that supports you, too.)
I'm guessing here (translation: pulling it out of my ass), but I expect that while most Americans get that there is a concept of freedom which America is supposed to stand for, few realise how it affects them in day-to-day life. There was NOT a massive rejection of the Repubs after the Patriot Act. There was NOT a massive rejection of the Dems after Obamacare or the TSA's groping powers (it was pretty slight so far, and I'm betting on term #2 for Obama regardless). Why should something so minor as a phone to garner a bigger reaction?
It's not opposites attract. It's the plebs eating cake.