I don't see any advertisement on facebook at all, and I'm just using Adblock Edge + the rule facebook.com##div[class="ego_column"]. SocialFixer does more than just block advertisements, but if all you want is to get rid of then adblock already does a very good job at it.
Examples of this working are x264, libavcodec and mplayer, for example. All of these probably break large numbers of patents and are quite high profile and I'm sure they are a thorn in the side for the video and audio format cartels. But they are doing just fine, and have been doing so for a long time. If they had been closed one-man projects by somebody in the USA, they life expectancy would probably be much shorter.
This was one of the most insightful comments I've read on slashdot.
And the success of kickstarter shows that we don't need a huge change in people's mindsets for this to work right now. (Too bad that most kickstarter projects still insist on enforcing copyright even after they have been fully funded from the beginning, though. They could at least add "free software" as a stretch goal.)
I do this occasionally, when my internet connection is too flaky. I don't know anybody else who does. I'm sure less than 1% of youtube's users has ever saved a youtube-video to disk. I doubt google notices any impact from this at all. The reason for any arms race is paranoia, not actual risk.
What I and a much greater (but still small) number of people do, which probably *is* noticable for google, is to block advertisements on youtube. It works perfectly out of the box, and removes not only the general advertisements on the page, but also the video advertisements. I was disgusted recently when I tried to use youtube on a friend's computer, and had to wait through 10 seconds of an obnoxious, unseekable commercial before getting to the actual video.
Needless to say, I don't feel bad at all for using youtube this way. This is like a restaurant which gives you a side dish of heroin even though you didn't order it, and then gets upset when you leave it behind and only eat the main dish. Sure some people like heroin, but sensible people recognize that it is unhealthy and detracts from the experience of the real food.
If enough people start blocking advertisement in youtube for google to shut it down, then that would be a loss - I think youtube is filling an important niche. But it should be possible to fill this niche in other ways too. BitTorrent demonstrated how peer-to-peer distribution can scale cheaply for a relatively low number of relatively large files. It is not suitable as a youtube-replacements, but it would probably be possible to design a different distributed system that is. Something like freenet, but without the huge overhead of anonymous routing. That isn't necessary right now due to youtube filling the niche. But if youtube were to go under, the niche itself wouldn't disappear.
So to answer your question. yes, you could say that "Canada was still frozen while Greenland was basking in warmth". Though temperatures slightly elevated in some parts of Canada, most of it was cold. And none of them were anywhere near as hot as they are now.
I, too, work in HPC computing, and while I found "codes" very jarring to begin with, I've learned to live with it. I am not sure the "code" vs. "codes" issue it is more grammatically problematic than "people" vs. "peoples". A people (countable) is made up of people (uncountable). Similarly "a code" (countable, but nonstandard) is made up of code (uncountable). Personally I would use "a program" or "a library" instead of "a code", though.
Another related issue is whether "data" is countable or not. I'm used to it being uncountable, with there being more or less of it, but not "several data". But scientific journals in my field prefer the countable version "a datum", "several data", which is arguably more historically correct. That, too, took some getting used to.
A capacitance fingerprint reader leverages a handy property of your skin: The outer layer of your skin (your dermis), where your fingerprint is, is non-conductive, while the subdermal layer behind it is conductive. When you touch the iPhone’s fingerprint sensor, it measures the minuscule differences in conductivity caused by the raised parts of your fingerprint, and it uses those measurements to form an image..
A capacitor works by having an insulator sandwitched between two conductors. The thinner the insulator is, the higher the capacitance. In the case of a capacitive fingerprint reader, the conductors are the reader itself on one side, and the subdermal layers on the other side. In between them, the skin works as an insulator. Hence, by measuring the capacitance, one is effectively measuring the thickness of the skin. I.e. the pattern of ridges and valleys visible on your fingers. This is the layer you claimed wasn't measured in the first place.
What is your source for claiming that the sensor reads a different pattern than the normal fingerprints you leave behind? A capacitive fingerprint reader works by measuring the difference in capacitance between the ridges and valleys of your fingerprint. In the ridges, the distance to the more conductive layers beneath the skin (the sub-dermal layers you've heard about) is greater than in the valleys, which gives these regions higher capacitance. I guess the pattern you get this way could be different from the visible fingerprint if the underside of the skin has a significant, different pattern than the overside, but I have not heard that that is supposed to be the case.
To simplify things a bit, the much touted sub-dermal layers work as a sort of capacitive back-light which highlights the differences in thickness of the fingerprint above it. It is, to the best of my knowledge, simply another way of measuring the same fingerprint we see when we look at our fingers.
The source you site seems to be saying the opposite of what you claim:
When you touch the iPhone’s fingerprint sensor, it measures the minuscule differences in conductivity caused by the raised parts of your fingerprint, and it uses those measurements to form an image..
Those raised parts of the fingerprint are exactly the ones that deposit fat stains on every surface you touch.
Of course, it is possible that the macworld article is misleading, and that the fingerprint reader reads some other pattern after all. If so, it would be nice to see a source that backs that up. This has been brought up in previous slashdot discussions too, but I have never seen any evidence backing it up, even after explicitly asking for it.
While I think this particular project (and the one to mars) are unrealistic due to not being robustly funded and planned, I think the concept that some humans may be willing to greatly shorten their lives in order to go off and explore something new shows that there is still hope for humanity. Isn't it nice to hear that some are curious enough and adventurous enough to do this, instead of simply being focused on entertainment and producing more humans?
I would not be brave enough to do this myself, even if I actually believed these projects were serious, but I wouldn't belittle those who are willing to sacrifice themselves like this. It seems to be implicit in your post that people only have one objective: to be as comfortable as possible, and so the only reason they might have to leave the Earth is if it is too uncomfortable here. I.e. no nobler motives exist. That seems like a pretty depressing world view, and I am glad to see indications that it might be false, such as this story.
Patents slowing down progress by discouraging use and hindering competitors is not new, but has been going on since the beginning of the patent system. The beginning chapter of Against Intellectual Monopoly details the case of the steam engine, where progress in efficiency and adoption of the steam engine was effectively halted for the duration of Watt's patents, only to take off right after they expired.
Once Watt’s patents were secured and production started, a substantial portion of his energy was devoted to fending off rival inventors. In 1782, Watt secured an additional patent, made “necessary in consequence of... having been so unfairly anticipated, by [Matthew] Wasborough in the crank motion.” More dramatically, in the 1790s, when the superior Hornblower engine was put into production, Boulton and Watt went after him with the full force of the legal system. During the period of Watt’s patents the U.K. added about 750 horsepower of steam engines per year. In the thirty years following Watt’s patents, additional horsepower was added at a rate of more than 4,000 per year. Moreover, the fuel efficiency of steam engines changed little during the period of Watt’s patent; while between 1810 and 1835 it is estimated to have increased by a factor of five. After the expiration of Watt’s patents, not only was there an explosion in the production and efficiency of engines, but steam power came into its own as the driving force of the industrial revolution. Over a thirty year period steam engines were modified and improved as crucial innovations such as the steam train, the steamboat and the steam jenny came into wide usage. The key innovation was the high-pressure steam engine – development of which had been blocked by Watt’s strategic use of his patent.
The above is just a short section, they go through the case very thoroughly (with references), and it is worth a read. Interestingly, the steam engine is often quoted by patent proponents as an example of patents working like they are supposed to.
I skimmed through the paper itself, and it seems like flies are only mentioned in passing. The paper mainly concerns itself with vertebrates, and their new result is that they have tested the hypothesis that smallness of body and high metabolism correlate with the flicker fusion frequency of the visual system, i.e. how fast a light has to flash before the flashing becomes invisible. They find the hypothesis to hold (like your teacher suspected).
The fact that flies have a very high flicker fusion frequency (270 Hz vs. 60 for humans under ideal lighting), has, however, been known for a long time, and is not a new result from this paper. In fact, houseflies have 2.5 times higher flicker fusion frequency than even the smallest and most active vertebrates tested in this study (actually, looking at their graphs, it seems like the housefly would be a huge outlier if they had included it).
The flicker fusion frequency is related to, but not the same thing, as how often an image needs to change in order to be percieved as motion. This difference is why 50-60 Hz CRT screens are annoyingly flashy to many, while 25 fps movies look fine. In the latter case, each image only changes slightly.
For a fly, watching a 25 fps movie would probably be similar to watching an 8 fps movie for a human.
I agree. The Wayland developers have convincingly argued that X11 is a broken heap of cruft, and non-backwards compatible changes seem to be needed to fix that. But as you say, that could be done by updating the protocol itself. Judging from the X version number, that has happened many times in the past, but somehow we've gotten stuck at version 11 for a long while now. Make X13 or X14 the redesigned version with all the cruft removed, and use the versions between there as a deprecation buffer.
I just read the chapter on symmetry, and that is a bit out of date in that while it correctly explains that parity symmetry is broken, it still incorrectly claims that parity-charge symmetry holds, which we now know is false.
The lectures are very educational and engagingly written, so I recommend that you give it a go anyway. If you take it all on face value, you will end up with only a very few, minor misunderstandings.
That article does not show that one couldn't use a fingerprint recovered from the outside of the iPhone itself to authenticate. It uses the same pattern of ridges that you leave behind anytime you touch something. However, it reads it as a capacitance pattern rather than a visual image. So to fool it, it wouldn't be enough to display it on a normal display, or print it using normal ink. One would need a display/ink/whatever that creates features with capacitance that the fingerprint reader will recognize. I'm sure it would be possible to create ink with the right properties, and with a lot of more work, it should also be possible to create a capacitance 'display', which assigns a programmable capacitance to each pixel.
Of course, all of this would be a lot of work, since these tools aren't available now. But it is a weakness that the device has its own password plastered around on its outside (and on your books, windows, glass, table, etc.). I would use it to complement passwords, not replace them.
That's a good question. I think, in theory, that a distributed solution could handle it. You would need a distributed peer-to-peer protocol which also handles storage and discovery. The actual data transfer could be done though something like bittorrent, but with blocks downloaded chronologically.
Services like gnutella, e-donkey etc. actually do something relatively similar to youtube when you think of it.
You're completely right that a low value on a die roll does not predict that the next roll will be below average. After, all, each roll of the dice is independent. However, we are asking a slightly different question: What will the next value be relative to the previous one? And this comparison with the previous value introduces correlations. For example, if you roll a 1, the probability that the next value will be smaller than that is 0. But if you roll a 6, the probabiliy that the next value is smaller will be 5/6. And if you roll a 1 or a 6, the probability that the next value will be less extreme is 2/3.
Similarly, if you draw a random standard normal distributed number and get -3, then you will be right 99.87% of the time if you bet that the next number will be higher. Even though all the numbers are completely independent. In general, if you see an extreme value, it is a good bet that the next value will be less extreme, because that is where most of the probability volume is. And this observation is what is called regression to the mean.
I think the way the term was used in that article was completely consistent with this. It was a pretty safe bet that this year would see a rise in the ice coverage because last year was a large negative fluctuation compared to the long-term trend, and most fluctuations deviate less. So regression towards the mean can be used to make certain kinds of predictions, though they all boil down to the banal "the next value will probably be inside the most probable range of values".
80% of climate scientists who were asked last year expected more ice this year than 2013. So this is hardly an unforseen event. The blog link mentioned in the summary explains why, but I'll repeat it since you didn't read it.
Arctic ice volume has a falling long-term trends, but on top of that there are short-term year-by-year changes. You effectively have a long-term signal with short-term noise on it. As you can see from this figure, the trend is about -0.065 million square kilometers per year, while the year-to-year variations are 0.5-1 million square kilometers. Hence, on a short timescale you can basically only see the yearly random variations. If you suddenly see a large jump, it is much more likely to be a short-term change than a long term one, and several years of observations are needed to see if the long-term behavior has changed or not.
The point now is that if you happen to get a particularly low value of the random yearly variations one year, you are likely to get a larger value the next year. Much like if you roll a die and get a 1, you are likely to get a larger value the next time you roll, simply because there are more values (2,3,4,5,6) that are larger than 1 than those that aren't (1). In general, extreme values are unlikely, and the chance of getting several of them in a row is much lower than getting one of them followed by less extreme values. This is called regression toward the mean.
So to summarize, this was expected, and predicted, and no models will have to be changed based on this observation.
You bring up one good point, which is the ocean mixing timescale. During the short-term future, only the top layer will have time to mix. A typical depth seems to be about 50 m, which is surprisingly shallow. That brings the effective ocean mass down from 1.3e18 kg to 1.8e16 kg, or a factor of 72. My other numbers would then need to be multiplied by this factor. Hence, the surface layer would get an effective exposure of 0.45 mSv/year.
Taking bioaccumulation into effect, assuming most of the biomass is in the top mixed layer, and still using the factor 640 estimate (the highest Cesium bioaccumulation factors I could find in a quick search were about 10, but I might have missed something higher), that would give humans an effective exposure of 286 mSv/year. That is definitely big, and would probably be enough to significantly reduce average life expectancy. I couldn't find any numbers for how much, though.
What I don't understand is why you didn't try to do the calculation yourself. You seem very interested in this issue, after all. Your answer would have been a good one if you had brougt up the issue of ocean mixing and then proceeded to show its effect rather than pouring out angry accusations.
What will happen if all fuel that is now in reactors will be washed out to the ocean and spread worldwide by currents?
That's a really good question! I'd like to know the answer myself. Let's try a rough back-of-the-envelope calculation.
According to this, the risky fuel pond 4 has 1.4e18 Bq of mostly Cesium 137 which decays by beta and gamma radiation, releasing 1.176MeV per decay, giving 264 kW total. If dissipated uniformly in the ocean, this would result in 264 kW/1.3e18kg = 2.0e-13 W/kg = 2.0e-13 Gray/s. Since it is beta and gamma radiation and uniformly permeating, we can translate this directly into 2.0e-13 Sv/s = 6.4 uSv/year. This can be compared to the natural background radiation, which is about 2.4 mSv/year.
However, Cesium may be subject to bioaccumulation. If we assume perfect bioaccumulation, then all the cesium at the bottom of the food chain will end up at the top (i.e. humans). This is a huge exaggeration (think many orders of magnitude), but let's see what we get. The total ocean biomass is about 2.24e14 kg, while the total human biomass is about 3.5e11 kg. So after all the ocean biomass has passed through humans, and if all the Cesium is retained (whch it won't be), then we would have magnification of 640, bringing us to 4.1 mSv/year, which is almost double the natural background radiation. So that might give measurable effects, but is still not dangerous.
So unless I have made any huge errors, it seems like Fukushima will not be able to threaten humanity. That doesn't mean it wouldn't be a local problem, though.
How much radiation are we talking about here, in the worst case scenario you present? How radioactive would Tokyo become? I don't know how much radioactive material in total is in the spent fuel tanks, but it would be interesting to see some numbers about how many mSv/hour one would get in Tokyo if the spent fuel tanks powdered and smeared evenly across the city. My intuition on this is pretty useless. I can't say whether it would be at the level 0.1% risk of cancer per person per year or everybody dead in a week. You make it sounds more like the latter, but have you actually checked the numbers?
If you have a political target for the amount of research and engineering that should be done, then it is not necessarily inconsistent that there are too few people taking education in these fields while few of them can get any jobs. That situation would be consistent with there being too few STEM jobs *and* too few STEM workers. This is not unrealistic - countries like to compare themselves with others based on how much research they do, so it is only natural to set political goals for research.
When India claims it needs 800 new universities, that could be based not on how many STEM jobs are available in India at the moment, but on how much research India wants to do. I haven't checked if that is the case here, but if it is, I would accept the statement to be accompanied by something like "and we will also need 200000 new STEM jobs to reach our research goals".
As mentioned in http://www.tepco.co.jp/en/announcements/2013/1230191_5502.html (mentioned by another commenter), the high radiation was 5 cm off the bottom, and fell very quickly with height. So this seems to be almost exactly the same situation as that in the xkcd strip.
Slashdot Mirror
I don't see any advertisement on facebook at all, and I'm just using Adblock Edge + the rule facebook.com##div[class="ego_column"]. SocialFixer does more than just block advertisements, but if all you want is to get rid of then adblock already does a very good job at it.
Examples of this working are x264, libavcodec and mplayer, for example. All of these probably break large numbers of patents and are quite high profile and I'm sure they are a thorn in the side for the video and audio format cartels. But they are doing just fine, and have been doing so for a long time. If they had been closed one-man projects by somebody in the USA, they life expectancy would probably be much shorter.
This was one of the most insightful comments I've read on slashdot.
And the success of kickstarter shows that we don't need a huge change in people's mindsets for this to work right now. (Too bad that most kickstarter projects still insist on enforcing copyright even after they have been fully funded from the beginning, though. They could at least add "free software" as a stretch goal.)
I do this occasionally, when my internet connection is too flaky. I don't know anybody else who does. I'm sure less than 1% of youtube's users has ever saved a youtube-video to disk. I doubt google notices any impact from this at all. The reason for any arms race is paranoia, not actual risk.
What I and a much greater (but still small) number of people do, which probably *is* noticable for google, is to block advertisements on youtube. It works perfectly out of the box, and removes not only the general advertisements on the page, but also the video advertisements. I was disgusted recently when I tried to use youtube on a friend's computer, and had to wait through 10 seconds of an obnoxious, unseekable commercial before getting to the actual video.
Needless to say, I don't feel bad at all for using youtube this way. This is like a restaurant which gives you a side dish of heroin even though you didn't order it, and then gets upset when you leave it behind and only eat the main dish. Sure some people like heroin, but sensible people recognize that it is unhealthy and detracts from the experience of the real food.
If enough people start blocking advertisement in youtube for google to shut it down, then that would be a loss - I think youtube is filling an important niche. But it should be possible to fill this niche in other ways too. BitTorrent demonstrated how peer-to-peer distribution can scale cheaply for a relatively low number of relatively large files. It is not suitable as a youtube-replacements, but it would probably be possible to design a different distributed system that is. Something like freenet, but without the huge overhead of anonymous routing. That isn't necessary right now due to youtube filling the niche. But if youtube were to go under, the niche itself wouldn't disappear.
There is some discussion on this here.
In particular, these two images from the same article are interesting: Temperature anomaly for the medieval warm period and temperature anomaly for the period 1999-2008. Both are anomalies relative to the same 1961-1990 average, so they should be directly comparable, though of course the medieval warm period is a reconstruction with significant uncertainties.
So to answer your question. yes, you could say that "Canada was still frozen while Greenland was basking in warmth". Though temperatures slightly elevated in some parts of Canada, most of it was cold. And none of them were anywhere near as hot as they are now.
I, too, work in HPC computing, and while I found "codes" very jarring to begin with, I've learned to live with it. I am not sure the "code" vs. "codes" issue it is more grammatically problematic than "people" vs. "peoples". A people (countable) is made up of people (uncountable). Similarly "a code" (countable, but nonstandard) is made up of code (uncountable). Personally I would use "a program" or "a library" instead of "a code", though.
Another related issue is whether "data" is countable or not. I'm used to it being uncountable, with there being more or less of it, but not "several data". But scientific journals in my field prefer the countable version "a datum", "several data", which is arguably more historically correct. That, too, took some getting used to.
Higher distance gives lower capacitance, not higher. This does not change the argument, though.
From this macworld article on the subject:
A capacitance fingerprint reader leverages a handy property of your skin: The outer layer of your skin (your dermis), where your fingerprint is, is non-conductive, while the subdermal layer behind it is conductive. When you touch the iPhone’s fingerprint sensor, it measures the minuscule differences in conductivity caused by the raised parts of your fingerprint, and it uses those measurements to form an image..
A capacitor works by having an insulator sandwitched between two conductors. The thinner the insulator is, the higher the capacitance. In the case of a capacitive fingerprint reader, the conductors are the reader itself on one side, and the subdermal layers on the other side. In between them, the skin works as an insulator. Hence, by measuring the capacitance, one is effectively measuring the thickness of the skin. I.e. the pattern of ridges and valleys visible on your fingers. This is the layer you claimed wasn't measured in the first place.
What is your source for claiming that the sensor reads a different pattern than the normal fingerprints you leave behind? A capacitive fingerprint reader works by measuring the difference in capacitance between the ridges and valleys of your fingerprint. In the ridges, the distance to the more conductive layers beneath the skin (the sub-dermal layers you've heard about) is greater than in the valleys, which gives these regions higher capacitance. I guess the pattern you get this way could be different from the visible fingerprint if the underside of the skin has a significant, different pattern than the overside, but I have not heard that that is supposed to be the case.
To simplify things a bit, the much touted sub-dermal layers work as a sort of capacitive back-light which highlights the differences in thickness of the fingerprint above it. It is, to the best of my knowledge, simply another way of measuring the same fingerprint we see when we look at our fingers.
The source you site seems to be saying the opposite of what you claim:
When you touch the iPhone’s fingerprint sensor, it measures the minuscule differences in conductivity caused by the raised parts of your fingerprint, and it uses those measurements to form an image..
Those raised parts of the fingerprint are exactly the ones that deposit fat stains on every surface you touch.
Of course, it is possible that the macworld article is misleading, and that the fingerprint reader reads some other pattern after all. If so, it would be nice to see a source that backs that up. This has been brought up in previous slashdot discussions too, but I have never seen any evidence backing it up, even after explicitly asking for it.
While I think this particular project (and the one to mars) are unrealistic due to not being robustly funded and planned, I think the concept that some humans may be willing to greatly shorten their lives in order to go off and explore something new shows that there is still hope for humanity. Isn't it nice to hear that some are curious enough and adventurous enough to do this, instead of simply being focused on entertainment and producing more humans?
I would not be brave enough to do this myself, even if I actually believed these projects were serious, but I wouldn't belittle those who are willing to sacrifice themselves like this. It seems to be implicit in your post that people only have one objective: to be as comfortable as possible, and so the only reason they might have to leave the Earth is if it is too uncomfortable here. I.e. no nobler motives exist. That seems like a pretty depressing world view, and I am glad to see indications that it might be false, such as this story.
Patents slowing down progress by discouraging use and hindering competitors is not new, but has been going on since the beginning of the patent system. The beginning chapter of Against Intellectual Monopoly details the case of the steam engine, where progress in efficiency and adoption of the steam engine was effectively halted for the duration of Watt's patents, only to take off right after they expired.
Once Watt’s patents were secured and production started, a substantial portion of his energy was devoted to fending off rival inventors. In 1782, Watt secured an additional patent, made “necessary in consequence of ... having been so unfairly anticipated, by [Matthew] Wasborough in the crank motion.” More dramatically, in the 1790s, when the superior Hornblower engine was put into production, Boulton and Watt went after him with the full force of the legal system. During the period of Watt’s patents the U.K. added about 750 horsepower of steam engines per year. In the thirty years following Watt’s patents, additional horsepower was added at a rate of more than 4,000 per year. Moreover, the fuel efficiency of steam engines changed little during the period of Watt’s patent; while between 1810 and 1835 it is estimated to have increased by a factor of five. After the expiration of Watt’s patents, not only was there an explosion in the production and efficiency of engines, but steam power came into its own as the driving force of the industrial revolution. Over a thirty year period steam engines were modified and improved as crucial innovations such as the steam train, the steamboat and the steam jenny came into wide usage. The key innovation was the high-pressure steam engine – development of which had been blocked by Watt’s strategic use of his patent.
The above is just a short section, they go through the case very thoroughly (with references), and it is worth a read. Interestingly, the steam engine is often quoted by patent proponents as an example of patents working like they are supposed to.
I skimmed through the paper itself, and it seems like flies are only mentioned in passing. The paper mainly concerns itself with vertebrates, and their new result is that they have tested the hypothesis that smallness of body and high metabolism correlate with the flicker fusion frequency of the visual system, i.e. how fast a light has to flash before the flashing becomes invisible. They find the hypothesis to hold (like your teacher suspected).
The fact that flies have a very high flicker fusion frequency (270 Hz vs. 60 for humans under ideal lighting), has, however, been known for a long time, and is not a new result from this paper. In fact, houseflies have 2.5 times higher flicker fusion frequency than even the smallest and most active vertebrates tested in this study (actually, looking at their graphs, it seems like the housefly would be a huge outlier if they had included it).
The flicker fusion frequency is related to, but not the same thing, as how often an image needs to change in order to be percieved as motion. This difference is why 50-60 Hz CRT screens are annoyingly flashy to many, while 25 fps movies look fine. In the latter case, each image only changes slightly.
For a fly, watching a 25 fps movie would probably be similar to watching an 8 fps movie for a human.
I agree. The Wayland developers have convincingly argued that X11 is a broken heap of cruft, and non-backwards compatible changes seem to be needed to fix that. But as you say, that could be done by updating the protocol itself. Judging from the X version number, that has happened many times in the past, but somehow we've gotten stuck at version 11 for a long while now. Make X13 or X14 the redesigned version with all the cruft removed, and use the versions between there as a deprecation buffer.
I just read the chapter on symmetry, and that is a bit out of date in that while it correctly explains that parity symmetry is broken, it still incorrectly claims that parity-charge symmetry holds, which we now know is false.
The lectures are very educational and engagingly written, so I recommend that you give it a go anyway. If you take it all on face value, you will end up with only a very few, minor misunderstandings.
That article does not show that one couldn't use a fingerprint recovered from the outside of the iPhone itself to authenticate. It uses the same pattern of ridges that you leave behind anytime you touch something. However, it reads it as a capacitance pattern rather than a visual image. So to fool it, it wouldn't be enough to display it on a normal display, or print it using normal ink. One would need a display/ink/whatever that creates features with capacitance that the fingerprint reader will recognize. I'm sure it would be possible to create ink with the right properties, and with a lot of more work, it should also be possible to create a capacitance 'display', which assigns a programmable capacitance to each pixel.
Of course, all of this would be a lot of work, since these tools aren't available now. But it is a weakness that the device has its own password plastered around on its outside (and on your books, windows, glass, table, etc.). I would use it to complement passwords, not replace them.
That's a good question. I think, in theory, that a distributed solution could handle it. You would need a distributed peer-to-peer protocol which also handles storage and discovery. The actual data transfer could be done though something like bittorrent, but with blocks downloaded chronologically.
Services like gnutella, e-donkey etc. actually do something relatively similar to youtube when you think of it.
You're completely right that a low value on a die roll does not predict that the next roll will be below average. After, all, each roll of the dice is independent. However, we are asking a slightly different question: What will the next value be relative to the previous one? And this comparison with the previous value introduces correlations. For example, if you roll a 1, the probability that the next value will be smaller than that is 0. But if you roll a 6, the probabiliy that the next value is smaller will be 5/6. And if you roll a 1 or a 6, the probability that the next value will be less extreme is 2/3.
Similarly, if you draw a random standard normal distributed number and get -3, then you will be right 99.87% of the time if you bet that the next number will be higher. Even though all the numbers are completely independent. In general, if you see an extreme value, it is a good bet that the next value will be less extreme, because that is where most of the probability volume is. And this observation is what is called regression to the mean.
I think the way the term was used in that article was completely consistent with this. It was a pretty safe bet that this year would see a rise in the ice coverage because last year was a large negative fluctuation compared to the long-term trend, and most fluctuations deviate less. So regression towards the mean can be used to make certain kinds of predictions, though they all boil down to the banal "the next value will probably be inside the most probable range of values".
80% of climate scientists who were asked last year expected more ice this year than 2013. So this is hardly an unforseen event. The blog link mentioned in the summary explains why, but I'll repeat it since you didn't read it.
Arctic ice volume has a falling long-term trends, but on top of that there are short-term year-by-year changes. You effectively have a long-term signal with short-term noise on it. As you can see from this figure, the trend is about -0.065 million square kilometers per year, while the year-to-year variations are 0.5-1 million square kilometers. Hence, on a short timescale you can basically only see the yearly random variations. If you suddenly see a large jump, it is much more likely to be a short-term change than a long term one, and several years of observations are needed to see if the long-term behavior has changed or not.
The point now is that if you happen to get a particularly low value of the random yearly variations one year, you are likely to get a larger value the next year. Much like if you roll a die and get a 1, you are likely to get a larger value the next time you roll, simply because there are more values (2,3,4,5,6) that are larger than 1 than those that aren't (1). In general, extreme values are unlikely, and the chance of getting several of them in a row is much lower than getting one of them followed by less extreme values. This is called regression toward the mean.
So to summarize, this was expected, and predicted, and no models will have to be changed based on this observation.
You bring up one good point, which is the ocean mixing timescale. During the short-term future, only the top layer will have time to mix. A typical depth seems to be about 50 m, which is surprisingly shallow. That brings the effective ocean mass down from 1.3e18 kg to 1.8e16 kg, or a factor of 72. My other numbers would then need to be multiplied by this factor. Hence, the surface layer would get an effective exposure of 0.45 mSv/year.
Taking bioaccumulation into effect, assuming most of the biomass is in the top mixed layer, and still using the factor 640 estimate (the highest Cesium bioaccumulation factors I could find in a quick search were about 10, but I might have missed something higher), that would give humans an effective exposure of 286 mSv/year. That is definitely big, and would probably be enough to significantly reduce average life expectancy. I couldn't find any numbers for how much, though.
What I don't understand is why you didn't try to do the calculation yourself. You seem very interested in this issue, after all. Your answer would have been a good one if you had brougt up the issue of ocean mixing and then proceeded to show its effect rather than pouring out angry accusations.
What will happen if all fuel that is now in reactors will be washed out to the ocean and spread worldwide by currents?
That's a really good question! I'd like to know the answer myself. Let's try a rough back-of-the-envelope calculation.
According to this, the risky fuel pond 4 has 1.4e18 Bq of mostly Cesium 137 which decays by beta and gamma radiation, releasing 1.176MeV per decay, giving 264 kW total. If dissipated uniformly in the ocean, this would result in 264 kW/1.3e18kg = 2.0e-13 W/kg = 2.0e-13 Gray/s. Since it is beta and gamma radiation and uniformly permeating, we can translate this directly into 2.0e-13 Sv/s = 6.4 uSv/year. This can be compared to the natural background radiation, which is about 2.4 mSv/year.
However, Cesium may be subject to bioaccumulation. If we assume perfect bioaccumulation, then all the cesium at the bottom of the food chain will end up at the top (i.e. humans). This is a huge exaggeration (think many orders of magnitude), but let's see what we get. The total ocean biomass is about 2.24e14 kg, while the total human biomass is about 3.5e11 kg. So after all the ocean biomass has passed through humans, and if all the Cesium is retained (whch it won't be), then we would have magnification of 640, bringing us to 4.1 mSv/year, which is almost double the natural background radiation. So that might give measurable effects, but is still not dangerous.
So unless I have made any huge errors, it seems like Fukushima will not be able to threaten humanity. That doesn't mean it wouldn't be a local problem, though.
How much radiation are we talking about here, in the worst case scenario you present? How radioactive would Tokyo become? I don't know how much radioactive material in total is in the spent fuel tanks, but it would be interesting to see some numbers about how many mSv/hour one would get in Tokyo if the spent fuel tanks powdered and smeared evenly across the city. My intuition on this is pretty useless. I can't say whether it would be at the level 0.1% risk of cancer per person per year or everybody dead in a week. You make it sounds more like the latter, but have you actually checked the numbers?
Source: http://www.tepco.co.jp/en/announcements/2013/1230191_5502.html
If you have a political target for the amount of research and engineering that should be done, then it is not necessarily inconsistent that there are too few people taking education in these fields while few of them can get any jobs. That situation would be consistent with there being too few STEM jobs *and* too few STEM workers. This is not unrealistic - countries like to compare themselves with others based on how much research they do, so it is only natural to set political goals for research.
When India claims it needs 800 new universities, that could be based not on how many STEM jobs are available in India at the moment, but on how much research India wants to do. I haven't checked if that is the case here, but if it is, I would accept the statement to be accompanied by something like "and we will also need 200000 new STEM jobs to reach our research goals".
http://what-if.xkcd.com/29/
As mentioned in http://www.tepco.co.jp/en/announcements/2013/1230191_5502.html (mentioned by another commenter), the high radiation was 5 cm off the bottom, and fell very quickly with height. So this seems to be almost exactly the same situation as that in the xkcd strip.