Because you're not in possession of the copyrighted image, but your mirror (the physical kind, not the server kind) is reflecting its image. Therefore, in the court's opinion it's a copyright violation. Brilliant reasoning.
The FAA limits flight time to 1000 hours per year. Which translated into the typical 5 days/week, 50 weeks/yr most people are familiar with, is an average of 4 hours per workday. As you point out, other factors lead to increased time spent working on the ground, which is probably part of the problem (not just for pilots but for air travel overall).
I think the low pay at regional airlines you point out is the bigger culprit. Airlines have tried to cut costs by shifting operations to regional airlines, where they have more leeway to cut costs and wages. This has resulted in them viewing regional airlines as a place to reduce the cost of pilot labor, rather than as "training grounds" for pilots who'll eventually work at the major airlines. I suspect airlines could get away with doing this because for most of the 20th century, there was a large supply of experienced ex-military pilots.
However, we haven't had any extended wars for over 4 decades. Throughout the 20th century, the military was churning out a constant stream of experienced pilots who logged hundreds if not thousands of hours flying during WWII, the Korean War, and the Vietnam War. Since then, the wars the U.S. has been involved in have only lasted days or weeks - not enough time for pilots to log a lot of hours. Consequently, the crutch of experienced ex-military pilots the airlines had been relying on to justify driving down wages for regional airline pilots, has evaporated. If you consider a military pilot who logged hours over Vietnam, figure he left service at age 25 when the war ended (1975). Today he would be 68, which is right around retirement age. So yeah, right now is about the time you would expect the problem to come to a head.
The obvious solution is to raise pilot salaries at regional airlines. That'll get more civilian pilots to seriously consider it as a career, solving the supply shortage. Unfortunately there's a lag factor at play here. The FAA requires 1500 hours flight time before you can fly for the big airlines. Which combined with the 1000 hours/year limit means a minimum 1.5 year lag between the effect of any changes you make at the regional airline level trickling up to the major airlines.
No country has free medical care. The only difference is in how it's managed and paid for. The U.S. uses a combination of government programs, private insurance, and personal funding. A lot of countries use government programs and personal funding. So in countries with "free" health care, it's not free, you are paying for it via your taxes.
As for the hypothesis that it's the lack of government-funded health care which drives up prices in the U.S., in 2009 prior to Obamacare passing, the U.S. government was alreadyspending more per capita on health care than any country except Norway, Luxembourg, and the Netherlands. If government-funded health care were truly the solution, the U.S. government was already spending enough to copy Canada's health care system wholesale before Obamacare was even proposed. The problems with the U.S. health care system are deeper and more complex than "it's because you don't have a government-funded system."
The reason they're not voting for statehood is because their current status as a U.S. territory means residents don't pay Federal income tax. As a territory, they get the benefits of being a state (Federal assistance and participation in Federal programs), without having to pay for it.
The Trump administration foot-dragging on the hurricane disaster response in Puerto Rico is a tactic to pressure them to hurry up and decide - do they want to become a state, or become independent as a new country? Right now, the limbo state they're in - which is supposed to be temporary to give them time to decide - is more attractive than either option, resulting in them putting off the decision for 118 years.
So Trump's handling of Puerto Rico is in fact impartial. He's putting the country's best interest ahead of his party's best interest (since Puerto Rico would almost certainly end up a blue state).
It's no different from the defendant claiming he's lost the key, while the judge thinks he's just carefully hidden it somewhere and isn't telling. Both (the password, or location of the physical key) are "mind reading" aspects. The case law does in fact apply here.
This is what I've been trying to caution people against when they cite 4th Amendment protection for passwords. The 4th Amendment isn't a bulletproof shield. Once a warrant has been issued (as was in this case), pretty much all of your 4th Amendment protections evaporate. Failing to obey the warrant puts you at risk of being jailed indefinitely for contempt of court. No trial, no jury, the judge just sends you to jail because you didn't obey a court order.
Real wages (wages with inflation and purchasing power factored in) depends on productivity. You add up the productivity of all your citizens, divide it by the number of citizens you have, and that's the average (mean) wage per citizen in real (purchasing power) dollars. While money is not conserved ("economics is not a zero-sum game"), productivity is. Everything that's consumed must be produced. So productivity is zero-sum (negative-sum if you account for losses like crop spoilage, broken windows, etc).
Raising the minimum wage doesn't directly affect productivity. But iself, a minimum wage redistributes income (takes some of the income that would've gone to company owners, and gives it to employees as increased wages). If the sale of an item produces $x of productivity, increasing the minimum wage means the worker gets a higher percentage of x, and the company owner gets a smaller percentage.
The only way an increased minimum wage can help raise productivity (increase real wages and the standard of living) is if workers are already being underpaid. A market economy wants to allocate wages (price for labor) in direct proportion to how much productivity each laborer is actually generating. If a business owner is keeping too much of the company's income for himself instead of paying his employees fair wages, that represents an economic inefficiency. Henry Ford stumbled upon this when he began paying his workers far above the prevailing wage at the time. Because workers were being underpaid, the economy was artificially stunted. When he paid his workers higher wages which turned out to be fair, that created an economic feedback loop. His employees could suddenly afford to buy the cars they were producing, which increased demand for cars, which increased the amount of work for these employees, which increased their productivity, which increased their wages further.
Crucially, Ford's income also increased as a result of this, making him one of the richest men on earth. Even though he got to keep a smaller percentage of his company's income, because the wage increase was fair, it increased his company's income by a larger percentage than his percentage share declined, and he ended up making more money (a lot more). That's the important part most minimum wage proponents miss. That if a situation exists where a minimum wage would help, not only would it increase employee's incomes, it would also increase employer's incomes. In other words, if a minimum wage increases employees' income at the expense of business owners' income, then it's bad for the economy (it's forcing a wage that's too high for the productivity the employees generate, and thus reducing overall productivity for the country). If a minimum wage increases both employees' and owners' income, then it's good for the economy (it's forcing a fairer distribution of business income to employee vs employer wages).
EV sales aren't being driven by the market though. They're being driven by California's ZEV mandate, which requires a certain percentage of each automaker's sales be EVs (or they buy credits from an automaker who exceeds their quota). If they fail, they're banned from selling cars in California. And since about a dozen states automatically adopt California's guidelines, they'd end up banned from selling cars from about a third of the U.S. by population. Nobody wants to be cut off from a third of the U.S. market, so they're all working to push out EVs, and offer incentives and sales to make sure they sell enough of them to meet their quota.
Once you understand that, you realize that giving each automaker a reserved slice of the subsidy is the only way to make it fair.
The ironic thing about this entire incident to me is that what Price experienced, fair or not, is basically the employment world that men have lived in for the last couple decades. Stepping on eggshells for fear of doing something that might offend some woman, who could file a sexual harassment claim against them and get them fired.
I'm a firm believer that offense has to be intended by the person making the comment for it to be discrimination (be it misogyny or misandry). If you set the benchmark as whether the person hearing the comment was offended, you create a logically inconsistent process. As your program to stamp out discrimination succeeds, more people grow up never learning certain slurs and stereotypes. But because your system is based on whether someone was offended, someone who does remember those slurs gets offended. When they report their offense, an innocent person is punished for violating some rule they never knew existed. That's already happened with racism. And you end up punishing people whose only crime was that your program to stamp out discrimination was successful . The only rule which is logically consistent is to punish behavior intended to be offensive. A much more difficult standard to prove, but the only one which makes logical sense.
Unfortunately, we live in a world (or at least a media) that's decided that offense should measured based on how listeners interpret what you say. And by that standard, Price was hung with the same rope used to hang many a man before her (someone interpreted her statements as sexist and were offended).
1. The same rules apply to everyone. Media companies don't get a free passdon't get a free pass just because they own a lot of copyrighted content.
2. The company(ies) requesting the block have to financially compensate the blocked website if it's later discovered that their claim of copyright violation was in error. Unlike is currently done under the DMCA where media companies regularly claim copyright violation on YouTube videos and get them defended. And when the person who posts the video is finally able to prove that there was no copyright violation, the media companies only have to say "oopsies, sorry."
If you follow these common-sense guidelines, you'll quickly find that the problem with blocking websites for repeated copyright violations is that the websites which feature large amounts of media (e.g. news sites, art/photo sharing sites, etc) are the ones which get accused of copyright violation the most. And you'll conclude that an outright ban based on a handful of accusations ends up hurting some of the most useful sites disproportionately.
Explanation given in TFA (and omitted in summary and other articles crowing over this):
To be fair, Apple's relatively new APFS file system is designed to speed up file file copies using a technology Apple calls Instant Cloning. But a win is a win.
the technology used in the new cloning feature makes it easier to store multiple versions of a file in a minimum of space
In other words, the files weren't copied. A hard link (similar to a shortcut for you Windows users) was created. The whole story is an error by non-techie journalists who noticed something wildly odd in their test results, and rather than spend 30 seconds researching it online like I did, decided "it must be because it's Apple!" and published it. The reality distortion field is alive and well.
Apple has been using Sandisk NAND lately as a bid to try to reduce dependence on Samsung. Both Sandisk and Toshiba SSDs (also used frequently by Apple) regularly benchmark slower than Samsung SSDs.
2016 was different from 2000. In 2000, Gore won a plurality of the popular vote (but not a majority), and liberal parties (Gore + Nader) won a majority of the popular vote. The Electoral college did not reflect a majority of the popular vote in 2000.
In 2016, Clinton won a plurality (but not a majority) of the popular vote. However, Conservative parties combined won more popular votes than liberal parties combined (49.88% vs 49.13%, with the rest being cast for candidates without a party affiliation). So in 2016, the Electoral College awarded the election to the candidate whose ideology came closest to winning a majority of votes, rather than the individual who came closest to winning a majority. I didn't vote for Trump, but he was probably the correct winner in 2016.
People like to criticize the Electoral College. But IMHO the plurality-wins system they propose be used instead is nearly as bad (consider the California primaries where some candidates won one of two slots in the general election with barely 20% of the vote). We really need to switch to instant run-off voting, which is designed so that a candidate always gets a majority of votes before being declared the winner.
PC sales began leveling off in the late 1990s, more steeply after 2000. Long before smartphones and tablets.
What happened was Intel and AMD ran headfirst into physics. Prior to 2000, CPU clock speeds had been doubling roughly every 18 months. But the power a processor needs increases non-linearly with frequency. Past about 3 GHz (roughly 2002), CPUs began to require exorbitant amounts of additional power for little gains in clock speed.
Consequently, the rate of clock speed increases nearly stalled after 2002 (at a bit above 3 GHz). Before 2000, each new gen of Intel CPU roughly doubled performance. Today, each new gen only nets about a 5%-15% performance improvement, and most of that has been due to improvements in parallel processing (more cores, speculative execution, hyperthreading, all the goodies which made the news last year as avenues for new exploits).
Up til about 20002, software makers had been counting on increased CPU performance to support the new features they were adding. They relied on people upgrading their PCs to be able to run the latest version of their software. Now that an upgraded PC was barely faster than the PC it replaced, software makers were forced to do something they'd given little thought to in the past - optimize.
Why not just add a flexible bellows around each key, sealing it against its cutout and prevent crud from getting underneath the key?
The only reason we live with crud getting underneath laptop keys is because most of them have a big cut-out for the entire keyboard. So the nearest neighboring structure to a key are its neighboring keys, so there's no way to seal them since the seal would bridge multiple keys together, inhibiting them from moving independently. But if you've got an individual cutout for each key, then it's possible to seal each individual key. As a bonus, the cutout doesn't need as tight a tolerance (in fact you want a little more space between it and the key so crud can be picked out).
By that reasoning, shouldn't a government not have investments in any company? Or invest in broad index funds (i.e. investment in all companies) if they insist on investing?
The role of government is to make laws which benefit society. Having investments (or lack of investments) in particular companies or industries compromises its ability to be impartial when making laws which affect those industries. (Ignoring for the moment the many other ways corruption can creep in.)
That was the same problem JEDEC ran into with Rambus. According to JEDEC rules, members weren't allowed to patent the technologies being discussed. The members were laying the framework for DDR2, and Rambus simply lifted it and patented it. After a lot of court cases, the courts finally decided that Rambus had violated the terms they'd agreed to when they joined JEDEC. However, since the membership terms didn't lay out any penalties for what would happen if someone violated the rules, the only recourse JEDEC had was to kick Rambus out. Outside of JEDEC, Rambus' patents were still valid.
If there's no guaranteed penalty for breaking a promise, the promise is worthless.
Yeah, I feel for you guys up there. I worked in Canada for a couple years. When I went shopping for a Canadian cell phone, it turned out to actually be cheaper for me to add Canada roaming to my U.S. plan for $10/mo (calls while in Canada would be billed as if they were made in the U.S.), than it was for me to replace my U.S. plan with a Canadian plan. And I know from traveling to other parts of the world that U.S. cell phone plans are already quite expensive.
The device doesn't actually measure the color of light the tissue reflects (can't due to no light reaching some of these tissues). It determines what type of tissue it is, then colors it in the images based on what we know the color of that type of tissue to be. Like adding color to old black and white movies based on our best estimate of skin tones, grass, roads, etc.
If you start a program to fertilize coral reefs, and it turns out the fertilizer has other unintended consequences (like causing algal blooms), you become financially and possibly criminally liable for those consequences.
If birds fertilize coral reefs with their droppings, and their droppings have other unintended consequences, nobody sues the birds in court.
Gives multiple meanings to the phrase, "shit happens."
The better (more cost-effective) solution to the problem is a simple 90-degree right-angle connector for the power cord plugging into the laptop. The straight plugs turned into a lever whenever you yanked the cord sideways, which would twist the internal socket and eventually break it. If you used the laptop at the edge of a table with the power cord draping down, then the weight of the cord on this lever was constantly twisting the socket down. By using a right-angle connector, you limit the maximum length of this lever and thus the maximum torque it can place on the socket. The forces instead get transferred into the plastic/metal housing at the edge of the socket, which is easy to build strong enough to withstand the loads.
The MagSafe connectors do help prevent you from sending the laptop flying if you trip over the power cable. But with most laptops coming with SSDs nowadays, that's not a problem. There are virtually no moving parts inside a modern laptop, so they can withstand very high accelerations without sustaining damage. Most solid state electronics can be shot out of a canon without sustaining damage.
Many older cities, especially in Europe, followed a circular diagram, with ray streets outwards and round ones connecting them. At first that would be the result of being squeezed into city walls, then followed expansion outwards, along these streets and filling in the space in between.
It's nowhere near as structured as circles, rays, and hoops. What happened was there was an important location here, and an important location there, and some people decided "hey we should build a road connecting these important locations!" because people were traveling between them all the time. Repeat with enough important locations and you end up with the rays spreading out from the important locations to each other. The rest of the city then fills in around these important locations, with side streets shooting off these rays, approximating hoops but nowhere near as exact nor complete. Eventually one fill-in meets another, with corresponding discordance between street orientation.
The place names in Boston still reflect this. The "important places" are called squares (Central Square, Haymarket Square, Kenmore Square, Harvard Square, etc). And the crazy street orientation is due to the major streets connecting these squares with each other. Same for Asia - many Asian cities don't even give streets names. They name the important locations, and the street leading to an important location is simply named "road to [important location]."
While I agree the program is abused, the program was started due to other countries running similar programs. A study found that the U.S. was suffering a net drain of talented graduates leaving for jobs in these other countries. These work visa programs are basically ways for countries to poach talent from each other, and the U.S. had been on the losing end. So it started its own work visa program.
Both iOS and Android already give the device owner control over what functions an app is able to access. For example, Android notified me that an update to one of the games on my tablet was asking for access to the microphone and camera. I of course denied those permissions (the game seems to run just fine without them). Since my tablet is rooted, I also get control over which apps are allowed to use the network. So even with the few programs which need such access (like a photo-to-PDF converter), I'm confident it isn't transmitting info about me back to the app maker.
There are two reasons for the problem.
Certain apps need such permissions. The voice input app mentioned in the summary requires access to the microphone to function. The maker of the app can then abuse that permission to use the microphone to record conversations and transmit them back to the mothership. This is even more insidious with voice recognition apps, which have to record conversations and transmit them back for the recognition stage anyway. At that point the difference between legitimate and illegitimate use becomes whether the company keeps the recordings on file, or deletes them after the recognition is completed (which is why I've long advocated that voice recognition be moved to the device itself now that processors are getting to the point where that's feasible). It's impossible for OS-level restrictions to prevent this type of abuse.
China has encouraged forking Android and developing its own version for use in the Chinese market. Ostensibly this is to reduce the amount of control foreign companies (namely Google) have over products used within China. Most people however suspect that it's done so the Chinese government can insert its own monitoring software directly within the OS itself. The kind of stuff the NSA only dreams it could do. The maker of an open-source OS has no control over what happens to forks.
Further proof that machine learning and AI has real world use.
That's the catch - it's learned. For it to learn correctly, the initial data set be based on materials actually tested on animals (or people).
I'm not saying this isn't useful. But understand that it's basically just interpolating within regimes already covered by empirical test data. It's not extrapolation outside the bounds of that empirical data set. It's inherently limited to biological reactions to known substances and chemically similar substances. It won't be able to accurately predict how biology will react to a new material (though it might be able to make a statistically better-than-chance guess).
Looking at the samples, it's pretty clear that most of the noise was high-frequency (i.e. pixel-level), high-contrast. That's relatively easy to filter out even without AI. Back in the days of analog TVs, if you had a poor signal the image ended up with a lot of snow. Since the snow was high-frequency, high-contrast (black and white dots), a trick to filtering it out was to cover the screen with pieces of tissue paper. The static (CRTs used electron guns) held the tissues up against the screen, and they were translucent enough that they acted as a lowpass filter blocking out high-frequency info. They averaged out the high-frequency noise to where the image was actually improved. You lost a little bit of high-resolution detail, but you couldn't see it anyway in the first place because of all the snow. I guarantee you there was no AI in the tissue paper.
They've got one sample in the paper with colored text overlaid on top of the picture. Unfortunately it's too small for me to see clearly, but I suspect the AI was able to distinguish the text simply because the colors were so different from the rest of the picture. It shows up as deviations in the RGB channel histograms as discussed earlier in the paper.
The tough noise to remove is stuff that's ambiguous. A friend once brought me a photo printed on textured paper. She'd lost the negative, and wanted a clean copy of the photo. But the paper's texture (a repeated pattern of subtle whorls which could've easily been mistaken for the texture of her clothes if it didn't also show up on her face and background) showed up on all the scans. I threw a bunch of filters at it without success. I ended up editing it by hand to manually blur out the whorls (she was willing to pay to have the photo cleaned). Now *that* is something you could probably train an AI to remove based on sampling the whorls on featureless background.
Slashdot Mirror
Because you're not in possession of the copyrighted image, but your mirror (the physical kind, not the server kind) is reflecting its image. Therefore, in the court's opinion it's a copyright violation. Brilliant reasoning.
The FAA limits flight time to 1000 hours per year. Which translated into the typical 5 days/week, 50 weeks/yr most people are familiar with, is an average of 4 hours per workday. As you point out, other factors lead to increased time spent working on the ground, which is probably part of the problem (not just for pilots but for air travel overall).
I think the low pay at regional airlines you point out is the bigger culprit. Airlines have tried to cut costs by shifting operations to regional airlines, where they have more leeway to cut costs and wages. This has resulted in them viewing regional airlines as a place to reduce the cost of pilot labor, rather than as "training grounds" for pilots who'll eventually work at the major airlines. I suspect airlines could get away with doing this because for most of the 20th century, there was a large supply of experienced ex-military pilots.
However, we haven't had any extended wars for over 4 decades. Throughout the 20th century, the military was churning out a constant stream of experienced pilots who logged hundreds if not thousands of hours flying during WWII, the Korean War, and the Vietnam War. Since then, the wars the U.S. has been involved in have only lasted days or weeks - not enough time for pilots to log a lot of hours. Consequently, the crutch of experienced ex-military pilots the airlines had been relying on to justify driving down wages for regional airline pilots, has evaporated. If you consider a military pilot who logged hours over Vietnam, figure he left service at age 25 when the war ended (1975). Today he would be 68, which is right around retirement age. So yeah, right now is about the time you would expect the problem to come to a head.
The obvious solution is to raise pilot salaries at regional airlines. That'll get more civilian pilots to seriously consider it as a career, solving the supply shortage. Unfortunately there's a lag factor at play here. The FAA requires 1500 hours flight time before you can fly for the big airlines. Which combined with the 1000 hours/year limit means a minimum 1.5 year lag between the effect of any changes you make at the regional airline level trickling up to the major airlines.
No country has free medical care. The only difference is in how it's managed and paid for. The U.S. uses a combination of government programs, private insurance, and personal funding. A lot of countries use government programs and personal funding. So in countries with "free" health care, it's not free, you are paying for it via your taxes.
As for the hypothesis that it's the lack of government-funded health care which drives up prices in the U.S., in 2009 prior to Obamacare passing, the U.S. government was already spending more per capita on health care than any country except Norway, Luxembourg, and the Netherlands. If government-funded health care were truly the solution, the U.S. government was already spending enough to copy Canada's health care system wholesale before Obamacare was even proposed. The problems with the U.S. health care system are deeper and more complex than "it's because you don't have a government-funded system."
The reason they're not voting for statehood is because their current status as a U.S. territory means residents don't pay Federal income tax. As a territory, they get the benefits of being a state (Federal assistance and participation in Federal programs), without having to pay for it.
The Trump administration foot-dragging on the hurricane disaster response in Puerto Rico is a tactic to pressure them to hurry up and decide - do they want to become a state, or become independent as a new country? Right now, the limbo state they're in - which is supposed to be temporary to give them time to decide - is more attractive than either option, resulting in them putting off the decision for 118 years.
So Trump's handling of Puerto Rico is in fact impartial. He's putting the country's best interest ahead of his party's best interest (since Puerto Rico would almost certainly end up a blue state).
It's no different from the defendant claiming he's lost the key, while the judge thinks he's just carefully hidden it somewhere and isn't telling. Both (the password, or location of the physical key) are "mind reading" aspects. The case law does in fact apply here.
This is what I've been trying to caution people against when they cite 4th Amendment protection for passwords. The 4th Amendment isn't a bulletproof shield. Once a warrant has been issued (as was in this case), pretty much all of your 4th Amendment protections evaporate. Failing to obey the warrant puts you at risk of being jailed indefinitely for contempt of court. No trial, no jury, the judge just sends you to jail because you didn't obey a court order.
Real wages (wages with inflation and purchasing power factored in) depends on productivity. You add up the productivity of all your citizens, divide it by the number of citizens you have, and that's the average (mean) wage per citizen in real (purchasing power) dollars. While money is not conserved ("economics is not a zero-sum game"), productivity is. Everything that's consumed must be produced. So productivity is zero-sum (negative-sum if you account for losses like crop spoilage, broken windows, etc).
Raising the minimum wage doesn't directly affect productivity. But iself, a minimum wage redistributes income (takes some of the income that would've gone to company owners, and gives it to employees as increased wages). If the sale of an item produces $x of productivity, increasing the minimum wage means the worker gets a higher percentage of x, and the company owner gets a smaller percentage.
The only way an increased minimum wage can help raise productivity (increase real wages and the standard of living) is if workers are already being underpaid. A market economy wants to allocate wages (price for labor) in direct proportion to how much productivity each laborer is actually generating. If a business owner is keeping too much of the company's income for himself instead of paying his employees fair wages, that represents an economic inefficiency. Henry Ford stumbled upon this when he began paying his workers far above the prevailing wage at the time. Because workers were being underpaid, the economy was artificially stunted. When he paid his workers higher wages which turned out to be fair, that created an economic feedback loop. His employees could suddenly afford to buy the cars they were producing, which increased demand for cars, which increased the amount of work for these employees, which increased their productivity, which increased their wages further.
Crucially, Ford's income also increased as a result of this, making him one of the richest men on earth. Even though he got to keep a smaller percentage of his company's income, because the wage increase was fair, it increased his company's income by a larger percentage than his percentage share declined, and he ended up making more money (a lot more). That's the important part most minimum wage proponents miss. That if a situation exists where a minimum wage would help, not only would it increase employee's incomes, it would also increase employer's incomes. In other words, if a minimum wage increases employees' income at the expense of business owners' income, then it's bad for the economy (it's forcing a wage that's too high for the productivity the employees generate, and thus reducing overall productivity for the country). If a minimum wage increases both employees' and owners' income, then it's good for the economy (it's forcing a fairer distribution of business income to employee vs employer wages).
EV sales aren't being driven by the market though. They're being driven by California's ZEV mandate, which requires a certain percentage of each automaker's sales be EVs (or they buy credits from an automaker who exceeds their quota). If they fail, they're banned from selling cars in California. And since about a dozen states automatically adopt California's guidelines, they'd end up banned from selling cars from about a third of the U.S. by population. Nobody wants to be cut off from a third of the U.S. market, so they're all working to push out EVs, and offer incentives and sales to make sure they sell enough of them to meet their quota.
Once you understand that, you realize that giving each automaker a reserved slice of the subsidy is the only way to make it fair.
The ironic thing about this entire incident to me is that what Price experienced, fair or not, is basically the employment world that men have lived in for the last couple decades. Stepping on eggshells for fear of doing something that might offend some woman, who could file a sexual harassment claim against them and get them fired.
I'm a firm believer that offense has to be intended by the person making the comment for it to be discrimination (be it misogyny or misandry). If you set the benchmark as whether the person hearing the comment was offended, you create a logically inconsistent process. As your program to stamp out discrimination succeeds, more people grow up never learning certain slurs and stereotypes. But because your system is based on whether someone was offended, someone who does remember those slurs gets offended. When they report their offense, an innocent person is punished for violating some rule they never knew existed. That's already happened with racism. And you end up punishing people whose only crime was that your program to stamp out discrimination was successful . The only rule which is logically consistent is to punish behavior intended to be offensive. A much more difficult standard to prove, but the only one which makes logical sense.
Unfortunately, we live in a world (or at least a media) that's decided that offense should measured based on how listeners interpret what you say. And by that standard, Price was hung with the same rope used to hang many a man before her (someone interpreted her statements as sexist and were offended).
1. The same rules apply to everyone. Media companies don't get a free passdon't get a free pass just because they own a lot of copyrighted content.
2. The company(ies) requesting the block have to financially compensate the blocked website if it's later discovered that their claim of copyright violation was in error. Unlike is currently done under the DMCA where media companies regularly claim copyright violation on YouTube videos and get them defended. And when the person who posts the video is finally able to prove that there was no copyright violation, the media companies only have to say "oopsies, sorry."
If you follow these common-sense guidelines, you'll quickly find that the problem with blocking websites for repeated copyright violations is that the websites which feature large amounts of media (e.g. news sites, art/photo sharing sites, etc) are the ones which get accused of copyright violation the most. And you'll conclude that an outright ban based on a handful of accusations ends up hurting some of the most useful sites disproportionately.
Some research turns up that:
In other words, the files weren't copied. A hard link (similar to a shortcut for you Windows users) was created. The whole story is an error by non-techie journalists who noticed something wildly odd in their test results, and rather than spend 30 seconds researching it online like I did, decided "it must be because it's Apple!" and published it. The reality distortion field is alive and well.
Apple has been using Sandisk NAND lately as a bid to try to reduce dependence on Samsung. Both Sandisk and Toshiba SSDs (also used frequently by Apple) regularly benchmark slower than Samsung SSDs.
2016 was different from 2000. In 2000, Gore won a plurality of the popular vote (but not a majority), and liberal parties (Gore + Nader) won a majority of the popular vote. The Electoral college did not reflect a majority of the popular vote in 2000.
In 2016, Clinton won a plurality (but not a majority) of the popular vote. However, Conservative parties combined won more popular votes than liberal parties combined (49.88% vs 49.13%, with the rest being cast for candidates without a party affiliation). So in 2016, the Electoral College awarded the election to the candidate whose ideology came closest to winning a majority of votes, rather than the individual who came closest to winning a majority. I didn't vote for Trump, but he was probably the correct winner in 2016.
People like to criticize the Electoral College. But IMHO the plurality-wins system they propose be used instead is nearly as bad (consider the California primaries where some candidates won one of two slots in the general election with barely 20% of the vote). We really need to switch to instant run-off voting, which is designed so that a candidate always gets a majority of votes before being declared the winner.
PC sales began leveling off in the late 1990s, more steeply after 2000. Long before smartphones and tablets.
What happened was Intel and AMD ran headfirst into physics. Prior to 2000, CPU clock speeds had been doubling roughly every 18 months. But the power a processor needs increases non-linearly with frequency. Past about 3 GHz (roughly 2002), CPUs began to require exorbitant amounts of additional power for little gains in clock speed.
Consequently, the rate of clock speed increases nearly stalled after 2002 (at a bit above 3 GHz). Before 2000, each new gen of Intel CPU roughly doubled performance. Today, each new gen only nets about a 5%-15% performance improvement, and most of that has been due to improvements in parallel processing (more cores, speculative execution, hyperthreading, all the goodies which made the news last year as avenues for new exploits).
Up til about 20002, software makers had been counting on increased CPU performance to support the new features they were adding. They relied on people upgrading their PCs to be able to run the latest version of their software. Now that an upgraded PC was barely faster than the PC it replaced, software makers were forced to do something they'd given little thought to in the past - optimize.
Why not just add a flexible bellows around each key, sealing it against its cutout and prevent crud from getting underneath the key?
The only reason we live with crud getting underneath laptop keys is because most of them have a big cut-out for the entire keyboard. So the nearest neighboring structure to a key are its neighboring keys, so there's no way to seal them since the seal would bridge multiple keys together, inhibiting them from moving independently. But if you've got an individual cutout for each key, then it's possible to seal each individual key. As a bonus, the cutout doesn't need as tight a tolerance (in fact you want a little more space between it and the key so crud can be picked out).
By that reasoning, shouldn't a government not have investments in any company? Or invest in broad index funds (i.e. investment in all companies) if they insist on investing?
The role of government is to make laws which benefit society. Having investments (or lack of investments) in particular companies or industries compromises its ability to be impartial when making laws which affect those industries. (Ignoring for the moment the many other ways corruption can creep in.)
That was the same problem JEDEC ran into with Rambus. According to JEDEC rules, members weren't allowed to patent the technologies being discussed. The members were laying the framework for DDR2, and Rambus simply lifted it and patented it. After a lot of court cases, the courts finally decided that Rambus had violated the terms they'd agreed to when they joined JEDEC. However, since the membership terms didn't lay out any penalties for what would happen if someone violated the rules, the only recourse JEDEC had was to kick Rambus out. Outside of JEDEC, Rambus' patents were still valid.
If there's no guaranteed penalty for breaking a promise, the promise is worthless.
The Onion beat you to that punchline almost a decade ago.
Yeah, I feel for you guys up there. I worked in Canada for a couple years. When I went shopping for a Canadian cell phone, it turned out to actually be cheaper for me to add Canada roaming to my U.S. plan for $10/mo (calls while in Canada would be billed as if they were made in the U.S.), than it was for me to replace my U.S. plan with a Canadian plan. And I know from traveling to other parts of the world that U.S. cell phone plans are already quite expensive.
The device doesn't actually measure the color of light the tissue reflects (can't due to no light reaching some of these tissues). It determines what type of tissue it is, then colors it in the images based on what we know the color of that type of tissue to be. Like adding color to old black and white movies based on our best estimate of skin tones, grass, roads, etc.
If you start a program to fertilize coral reefs, and it turns out the fertilizer has other unintended consequences (like causing algal blooms), you become financially and possibly criminally liable for those consequences.
If birds fertilize coral reefs with their droppings, and their droppings have other unintended consequences, nobody sues the birds in court.
Gives multiple meanings to the phrase, "shit happens."
The better (more cost-effective) solution to the problem is a simple 90-degree right-angle connector for the power cord plugging into the laptop. The straight plugs turned into a lever whenever you yanked the cord sideways, which would twist the internal socket and eventually break it. If you used the laptop at the edge of a table with the power cord draping down, then the weight of the cord on this lever was constantly twisting the socket down. By using a right-angle connector, you limit the maximum length of this lever and thus the maximum torque it can place on the socket. The forces instead get transferred into the plastic/metal housing at the edge of the socket, which is easy to build strong enough to withstand the loads.
The MagSafe connectors do help prevent you from sending the laptop flying if you trip over the power cable. But with most laptops coming with SSDs nowadays, that's not a problem. There are virtually no moving parts inside a modern laptop, so they can withstand very high accelerations without sustaining damage. Most solid state electronics can be shot out of a canon without sustaining damage.
It's nowhere near as structured as circles, rays, and hoops. What happened was there was an important location here, and an important location there, and some people decided "hey we should build a road connecting these important locations!" because people were traveling between them all the time. Repeat with enough important locations and you end up with the rays spreading out from the important locations to each other. The rest of the city then fills in around these important locations, with side streets shooting off these rays, approximating hoops but nowhere near as exact nor complete. Eventually one fill-in meets another, with corresponding discordance between street orientation.
The place names in Boston still reflect this. The "important places" are called squares (Central Square, Haymarket Square, Kenmore Square, Harvard Square, etc). And the crazy street orientation is due to the major streets connecting these squares with each other. Same for Asia - many Asian cities don't even give streets names. They name the important locations, and the street leading to an important location is simply named "road to [important location]."
While I agree the program is abused, the program was started due to other countries running similar programs. A study found that the U.S. was suffering a net drain of talented graduates leaving for jobs in these other countries. These work visa programs are basically ways for countries to poach talent from each other, and the U.S. had been on the losing end. So it started its own work visa program.
There are two reasons for the problem.
That's the catch - it's learned. For it to learn correctly, the initial data set be based on materials actually tested on animals (or people).
I'm not saying this isn't useful. But understand that it's basically just interpolating within regimes already covered by empirical test data. It's not extrapolation outside the bounds of that empirical data set. It's inherently limited to biological reactions to known substances and chemically similar substances. It won't be able to accurately predict how biology will react to a new material (though it might be able to make a statistically better-than-chance guess).
Looking at the samples, it's pretty clear that most of the noise was high-frequency (i.e. pixel-level), high-contrast. That's relatively easy to filter out even without AI. Back in the days of analog TVs, if you had a poor signal the image ended up with a lot of snow. Since the snow was high-frequency, high-contrast (black and white dots), a trick to filtering it out was to cover the screen with pieces of tissue paper. The static (CRTs used electron guns) held the tissues up against the screen, and they were translucent enough that they acted as a lowpass filter blocking out high-frequency info. They averaged out the high-frequency noise to where the image was actually improved. You lost a little bit of high-resolution detail, but you couldn't see it anyway in the first place because of all the snow. I guarantee you there was no AI in the tissue paper.
They've got one sample in the paper with colored text overlaid on top of the picture. Unfortunately it's too small for me to see clearly, but I suspect the AI was able to distinguish the text simply because the colors were so different from the rest of the picture. It shows up as deviations in the RGB channel histograms as discussed earlier in the paper.
The tough noise to remove is stuff that's ambiguous. A friend once brought me a photo printed on textured paper. She'd lost the negative, and wanted a clean copy of the photo. But the paper's texture (a repeated pattern of subtle whorls which could've easily been mistaken for the texture of her clothes if it didn't also show up on her face and background) showed up on all the scans. I threw a bunch of filters at it without success. I ended up editing it by hand to manually blur out the whorls (she was willing to pay to have the photo cleaned). Now *that* is something you could probably train an AI to remove based on sampling the whorls on featureless background.