The unpredictable, unacceptable delay *is* for incremental re-encoding. Modern video codecs are stunningly clever predictive miracle engines, and encoding an entire episode of something takes minutes even on good hardware. Re-encoding the last 30 seconds on a mobile device would *still* take too long for a smooth user experience, and the only alternative would be to trade off speed against vastly increased requirements in space... which mobile devices are *also* low on.
...the terms of service state that you relinquish any and all exclusive rights on whatever documents you upload, and that Alphabet can use them for any conceivable purpose whatsoever. Am I right? What do I win?
They pledged? How on Earth is this not already the law? How on Earth is this not already in their terms of service? Seriously, are these services only used by terminally naive people?
Dude, any car being designed from now on is going to be a computer. Behaviour will be changed via updates after the fact and there's nothing you can do about it.
Algorithms, unlike humans, are susceptible to a specific type of problem called an "adversarial example." These are specially designed optical illusions that fool computers[...]
In other words, just like the optical illusions that humans are notoriously susceptible to? Jesus. The phenomenon is actually somewhat interesting, but maybe you shouldn't start out with a blatant self-contradictory assertion.
No it isn't. It's inevitable. Unless you believe that human mastery of Go was somehow due to special, non-information-processing-related powers, there is no way that our superiority could last forever, since evolution works with glacial slowness while computer technology advances at breakneck speed. Everyone who professes themselves shocked that we cannot understand the inner workings of programs we ourselves wrote overlooks that we also can't understand the inner workings of the thought processes of Go grandmasters.
Remember, there is no such thing as "identity theft".
There is only fraud, committed between two parties neither of which is you. The notion that someone can "steal your identity" is a red herring invented by big companies, in the hope that this will make it sound as if it was your responsibility and you should bear the costs. It isn't - it's their responsibility to guard against fraudulent transactions and not to withdraw money from you under fraudulent circumstances. But so far they've been pretty successful in establishing the narrative that it's your fault if someone abuses the ridiculously inadequate safeguards against fraud. This is a prime example of "Establish the terms of the debate, and you've determined its outcome".
I should hope very much that they took that into account and created an artificially balanced test set, otherwise the study could indeed not be very informative in principle. But the article doesn't say and the paper is yet to appear, so we can really only reserve judgment and wait.
It gets weirder. There is evidence that banning DDT in agriculture was overall the right thing to do independent of how harmful to humans it is.
The reason is that it prevented mosquitos from developing immunity too fast through prolonged exposure, so exterminating them for malaria control kept working much longer. (This is rather like the case of using antibiotics on cattle preemptively, only this time we're doing the wrong thing).
Until now perhaps, thanks to the work of Professor Antonio Paris of St Petersburg College, Florida. Known as 266P/Christensen and 335P/Gibbs, they have never been investigated before
So in other words, pretty much the entire publishing caste, as well as the majority of the population, does not understand selection bias. That is a serious problem, but it's hardly surprising. An awful lot of the scientific studies that someone trumpets around as a confirmation of their pet worldview suffer from similar problems.
That's a common and logical-seeming wish, but reality works differently. Every language spoken by humans contains some irregularity, to the point where there is clearly an underlying reason for why perfect regularity isn't optimal for human processing. (Note that you only have difficulty with irregular verbs in foreign languages; no one forgets forms in their mother tongue, and if you're uncertain, both forms usually work equally well for speaker and listener.)
The phenomenon isn't understood completely, but it's too pervasive to be just accidental. Some aspects of it are quite well-understood. For instance, there is a reason why languages resist perfectly phonetic spelling: written text mediates between writers and readers, and while writers would prefer perfectly regular spelling, readers actually profit from a small amount of irregularity, because it allows them to use gestalt perception to recognize some words even faster than sounding them out would be. Note that the most frequent words in English tend to be those with the weirdest spelling, much like the most common verbs have the most irregular past forms.
Clearly, a huge amount of optimization has been going on to shape the language for ever-greater efficiency, at a scale that laughs at any attempt to impose a "simpler" or "better" standard in a top-down way. That doesn't prevent purists and politicians from trying, but you know how well their efforts usually turn out. I'm certain that even Esperanto would acquire a certain amount of irregularity to the extent that it was actually used prominently as a native tongue.
Two people don't get suicidally depressed in perfect sync to pull this kind of thing off. That scenario is so much more implausible that tightening the rules would be a huge win and, in fact, almost a 100% prevention.
How about you don't seal the back of the ATM but instead put vents on it and a blower continuously pushing fresh air in? If they thieves try to pump it full of explosive gas, it would blow back out.
The point of an automated teller machine is to save the considerable cost of a human teller. Requiring more always-on moving parts incurs terrible running costs that would offset that gain.
This complaint is very common. I agree with everything in it, since as a European I am just as affected.
But the solution is not for content owners to wake up and do the obvious. The era of internet-illiterate companies has passed; these days they are very well aware what the technical possibilities are and how they could profit from them. If they don't give us what they want, it is now either because they don't want to give up control, or because they can't because of tangled third-part obligations.
Remember how NVidia was always unable to open-source their graphics drivers, because there was code in them from multiple non-cooperative third parties? The media situation is like that, only much, much worse. Many companies would love to harvest these additional Dollars you and I would be willing to pay, but to do that would require extra negotiations with a near-infinite list of suppliers, who are quite content with the existing, old-fashioned, area-based license agreements. It will take another generation of technical sophistication in all parts of the supply chain to change that. And until that changes, Netflix or ABC will not sell to me and you because it wouldn't generate enough revenue to pay for the nightmare of renegotiation required.
In short, the situation is much worse than many think. Ignorance we can deal with; education does work eventually. Conflicting vested interests, not so much. We can live in hope - after all, eventually someone figured out that people would love a music player with an interface that didn't suck - but don't hold your breath.
Why would a man join this site compared with dating sites that let him see photos and don't make him jump through silly hoops?
Because the better user experience for women will attract more and more desirable ones than elsewhere? As everything in life, it's a trade-off. If a site deters a lot of the obnoxious guys that plague the internet, thus bringing the F-to-M ratio up to a sane value, a lot of the non-obnoxious ones might consider it worth to expend more effort to participate there.
Is there an updated cut-and-paste toy command line to establish quickly whether you're still vulnerable to these variants, or are the details still embargoed?
Slashdot Mirror
The unpredictable, unacceptable delay *is* for incremental re-encoding. Modern video codecs are stunningly clever predictive miracle engines, and encoding an entire episode of something takes minutes even on good hardware. Re-encoding the last 30 seconds on a mobile device would *still* take too long for a smooth user experience, and the only alternative would be to trade off speed against vastly increased requirements in space... which mobile devices are *also* low on.
...the terms of service state that you relinquish any and all exclusive rights on whatever documents you upload, and that Alphabet can use them for any conceivable purpose whatsoever. Am I right? What do I win?
The best thing about this extension is that you do not have to install it.
...yet.
They pledged? How on Earth is this not already the law? How on Earth is this not already in their terms of service? Seriously, are these services only used by terminally naive people?
If they can't justify processing my data under any of the numerous and rather broad bases, then they don't deserve to get them.
Close, but no cigar. When you have to throw the device away, then it's bricked.
Dude, any car being designed from now on is going to be a computer. Behaviour will be changed via updates after the fact and there's nothing you can do about it.
Algorithms, unlike humans, are susceptible to a specific type of problem called an "adversarial example." These are specially designed optical illusions that fool computers[...]
In other words, just like the optical illusions that humans are notoriously susceptible to? Jesus. The phenomenon is actually somewhat interesting, but maybe you shouldn't start out with a blatant self-contradictory assertion.
This is a scary path we are following.
No it isn't. It's inevitable. Unless you believe that human mastery of Go was somehow due to special, non-information-processing-related powers, there is no way that our superiority could last forever, since evolution works with glacial slowness while computer technology advances at breakneck speed. Everyone who professes themselves shocked that we cannot understand the inner workings of programs we ourselves wrote overlooks that we also can't understand the inner workings of the thought processes of Go grandmasters.
Bond villain? I thought it was established that Elon Musk is Batman?
Remember, there is no such thing as "identity theft". There is only fraud, committed between two parties neither of which is you. The notion that someone can "steal your identity" is a red herring invented by big companies, in the hope that this will make it sound as if it was your responsibility and you should bear the costs. It isn't - it's their responsibility to guard against fraudulent transactions and not to withdraw money from you under fraudulent circumstances. But so far they've been pretty successful in establishing the narrative that it's your fault if someone abuses the ridiculously inadequate safeguards against fraud. This is a prime example of "Establish the terms of the debate, and you've determined its outcome".
I should hope very much that they took that into account and created an artificially balanced test set, otherwise the study could indeed not be very informative in principle. But the article doesn't say and the paper is yet to appear, so we can really only reserve judgment and wait.
It's a sad day indeed when you have to explain the pigeonhole principle to the readers of Slashdot, of all places.
Left wing? What are you smoking? Merkel belongs to the CDU party (christian democratic union) and it is a center - right party.
Center-right for Germany. You must remember that compared to the U.S., Europe in general is radically left-wing w.r.t. social policy.
It gets weirder. There is evidence that banning DDT in agriculture was overall the right thing to do independent of how harmful to humans it is.
The reason is that it prevented mosquitos from developing immunity too fast through prolonged exposure, so exterminating them for malaria control kept working much longer. (This is rather like the case of using antibiotics on cattle preemptively, only this time we're doing the wrong thing).
What is missing here? Please fix the summary.
So in other words, pretty much the entire publishing caste, as well as the majority of the population, does not understand selection bias. That is a serious problem, but it's hardly surprising. An awful lot of the scientific studies that someone trumpets around as a confirmation of their pet worldview suffer from similar problems.
That's a common and logical-seeming wish, but reality works differently. Every language spoken by humans contains some irregularity, to the point where there is clearly an underlying reason for why perfect regularity isn't optimal for human processing. (Note that you only have difficulty with irregular verbs in foreign languages; no one forgets forms in their mother tongue, and if you're uncertain, both forms usually work equally well for speaker and listener.)
The phenomenon isn't understood completely, but it's too pervasive to be just accidental. Some aspects of it are quite well-understood. For instance, there is a reason why languages resist perfectly phonetic spelling: written text mediates between writers and readers, and while writers would prefer perfectly regular spelling, readers actually profit from a small amount of irregularity, because it allows them to use gestalt perception to recognize some words even faster than sounding them out would be. Note that the most frequent words in English tend to be those with the weirdest spelling, much like the most common verbs have the most irregular past forms.
Clearly, a huge amount of optimization has been going on to shape the language for ever-greater efficiency, at a scale that laughs at any attempt to impose a "simpler" or "better" standard in a top-down way. That doesn't prevent purists and politicians from trying, but you know how well their efforts usually turn out. I'm certain that even Esperanto would acquire a certain amount of irregularity to the extent that it was actually used prominently as a native tongue.
Two people don't get suicidally depressed in perfect sync to pull this kind of thing off. That scenario is so much more implausible that tightening the rules would be a huge win and, in fact, almost a 100% prevention.
How about you don't seal the back of the ATM but instead put vents on it and a blower continuously pushing fresh air in? If they thieves try to pump it full of explosive gas, it would blow back out.
The point of an automated teller machine is to save the considerable cost of a human teller. Requiring more always-on moving parts incurs terrible running costs that would offset that gain.
But the solution is not for content owners to wake up and do the obvious. The era of internet-illiterate companies has passed; these days they are very well aware what the technical possibilities are and how they could profit from them. If they don't give us what they want, it is now either because they don't want to give up control, or because they can't because of tangled third-part obligations.
Remember how NVidia was always unable to open-source their graphics drivers, because there was code in them from multiple non-cooperative third parties? The media situation is like that, only much, much worse. Many companies would love to harvest these additional Dollars you and I would be willing to pay, but to do that would require extra negotiations with a near-infinite list of suppliers, who are quite content with the existing, old-fashioned, area-based license agreements. It will take another generation of technical sophistication in all parts of the supply chain to change that. And until that changes, Netflix or ABC will not sell to me and you because it wouldn't generate enough revenue to pay for the nightmare of renegotiation required.
In short, the situation is much worse than many think. Ignorance we can deal with; education does work eventually. Conflicting vested interests, not so much. We can live in hope - after all, eventually someone figured out that people would love a music player with an interface that didn't suck - but don't hold your breath.
I manage Unix systems so having it be wide screen helps with longer lines.
No one can possibly read lines well that are so long as to require the current ultra-wide monitors.
Why would a man join this site compared with dating sites that let him see photos and don't make him jump through silly hoops?
Because the better user experience for women will attract more and more desirable ones than elsewhere? As everything in life, it's a trade-off. If a site deters a lot of the obnoxious guys that plague the internet, thus bringing the F-to-M ratio up to a sane value, a lot of the non-obnoxious ones might consider it worth to expend more effort to participate there.
Is there an updated cut-and-paste toy command line to establish quickly whether you're still vulnerable to these variants, or are the details still embargoed?
You wish you had always known "how to design a solution on my own time before I code a solution on company time"? Why?