This is the problem with affirmative action. It creates an atmosphere where certain types of discrimination (e.g. against males, whites and asians) are acceptable, when the actual goal is to eliminate discrimination. Don't get me wrong, affirmative action has a valid mathematical basis - in engineering (underdamped system) we use it to force a system to change states more quickly. But it comes at the cost of a twitchy response and overshoot, resulting in taking longer to settle down to a steady state.
Basically affirmative action lowers (or eliminates) the inhibition against discrimination against white males to force the system away from being white male-dominated more quickly (underdamped). The things in Exhibit B (a lot of anti-white, anti-male vitriol) are a natural (albeit undesirable) result, which is going to push the system to eventually overshoot. At which point there's going to be a pushback, which will cause the system to swing back the other way, overshoot again, repeat. This is what you get when you try to fight discrimination with discrimination (aka affirmative action). One type of discrimination which used to be tolerated (discrimination against non-whites and women) gets replaced by tolerance for a different type of discrimination (against whites and men), until it overshoots and we'll eventually go back to tolerating discrimination against non-whites and women again, repeat over and over until the oscillation dies out and we finally arrive at our goal of equality.
As an alternative, a comprehensive non-discrimination policy (critical or overdamping - i.e. all types of discrimination are bad, so both discrimination and affirmative action are banned) takes longer to reach the desired level (equality), but it doesn't suffer from oscillations and reaches a steady state more quickly. The time constant here is people's lifetimes. By having affirmative action policies, we've taught a generation from the time they were kids that it's OK to discriminate against whites and men. They will hold those opinions and attitudes until they die, and are replaced by a younger generation which was taught differently. So when the system eventually overshoots, it'll take another generation to correct for that, and so on with each overshoot. And after a few dozen generations we'll stabilize at the desired steady state. Whereas the comprehensive non-discrimination policy would stabilize at the desired steady state somewhere between 1-2 generations.
I know which one I prefer, but this is something we as a society have to decide which path we want to take.
So well researched and reasoned that the authors of the two papers he relies on the most have publicly stated that he didn't understand them, and that his conclusions are wrong.
The basic mistake he makes repeatedly is to assume that the variations the papers discuss have vastly more effect and influence than they actually do.
He didn't assume the magnitude and influence of the effect. The difference has been measured for decades in observed data over thousands of studies. If those opposed to Damore could only find two paper authors on the topic who disagreed with him, then that sounds like a pretty strong validation of his claims, not a rebuttal. Heck, I could throw a rock blindfolded and hit two climate change denying studies.
Why is stating that women have a higher rate of neurosis a fireable offense. But stating that men have a higher rate of schizophrenia is not?
The problem Damore's case shows us is that too many people are judging the merits of these statements based on which group they portray in a negative light. Not upon the objective validity of the statement. If you wanted to counter Damore's statements on neurosis and gender, the logical (quickest and easiest) way to do it would be as I've done above - showing that there are other psychological gender differences which work against men biologically dominating an occupation. Then you can claim that perhaps these effects cancel out so a 50/50 gender distribution really should be expected.
But that's not what Damore's opponents do. They instead try to debunk measurable, objective data that's well-established science. They cannot stand to hear anything negative said about a group they care for (i.e. non-white, non-male, non-conservative, non-religious). So their gut instinct is that the statement that neurosis is more common among women "must be" wrong, and they conclude disproving it will be the quickest route to disproving him.
The rate of population growth is proportional to the size of the female population, not the male. One male can fertilize multiple females. Consequently, males are not as important to sustaining or growing the population as females are. While an entirely female turtle population would be bad, the rate of population growth is maximized at some point above 50% female. The more often turtles are to encounter each other during mating season, the fewer males are required to maximize population growth. So if you want the turtle population to grow more quickly, you want more eggs to become female (up to a point)
(This is why we traditionally sent men off to fight wars, and not women. Any society which also sent its women to fight and die would stunt its own population growth, putting itself at a numeric disadvantage in future wars. It's also why tradition says to save the "women and children first" during a disaster. So long as enough males survive to fertilize the remaining female population, the rest of the men are expendable - their deaths will have little impact on the rate of population growth.)
You're trying to be funny, but I think this highlights part of the problem. It's not necessarily Google's image recognition software, it's also poor photographs. A self-driving car (if it needed to distinguish between gorillas and black people) wouldn't have as much problem because it can adjust the camera's exposure to where it can see enough detail to distinguish the two. But a lot of photos of black people are taken with the wrong exposure, and the black skin tones end up crushed down into just a few discrete color values near black, or clipped to 0. The AI then ends up trying to distinguish one black humanoid-shaped blob from another.
When I shot weddings on film, I had to use special low contrast film (had a larger dynamic range). That was the only way to retain detail in both the bride's white dress and the groom's black tuxedo. And even then, if the wedding was held in sunlight a lot of the detail might still be unrecoverable. Modern cameras are getting around the problem with automatic HDR photo mode (takes two photos at different exposures to preserve detail in both the highlights and shadows, then combines them nonlinearly). But there's still a huge library of badly-exposed photos out there (from the pre-HDR days and being added to by current photographers taking simple snapshots without really caring about exposure), just waiting to trip up any image recognition AI.
White skin tends to be slightly brighter than the average background, while black skin tends to be much darker. So to properly expose a portrait of a black person, you have to either make sure their skin dominates the camera's auto-exposure algorithm, and not the background or their clothing. Or put additional light specifically on their face (e.g. fill flash) so it's not so dark relative to the background. A professional photographer knows this. The average person taking a snapshot, and the auto-exposure algorithm in their camera, does not.
Colleges began using test scores because different high schools (and in fact different teachers) awarded grades differently. An 4.0 in one high school might only be worth a 3.2 in another high school. A 4.0 by a student who took the "easy" teachers' courses might only be worth a 3.5 by a student who took the harder courses at the same high school. The test scores were used to try to normalize the grades.
The problem started when some college admissions staff got lazy, and started using test scores as a cutoff for admissions. That way they could circular file a bunch of the applicants without even having to read their application (but we still get to cash your application fee check kthxbye).
many colleges just want to have many students - as the fees will pay the bills
I don't think that assumption is right. Nearly all colleges get way more applicants than they accept. If what you say were true, they'd simply accept every applicant. (The ones which do accept everyone are usually scam colleges - soaking up subsidized Federal loan dollars in exchange for a middling or useless education, then leaving the student stuck with the bills.)
I suspect colleges aim for a certain uniformity in their students' capabilities so the professors don't have to dumb down lessons or worry that the material is too easy for some of the students. By stratifying education on the basis of student intelligence and capability, you can tailor the lessons and speed at which material is presented to better suit the strata of student ability the college has selected, thus reducing time wasted on trying to help poor students catch up, and time wasted by exceptional students on too-easy lessons. That improved efficiency also leads to cost reduction (professors don't have to spend as much time creating lessons, helping students, don't need to hire as many TAs, etc).
In this regard, TFA is walking a very fine line between claiming that test scores shouldn't exclude an otherwise exceptional student, and claiming that test scores shouldn't matter at all.
We went through this with Rambus. They joined JEDEC (a consortium of memory manufacturers setting future memory standards) and agreed to its terms of membership - mainly, members are not allowed to patent the memory standards being discussed. DDR was being discussed within JEDEC. Rambus went ahead and patented it, and sued the other JEDEC members for violating "their" patents.
After years of legal battles, the courts found that yes Rambus was guilty of violating JEDEC's membership agreement, and they were subject to whatever punishment they agreed to when they joined JEDEC. But that had nothing to do with the law, so the patents were valid (Rambus being the first to file). Meanwhile, since the JEDEC membership agreement didn't outline any punishment for violating the agreement, the only thing JEDEC could do was kick Rambus out.
Same thing here. An EULA or ToS is just a contract. If you violate it, you become subject to whatever punishment you agreed to when agreed to the contract. That does not automatically make it a violation of law however. It's only a violation of the law if the act was otherwise illegal. In Rambus' case, patenting stuff freely presented to you is not illegal. In Remini's case, downloading stuff you've been authorized to download is not illegal.
That's the way it's done in parliamentary elections. In the U.S. however, the Founding Fathers decided that legislators needed to be more in-touch with the voters they represented. This resulted in the idea of geographic representation. Each Congressman represents a geographical district - a specific chunk of the country whose voters elected that specific candidate.
Even the proportional representation system used in parliaments has mathematical problems. Imagine a parliament with 4 parties. Party A has 5 seats, Party B has 4 seats, Party C has 3 seats, party D has 1 seat. That's 13 seats in total, with 7 needed to decide a majority. Even though D ostensibly has 8% of the seats, it has zero power. It cannot change the outcome of any vote. A+B can generate a majority (9). A+C can generate a majority (8). B+C can generate a majority (7). Party D cannot change any of these outcomes, and thus it holds no power. It might as well not exist.
Mathematically, a representation system where the proportion of votes received is mapped onto the equivalent amount of power in government (ability to change decisions) would be ideal. But that would violate the "one person, one vote" ideal that people really seem beholden to, so will probably never happen.
So when it comes to Gerrymandering, most of the more "liberal" or even "purple" States have long had laws against it. Many (including CA) have independent councils that must have representatives from both parties in roughly equal proportion drawing the maps.
Erm, the independent council in California for apolitically drawing congressional maps exists due to the work of Republicans. California was long gerrymandered by the Democrats. The stats I read in the 1980s were something like 55% of votes being cast in elections were for Democratic candidates, but 60% of its Assembly members and 62% of its Congressional representatives and State Senators were Democrats.
The Republicans finally managed to put a stop to it in 1990 by campaigning hard to get Pete Wilson elected as governor. After the 1990 census, the Democrats tried to gerrymander California's districts again. Wilson simply vetoed it. That led to a lot of drama which eventually ended up in the California Supreme Court, who ended up drawing the districts themselves just so there would be districts in time for the 1992 elections. Eventually, the two parties came up with an independent council with representatives from both parties drawing the maps, as a long-term solution.
Democrats gerrymander too, don't get me wrong. But they haven't abused it to the extend like the Republicans in NC did
Probably not to the extent it was done in NC (which was particularly egregious - I oppose election manipulation regardless of party). But Democrats gerrymandered their way into the longest continuous control of the House by one party in history - 40 years. That presented a quandary for the Republicans: How do you stop a system of electoral abuse which gives more votes to the abuser thus allowing them to propagate the abuse? The solution they came up with was to prioritize getting Republicans elected as governors in time for the 1990 census. Then as in California, those governors simply vetoed the districts drawn by the Democrats. (Pete Wilson was a shoe-in to be re-elected as Senator, but they decided winning the governor's office, even for just 4 years, was more important than a Senate seat in DC. They made a strategic gambit, giving up his Senate seat for a better chance at winning the governorship. His seat was eventually won by the Democrat who ran against him for governor - Feinstein, who still holds it today.)
That's why we have the situation today of gerrymandering favoring Republicans. Red governors of blue states are more common than blue governors of red states. The states where Democrats traditionally held power in the legislative branches, the Republicans managed to thwart their gerrymandering with vetoes after the 1990 or 2000 census. And the two parties were forced to compromise on an apolitical way to draw districts. So it's mostly the blue states which have these independent councils drawing districts (thanks to the work of Republicans). Meanwhile, the red states weren't affected as much by any Democrat governorship counter-strategy, so their districts are drawn the same old way as before - by the party controlling the state legislature. Which results in net gerrymandering favoring Republicans.
And as long as we're talking about electoral abuse, let's talk about illegal aliens and apportionment. You see, the Constitution simply states that House "Representatives shall be apportioned among the several States according to their respective numbers, counting the whole number of persons in each State, excluding Indians not taxed." Not voters, not citizens, but "persons." Non-taxed Indians are excluded, but not non-citizens. California's 2.3 million illegal aliens give it 3 extra House seats. And because they're almost unive
Or most of the OECD for that matter. The "Aristocracy" rich mostly exists in developing countries where they control the vast majority of the wealth. In those countries they're big fish in little ponds, and maintain their status not only by being rich, but by preventing others from becoming rich. The economy of these countries mostly consists of (by volume) the rich selling and buying to/from each other. The GDP per capita in these countries typically stagnates at around $15,000/yr or below.
The U.S., EU, etc. grew past this stage around the 1900s. Henry Ford accidentally stumbled upon this when he discovered that paying his workers above the prevailing wage actually resulted in more business for himself (because his workers could afford to buy the cars he was producing). That's what happens when you (1) put a worker in a productive job, and (2) pay them a fair wage for the productivity they're generating. Basically, when pay your workers less than a fair wage, you make money for yourself, but you stunt the economy. When you pay your workers a fair wage, you spend more money, but the economy blossoms. Usually more than enough to offset the extra money you spent paying your workers.
A market economy *wants* everyone to be as productive as they can, because the feedback effect of that maximizes average income. GDP per capita in these countries is typically $30,000/yr or higher because the vast majority of the population is contributing a meaningful amount of productivity to the economy. Consequently, the vast majority of the rich in these countries are rich from selling things to the middle class (who by population and aggregate income are much bigger than the richest 1%*). If the average income of the middle class decreases in these countries, it ends up hurting the rich too.
* IRS tax stats show that the top 1% only makes about 20% of the income in the U.S. So if they began buying and selling only amongst themselves and replacing everyone else with robots, that would result in about an 80% pay cut for themselves. The bulk of the country's income (73%) is in the $30k-$500k per year wage range, and it's in the best interest of the 1%ers in the U.S. to insure those people continue to have jobs.
That's basically what Solyndra tried (and failed). It's got two basic problems:
While the efficiency per square meter of sunlight goes up, the efficiency per square meter of solar panel goes down. So it ends up being more expensive than flat solar panels per unit of energy collected (lower cost efficiency). That extra up-front cost may be worth it on a constrained space such as solar panels aboard a sailboat. But if you've got lots of open land that you can't afford to completely cover with panels in the first place, flat panels are better.
The sunlight which isn't absorbed by the photovoltaics in the panel isn't just reflected to be absorbed by more PV. Some of it is absorbed by the non-PV portions of the molecular structure like the protective glass layer on top of the PV layer (heating the array up). So the above case (higher efficiency per square meter of sunlight) is only the ideal case. In most practical implementations (like in Solyndra's case), it actually ends up being lower efficiency per square meter of sunlight due to the sunlight on average hitting the panel surfaces at a more oblique angle than a flat panel (the light has to travel a greater distance through the glass, so more is absorbed by it). This loss is compounded for each time the light reflects.
Black feathers don't have this problem because you don't care what happens to the sunlight you're absorbing, only that it's all absorbed. You're not trying to direct the light at specific molecules within your structure like you are with PV panels. It would however work for thermal solar installations. All you care about in those is converting the sunlight into heat, and thus maximizing net absorption.
Executive orders in the face of a Congress & Senate that outright refuse to even talk about the issues you want heard is one thing. Writing more than all other presidents in the last 50 years because you really wish you were a dictator and not bothering to even ask the legislature who is your own party and holds majorities in both Congress and the Senate is quite another.
Amusingly, Wikipedia has a list of executive orders per year. Trump's 12-month executive order total equals Obama's last 15 months in office. Not 50 years. And Carter (37 years ago) was the last President to issue executive orders at a higher rate than Trump. If you're going to badmouth him, at least get your facts right.
Also, taking something you hated when the other guy did it and going completely wild with it when you're in power, is not winning. It's just being a ginormous hypocrite.
I suspect Trump's rate is high just because this was his first year, and many of his executive orders were rolling back or modifying Obama's executive orders. Nothing hypocritical about that. As much as Democrats would've loved it if Trump had kept all of Obama's executive orders in force, we all knew that simply wasn't realistic. The next 3 years will tell if he's a hypocrite about executive orders.
The time constant for earthquake building codes is several decades long. Most of the California earthquake building codes stem from the 1933 Long Beach earthquake (mag 6.4). Most schools back then were brick buildings. Brick walls have no lateral strength - any sideways motion makes them simply fall over. Most of the schools in the Los Angeles areas were severely damaged or destroyed by the quake. But fortunately it hit around 6pm well after school hours. The realization that thousands if not tens of thousands of schoolchildren could've been killed if it had struck earlier in the day is what got the state to start imposing strict earthquake building codes (and banning unreinforced brick buildings).
But older buildings were grandfathered in. Brick structures weren't required to be reinforced until after 1971 San Fernando earthquake. That's when they began requiring steel bars be drilled into brick structures with square retaining plates at the ends, to keep the bricks from falling over under lateral load. You can frequently see these square plates in brick buildings on TV shows and movies, since many of the scenes are shot around Los Angeles. Most of the newer "brick" buildings you see in California are actually wood or concrete structures with a fake brick facade.
But we're learning new things from each quake. The 1994 Northridge quake in particular was eye-opening because one nearby seismograph measured vertical accelerations in excess of 1g (others measured close to 1g indicating it wasn't a fluke or malfunction). Up until then, it was thought that the maximum vertical acceleration possible from an earthquake was about 0.1g. Northridge shattered that assumption, and revised building codes take the possibility of strong vertical accelerations into account.
Anyhow, because of grandfathering, the time constant for most buildings to comply with improved regulations is several decades (average time until an old building is torn down and rebuilt). The difference in dollar damage between Loma Prieta and Northridge is almost entirely attributable to location, not because lessons learned from Loma Prieta got applied within 5 years. The Loma Prieta quake was centered in a forest about 20 miles from downtown San Jose, nearly 60 miles from San Francisco and Oakland. The Northridge quake was centered directly underneath the San Fernando Valley suburbs, and about 20 miles from downtown Los Angeles. Casualties would've been much higher had it struck during the middle of the day with the freeways jammed with traffic (a section of the I-10 overpass in LA collapsed). Fortunately it struck at 4:30 in the morning when most people were asleep in their relatively safe 1- and 2-story homes (the frequency of most earthquake shaking matches the resonance frequency of 3- and 4-story buildings, which are the majority of the ones which collapsed in both quakes).
By the time lessons from previous disasters get applied on a wide scale, increases in population and corresponding infrastructure pretty much offset any reduction in damage due to improved safety standards. A better statistic for comparing over time would be average cost of natural disasters per capita for the year (basically canceling out population growth). You could even make an argument for using the ratio of natural disaster cost to GDP (basically canceling out infrastructure growth). (And of course, normalizing for inflation is a given.)
This is basically what Steam In-Home Streaming does. The PC playing the game converts it into a h.264 video stream in real-time (using dedicated encoding hardware present on all modern GPUs). The device displaying the game just thinks it's playing back a streamed h.264 movie (using its hardware h.264 decoder which is also present on all modern GPUs, even on phones). The only bits that are missing are getting input from the display device and feeding it back to the game PC. That's probably the only reason the client hasn't come to tablets (and phones) yet - they generally don't come with keyboard+mouse or gamepad support.
The STS was a bad left turn for launchers and set the expectation that launch costs would be in the range of $10,000/lb or more. I think that was the real crime - the shuttle's costs got out of control very quickly and nothing was done to reign them in. If the decision was made to drop the STS and keep with Apollo technology (just like the Russians that continued working with their 1960s/1970s technology), which was proven, reliable and cheap compared to the resulting STS and expendable boosters costs along with the same NASA budgets for space exploration, then I suspect a station of the ISS' capabilities could have been put up by the late 1970s as well as maybe an outpost on the moon by 2001 - and we would have avoided the long drought in government sponsored manned space exploration.
That's the thing most people (except the Russians) don't seem to get about rocket launches. It's not a technology-bound problem, it's a physics-bound problem. You throw mass out the rear, and that pushes your rocket forward. The calculations for pretty much every type of rocket fuel were already done in the 1940s-1970s, and the fuels we've been using for the last 50 years are pretty much optimal.
Any new launch vehicle you design is just a pretty wrapper to hold the same fuel and throw it backwards the same way as old designs. So there's little to no performance to be gained by a new design. You need to economically justify the new design in other ways (e.g. reusable stages). Otherwise the most cost-effective strategy is to continue to use the old designs. Which is exactly what the Russians have done. And heaven forbid you spend $35 billion making a new design - that means amortized over an estimated 134 future launches (number of Space Shuttle missions), you've just increased the cost per launch by a quarter billion dollars.
Incidentally, STS was approved with exactly this in mind. The expectation was that by re-using the shuttles, money could be saved thus making launches cheaper in the long term. Unfortunately, that accounting calculation was made assuming about 50 shuttle launches per year. In reality it was less than 5 per year, meaning its facilities, equipment, and personnel costs per launch were 10x higher than projected, turning it from a cheaper-than-Apollo launch vehicle into the most expensive launch vehicle ever created.
The web is dying. It's being replaced by a bunch of proprietary apps you need to install to gain full access to any particular site on a mobile device. Facebook app, Google Maps app, YouTube app, Amazon app, etc. The reason the web worked on the PC was because you only had to install one app - a browser. Then you just typed in a different address for each site you wanted to visit into the browser, and like magic it was almost like you'd installed an app specific to that site. No more. I had to go through and purge a bunch of apps from my phone this weekend because its performance was bogging down (I had over 350 apps installed). On PC about 40%-50% of those apps were just bookmarks in my browser so took up no additional resources unless I was visiting the site at that moment. But on my phone they were constantly using up storage space, and in most cases constantly using RAM and network bandwidth (Amazon is particularly bad about this).
TFA is correct in that regard, and the danger of the shift back to walled gardens (AOL, GEnie, Prodigy back in the days before the web). Basically, rather than adapt to people shifting to mobile devices by creating mobile-friendly websites, companies have taken advantage of the shift to convince people to install a proprietary app on their mobile device instead. That app can then constantly run, monitor, and spy on the user even when it's not being used, and report it all back to the mothership.
Then it goes into a non sequitur about mesh networks saving us from all this, That's kinda like claiming SUVs will save us from proprietary dashboard navigation systems which cost hundreds of dollars to update. Those things have very little to do with each other. There is nothing inherently broken with TCP/IP or HTTP(S), except maybe HTML could be updated with some features to make for a more seamless browsing experience on mobile touch devices. If people are spending 90% of their online time on Google, Facebook, Amazon, moving away from the Internet won't prevent that. If mesh networks become popular, those three companies will just eventually set up mesh network sites, or modify their apps to also work over mesh networks.
A human looks at that picture, sees the banana and "thing" are sitting on a flat surface, and decides they must be about the same distance so their size in their picture is their actual scale. The banana is a lot bigger, so the human decides it is more important than the "thing".
An AI looks at that picture, sees the banana and "thing", but crucially doesn't estimate distance. Since the "thing" has a lot more detail the AI decides it's must be further away, and its greater detail means its the more important part of the picture, and the banana is just fluff in the background. And it gets lost trying to analyze the "thing".
The bit quoted in the summary (quoted from TechCrunch so don't blame slashdot editors for once) is written terribly. But:
The Google Home Mini went on sale on October 19.
Since October 19 (6.75 million seconds), Google has sold more than one Google Home device (Mini or not) per second.
The original Google Home went on sale in November 2016.
Sum total of Google Home devices sold in 2017 (Google Home since January, Google Home Mini since October) is tens of millions (which likely just means over 10 million).
The remote for the Apple TV absolutely sucks. Designers can't seem to get it through their heads that the #1 priority for TV remote is for it to be usable without looking away from the TV. So touchscreens and touchpads are out (except maybe for keyboard entry). You want tactile buttons so people can find the proper button to press without looking away from the screen. (The Logitech Harmonies make this mistake too.)
Despite the widespread availability of guns, Americans killed each other by knife at a higher rate than other countries. So it's not the guns. There's just something about Americans which make them more likely to kill each other, period. In that light, it's not at all surprising that U.S. police response is more aggressive than in other countries.
Intel should have informed vendors a long time ago, like Google did, without of course making the issue a public story until a fix is installed. But Intel admitting the flaw would have triggered many compensation requests. This is one reason why the class action makes sense.
Um, that makes the class action not make sense. According to your reasoning, the threat of a class action caused the very behavior (Intel not informing vendors) the class action is purportedly trying to discourage.
Anyhow, more than likely nobody was harmed by the flaw, and nobody will be harmed by the flaw (unless they refuse to apply the patches, in which case it's on their own heads). Unfortunately for Intel, the harm will mainly come from reduced performance due to the fix. Still, the early reports I'm seeing say that the fix has little performance impact on everyday computing tasks. The large performance degradation seems to be limited mostly to cloud virtual data centers, so I'm not actually sure a class action (for the general population) is warranted.
Good press or bad press it doesn't much matter, he just wants to be in the news. Raising his profile both advances his business / agenda and simply feels good for him. There were 16 Republican candidates who were generally more classically qualified than him, yet he got all the attention, and that's a big part of what won him the presidency.
This right here is IMHO the problem with society today. Not Trump. It's the asinine media which turns inanity into a superstar simply because they value flamboyance over substance. And the public which just eats this stuff up.
Trump being elected President is a symptom, not the problem. The proper reaction to someone being an idiot is to ignore them. Not to post a video of them on Facebook or YouTube just so you can comment "OMG can you believe this guy?" and be the first one to collect a million likes.
Speculative execution requires the CPU to do operations which it might be called upon to do in the future, then throw away the results if it turns out the code doesn't call for that operation. There's got to be an energy cost associated with that since you're making it do a bunch of operations it doesn't need to do. ARM SoCs are mostly used on mobile devices whose only power source is a battery. So it makes sense that they wouldn't incorporate energy-wasting "features" like speculation.
When Tim Cook became CEO, there was one model iPhone, and one model iPad. Today there are 3 iPhone models and 4 iPad models. Meanwhile, the number of Apple employees (non-sales) has not tripled. So there are fewer designers, artists, engineers assigned to each product. More delays are inevitable.
Personally I think it's a step in the right direction. Apple's product line under Jobs was woefully thin. Even at the height of its success the company was literally a single bad iPhone model away from bankruptcy.. Diversifying their product line was exactly what they needed for stability and expanding their customer base (for example, the many Android phone users held off switching until Apple released an iPhone with a bigger screen).
You can't infringe a copyright that does not exist. White noise is not eligible for copyright protection due to there not being an actual author
You're 90% of the way there. The problem here isn't that someone filed a copyright claim against white noise - something that inherently cannot be copyrighted.
The problem is that there is no disincentive, no punishment, for filing bogus copyright claims. The law has put the burden of proof entirely upon the purported infringer to prove he is innocent, none upon the accuser to prove an actual crime was committed.
Slashdot Mirror
This is the problem with affirmative action. It creates an atmosphere where certain types of discrimination (e.g. against males, whites and asians) are acceptable, when the actual goal is to eliminate discrimination. Don't get me wrong, affirmative action has a valid mathematical basis - in engineering (underdamped system) we use it to force a system to change states more quickly. But it comes at the cost of a twitchy response and overshoot, resulting in taking longer to settle down to a steady state.
Basically affirmative action lowers (or eliminates) the inhibition against discrimination against white males to force the system away from being white male-dominated more quickly (underdamped). The things in Exhibit B (a lot of anti-white, anti-male vitriol) are a natural (albeit undesirable) result, which is going to push the system to eventually overshoot. At which point there's going to be a pushback, which will cause the system to swing back the other way, overshoot again, repeat. This is what you get when you try to fight discrimination with discrimination (aka affirmative action). One type of discrimination which used to be tolerated (discrimination against non-whites and women) gets replaced by tolerance for a different type of discrimination (against whites and men), until it overshoots and we'll eventually go back to tolerating discrimination against non-whites and women again, repeat over and over until the oscillation dies out and we finally arrive at our goal of equality.
As an alternative, a comprehensive non-discrimination policy (critical or overdamping - i.e. all types of discrimination are bad, so both discrimination and affirmative action are banned) takes longer to reach the desired level (equality), but it doesn't suffer from oscillations and reaches a steady state more quickly. The time constant here is people's lifetimes. By having affirmative action policies, we've taught a generation from the time they were kids that it's OK to discriminate against whites and men. They will hold those opinions and attitudes until they die, and are replaced by a younger generation which was taught differently. So when the system eventually overshoots, it'll take another generation to correct for that, and so on with each overshoot. And after a few dozen generations we'll stabilize at the desired steady state. Whereas the comprehensive non-discrimination policy would stabilize at the desired steady state somewhere between 1-2 generations.
I know which one I prefer, but this is something we as a society have to decide which path we want to take.
He didn't assume the magnitude and influence of the effect. The difference has been measured for decades in observed data over thousands of studies. If those opposed to Damore could only find two paper authors on the topic who disagreed with him, then that sounds like a pretty strong validation of his claims, not a rebuttal. Heck, I could throw a rock blindfolded and hit two climate change denying studies.
Why is stating that women have a higher rate of neurosis a fireable offense. But stating that men have a higher rate of schizophrenia is not?
The problem Damore's case shows us is that too many people are judging the merits of these statements based on which group they portray in a negative light. Not upon the objective validity of the statement. If you wanted to counter Damore's statements on neurosis and gender, the logical (quickest and easiest) way to do it would be as I've done above - showing that there are other psychological gender differences which work against men biologically dominating an occupation. Then you can claim that perhaps these effects cancel out so a 50/50 gender distribution really should be expected.
But that's not what Damore's opponents do. They instead try to debunk measurable, objective data that's well-established science. They cannot stand to hear anything negative said about a group they care for (i.e. non-white, non-male, non-conservative, non-religious). So their gut instinct is that the statement that neurosis is more common among women "must be" wrong, and they conclude disproving it will be the quickest route to disproving him.
The rate of population growth is proportional to the size of the female population, not the male. One male can fertilize multiple females. Consequently, males are not as important to sustaining or growing the population as females are. While an entirely female turtle population would be bad, the rate of population growth is maximized at some point above 50% female. The more often turtles are to encounter each other during mating season, the fewer males are required to maximize population growth. So if you want the turtle population to grow more quickly, you want more eggs to become female (up to a point)
(This is why we traditionally sent men off to fight wars, and not women. Any society which also sent its women to fight and die would stunt its own population growth, putting itself at a numeric disadvantage in future wars. It's also why tradition says to save the "women and children first" during a disaster. So long as enough males survive to fertilize the remaining female population, the rest of the men are expendable - their deaths will have little impact on the rate of population growth.)
You're trying to be funny, but I think this highlights part of the problem. It's not necessarily Google's image recognition software, it's also poor photographs. A self-driving car (if it needed to distinguish between gorillas and black people) wouldn't have as much problem because it can adjust the camera's exposure to where it can see enough detail to distinguish the two. But a lot of photos of black people are taken with the wrong exposure, and the black skin tones end up crushed down into just a few discrete color values near black, or clipped to 0. The AI then ends up trying to distinguish one black humanoid-shaped blob from another.
When I shot weddings on film, I had to use special low contrast film (had a larger dynamic range). That was the only way to retain detail in both the bride's white dress and the groom's black tuxedo. And even then, if the wedding was held in sunlight a lot of the detail might still be unrecoverable. Modern cameras are getting around the problem with automatic HDR photo mode (takes two photos at different exposures to preserve detail in both the highlights and shadows, then combines them nonlinearly). But there's still a huge library of badly-exposed photos out there (from the pre-HDR days and being added to by current photographers taking simple snapshots without really caring about exposure), just waiting to trip up any image recognition AI.
White skin tends to be slightly brighter than the average background, while black skin tends to be much darker. So to properly expose a portrait of a black person, you have to either make sure their skin dominates the camera's auto-exposure algorithm, and not the background or their clothing. Or put additional light specifically on their face (e.g. fill flash) so it's not so dark relative to the background. A professional photographer knows this. The average person taking a snapshot, and the auto-exposure algorithm in their camera, does not.
Colleges began using test scores because different high schools (and in fact different teachers) awarded grades differently. An 4.0 in one high school might only be worth a 3.2 in another high school. A 4.0 by a student who took the "easy" teachers' courses might only be worth a 3.5 by a student who took the harder courses at the same high school. The test scores were used to try to normalize the grades.
The problem started when some college admissions staff got lazy, and started using test scores as a cutoff for admissions. That way they could circular file a bunch of the applicants without even having to read their application (but we still get to cash your application fee check kthxbye).
I don't think that assumption is right. Nearly all colleges get way more applicants than they accept. If what you say were true, they'd simply accept every applicant. (The ones which do accept everyone are usually scam colleges - soaking up subsidized Federal loan dollars in exchange for a middling or useless education, then leaving the student stuck with the bills.)
I suspect colleges aim for a certain uniformity in their students' capabilities so the professors don't have to dumb down lessons or worry that the material is too easy for some of the students. By stratifying education on the basis of student intelligence and capability, you can tailor the lessons and speed at which material is presented to better suit the strata of student ability the college has selected, thus reducing time wasted on trying to help poor students catch up, and time wasted by exceptional students on too-easy lessons. That improved efficiency also leads to cost reduction (professors don't have to spend as much time creating lessons, helping students, don't need to hire as many TAs, etc).
In this regard, TFA is walking a very fine line between claiming that test scores shouldn't exclude an otherwise exceptional student, and claiming that test scores shouldn't matter at all.
We went through this with Rambus. They joined JEDEC (a consortium of memory manufacturers setting future memory standards) and agreed to its terms of membership - mainly, members are not allowed to patent the memory standards being discussed. DDR was being discussed within JEDEC. Rambus went ahead and patented it, and sued the other JEDEC members for violating "their" patents.
After years of legal battles, the courts found that yes Rambus was guilty of violating JEDEC's membership agreement, and they were subject to whatever punishment they agreed to when they joined JEDEC. But that had nothing to do with the law, so the patents were valid (Rambus being the first to file). Meanwhile, since the JEDEC membership agreement didn't outline any punishment for violating the agreement, the only thing JEDEC could do was kick Rambus out.
Same thing here. An EULA or ToS is just a contract. If you violate it, you become subject to whatever punishment you agreed to when agreed to the contract. That does not automatically make it a violation of law however. It's only a violation of the law if the act was otherwise illegal. In Rambus' case, patenting stuff freely presented to you is not illegal. In Remini's case, downloading stuff you've been authorized to download is not illegal.
That's the way it's done in parliamentary elections. In the U.S. however, the Founding Fathers decided that legislators needed to be more in-touch with the voters they represented. This resulted in the idea of geographic representation. Each Congressman represents a geographical district - a specific chunk of the country whose voters elected that specific candidate.
Even the proportional representation system used in parliaments has mathematical problems. Imagine a parliament with 4 parties. Party A has 5 seats, Party B has 4 seats, Party C has 3 seats, party D has 1 seat. That's 13 seats in total, with 7 needed to decide a majority. Even though D ostensibly has 8% of the seats, it has zero power. It cannot change the outcome of any vote. A+B can generate a majority (9). A+C can generate a majority (8). B+C can generate a majority (7). Party D cannot change any of these outcomes, and thus it holds no power. It might as well not exist.
Mathematically, a representation system where the proportion of votes received is mapped onto the equivalent amount of power in government (ability to change decisions) would be ideal. But that would violate the "one person, one vote" ideal that people really seem beholden to, so will probably never happen.
Erm, the independent council in California for apolitically drawing congressional maps exists due to the work of Republicans. California was long gerrymandered by the Democrats. The stats I read in the 1980s were something like 55% of votes being cast in elections were for Democratic candidates, but 60% of its Assembly members and 62% of its Congressional representatives and State Senators were Democrats.
The Republicans finally managed to put a stop to it in 1990 by campaigning hard to get Pete Wilson elected as governor. After the 1990 census, the Democrats tried to gerrymander California's districts again. Wilson simply vetoed it. That led to a lot of drama which eventually ended up in the California Supreme Court, who ended up drawing the districts themselves just so there would be districts in time for the 1992 elections. Eventually, the two parties came up with an independent council with representatives from both parties drawing the maps, as a long-term solution.
Probably not to the extent it was done in NC (which was particularly egregious - I oppose election manipulation regardless of party). But Democrats gerrymandered their way into the longest continuous control of the House by one party in history - 40 years. That presented a quandary for the Republicans: How do you stop a system of electoral abuse which gives more votes to the abuser thus allowing them to propagate the abuse? The solution they came up with was to prioritize getting Republicans elected as governors in time for the 1990 census. Then as in California, those governors simply vetoed the districts drawn by the Democrats. (Pete Wilson was a shoe-in to be re-elected as Senator, but they decided winning the governor's office, even for just 4 years, was more important than a Senate seat in DC. They made a strategic gambit, giving up his Senate seat for a better chance at winning the governorship. His seat was eventually won by the Democrat who ran against him for governor - Feinstein, who still holds it today.)
That's why we have the situation today of gerrymandering favoring Republicans. Red governors of blue states are more common than blue governors of red states. The states where Democrats traditionally held power in the legislative branches, the Republicans managed to thwart their gerrymandering with vetoes after the 1990 or 2000 census. And the two parties were forced to compromise on an apolitical way to draw districts. So it's mostly the blue states which have these independent councils drawing districts (thanks to the work of Republicans). Meanwhile, the red states weren't affected as much by any Democrat governorship counter-strategy, so their districts are drawn the same old way as before - by the party controlling the state legislature. Which results in net gerrymandering favoring Republicans.
And as long as we're talking about electoral abuse, let's talk about illegal aliens and apportionment. You see, the Constitution simply states that House "Representatives shall be apportioned among the several States according to their respective numbers, counting the whole number of persons in each State, excluding Indians not taxed." Not voters, not citizens, but "persons." Non-taxed Indians are excluded, but not non-citizens. California's 2.3 million illegal aliens give it 3 extra House seats. And because they're almost unive
Or most of the OECD for that matter. The "Aristocracy" rich mostly exists in developing countries where they control the vast majority of the wealth. In those countries they're big fish in little ponds, and maintain their status not only by being rich, but by preventing others from becoming rich. The economy of these countries mostly consists of (by volume) the rich selling and buying to/from each other. The GDP per capita in these countries typically stagnates at around $15,000/yr or below.
The U.S., EU, etc. grew past this stage around the 1900s. Henry Ford accidentally stumbled upon this when he discovered that paying his workers above the prevailing wage actually resulted in more business for himself (because his workers could afford to buy the cars he was producing). That's what happens when you (1) put a worker in a productive job, and (2) pay them a fair wage for the productivity they're generating. Basically, when pay your workers less than a fair wage, you make money for yourself, but you stunt the economy. When you pay your workers a fair wage, you spend more money, but the economy blossoms. Usually more than enough to offset the extra money you spent paying your workers.
A market economy *wants* everyone to be as productive as they can, because the feedback effect of that maximizes average income. GDP per capita in these countries is typically $30,000/yr or higher because the vast majority of the population is contributing a meaningful amount of productivity to the economy. Consequently, the vast majority of the rich in these countries are rich from selling things to the middle class (who by population and aggregate income are much bigger than the richest 1%*). If the average income of the middle class decreases in these countries, it ends up hurting the rich too.
* IRS tax stats show that the top 1% only makes about 20% of the income in the U.S. So if they began buying and selling only amongst themselves and replacing everyone else with robots, that would result in about an 80% pay cut for themselves. The bulk of the country's income (73%) is in the $30k-$500k per year wage range, and it's in the best interest of the 1%ers in the U.S. to insure those people continue to have jobs.
Black feathers don't have this problem because you don't care what happens to the sunlight you're absorbing, only that it's all absorbed. You're not trying to direct the light at specific molecules within your structure like you are with PV panels. It would however work for thermal solar installations. All you care about in those is converting the sunlight into heat, and thus maximizing net absorption.
Amusingly, Wikipedia has a list of executive orders per year. Trump's 12-month executive order total equals Obama's last 15 months in office. Not 50 years. And Carter (37 years ago) was the last President to issue executive orders at a higher rate than Trump. If you're going to badmouth him, at least get your facts right.
I suspect Trump's rate is high just because this was his first year, and many of his executive orders were rolling back or modifying Obama's executive orders. Nothing hypocritical about that. As much as Democrats would've loved it if Trump had kept all of Obama's executive orders in force, we all knew that simply wasn't realistic. The next 3 years will tell if he's a hypocrite about executive orders.
The time constant for earthquake building codes is several decades long. Most of the California earthquake building codes stem from the 1933 Long Beach earthquake (mag 6.4). Most schools back then were brick buildings. Brick walls have no lateral strength - any sideways motion makes them simply fall over. Most of the schools in the Los Angeles areas were severely damaged or destroyed by the quake. But fortunately it hit around 6pm well after school hours. The realization that thousands if not tens of thousands of schoolchildren could've been killed if it had struck earlier in the day is what got the state to start imposing strict earthquake building codes (and banning unreinforced brick buildings).
But older buildings were grandfathered in. Brick structures weren't required to be reinforced until after 1971 San Fernando earthquake. That's when they began requiring steel bars be drilled into brick structures with square retaining plates at the ends, to keep the bricks from falling over under lateral load. You can frequently see these square plates in brick buildings on TV shows and movies, since many of the scenes are shot around Los Angeles. Most of the newer "brick" buildings you see in California are actually wood or concrete structures with a fake brick facade.
But we're learning new things from each quake. The 1994 Northridge quake in particular was eye-opening because one nearby seismograph measured vertical accelerations in excess of 1g (others measured close to 1g indicating it wasn't a fluke or malfunction). Up until then, it was thought that the maximum vertical acceleration possible from an earthquake was about 0.1g. Northridge shattered that assumption, and revised building codes take the possibility of strong vertical accelerations into account.
Anyhow, because of grandfathering, the time constant for most buildings to comply with improved regulations is several decades (average time until an old building is torn down and rebuilt). The difference in dollar damage between Loma Prieta and Northridge is almost entirely attributable to location, not because lessons learned from Loma Prieta got applied within 5 years. The Loma Prieta quake was centered in a forest about 20 miles from downtown San Jose, nearly 60 miles from San Francisco and Oakland. The Northridge quake was centered directly underneath the San Fernando Valley suburbs, and about 20 miles from downtown Los Angeles. Casualties would've been much higher had it struck during the middle of the day with the freeways jammed with traffic (a section of the I-10 overpass in LA collapsed). Fortunately it struck at 4:30 in the morning when most people were asleep in their relatively safe 1- and 2-story homes (the frequency of most earthquake shaking matches the resonance frequency of 3- and 4-story buildings, which are the majority of the ones which collapsed in both quakes).
By the time lessons from previous disasters get applied on a wide scale, increases in population and corresponding infrastructure pretty much offset any reduction in damage due to improved safety standards. A better statistic for comparing over time would be average cost of natural disasters per capita for the year (basically canceling out population growth). You could even make an argument for using the ratio of natural disaster cost to GDP (basically canceling out infrastructure growth). (And of course, normalizing for inflation is a given.)
This is basically what Steam In-Home Streaming does. The PC playing the game converts it into a h.264 video stream in real-time (using dedicated encoding hardware present on all modern GPUs). The device displaying the game just thinks it's playing back a streamed h.264 movie (using its hardware h.264 decoder which is also present on all modern GPUs, even on phones). The only bits that are missing are getting input from the display device and feeding it back to the game PC. That's probably the only reason the client hasn't come to tablets (and phones) yet - they generally don't come with keyboard+mouse or gamepad support.
That's the thing most people (except the Russians) don't seem to get about rocket launches. It's not a technology-bound problem, it's a physics-bound problem. You throw mass out the rear, and that pushes your rocket forward. The calculations for pretty much every type of rocket fuel were already done in the 1940s-1970s, and the fuels we've been using for the last 50 years are pretty much optimal.
Any new launch vehicle you design is just a pretty wrapper to hold the same fuel and throw it backwards the same way as old designs. So there's little to no performance to be gained by a new design. You need to economically justify the new design in other ways (e.g. reusable stages). Otherwise the most cost-effective strategy is to continue to use the old designs. Which is exactly what the Russians have done. And heaven forbid you spend $35 billion making a new design - that means amortized over an estimated 134 future launches (number of Space Shuttle missions), you've just increased the cost per launch by a quarter billion dollars.
Incidentally, STS was approved with exactly this in mind. The expectation was that by re-using the shuttles, money could be saved thus making launches cheaper in the long term. Unfortunately, that accounting calculation was made assuming about 50 shuttle launches per year. In reality it was less than 5 per year, meaning its facilities, equipment, and personnel costs per launch were 10x higher than projected, turning it from a cheaper-than-Apollo launch vehicle into the most expensive launch vehicle ever created.
The web is dying. It's being replaced by a bunch of proprietary apps you need to install to gain full access to any particular site on a mobile device. Facebook app, Google Maps app, YouTube app, Amazon app, etc. The reason the web worked on the PC was because you only had to install one app - a browser. Then you just typed in a different address for each site you wanted to visit into the browser, and like magic it was almost like you'd installed an app specific to that site. No more. I had to go through and purge a bunch of apps from my phone this weekend because its performance was bogging down (I had over 350 apps installed). On PC about 40%-50% of those apps were just bookmarks in my browser so took up no additional resources unless I was visiting the site at that moment. But on my phone they were constantly using up storage space, and in most cases constantly using RAM and network bandwidth (Amazon is particularly bad about this).
TFA is correct in that regard, and the danger of the shift back to walled gardens (AOL, GEnie, Prodigy back in the days before the web). Basically, rather than adapt to people shifting to mobile devices by creating mobile-friendly websites, companies have taken advantage of the shift to convince people to install a proprietary app on their mobile device instead. That app can then constantly run, monitor, and spy on the user even when it's not being used, and report it all back to the mothership.
Then it goes into a non sequitur about mesh networks saving us from all this, That's kinda like claiming SUVs will save us from proprietary dashboard navigation systems which cost hundreds of dollars to update. Those things have very little to do with each other. There is nothing inherently broken with TCP/IP or HTTP(S), except maybe HTML could be updated with some features to make for a more seamless browsing experience on mobile touch devices. If people are spending 90% of their online time on Google, Facebook, Amazon, moving away from the Internet won't prevent that. If mesh networks become popular, those three companies will just eventually set up mesh network sites, or modify their apps to also work over mesh networks.
A human looks at that picture, sees the banana and "thing" are sitting on a flat surface, and decides they must be about the same distance so their size in their picture is their actual scale. The banana is a lot bigger, so the human decides it is more important than the "thing".
An AI looks at that picture, sees the banana and "thing", but crucially doesn't estimate distance. Since the "thing" has a lot more detail the AI decides it's must be further away, and its greater detail means its the more important part of the picture, and the banana is just fluff in the background. And it gets lost trying to analyze the "thing".
The remote for the Apple TV absolutely sucks. Designers can't seem to get it through their heads that the #1 priority for TV remote is for it to be usable without looking away from the TV. So touchscreens and touchpads are out (except maybe for keyboard entry). You want tactile buttons so people can find the proper button to press without looking away from the screen. (The Logitech Harmonies make this mistake too.)
In 2016:
1604 Americans were murdered by knife. With a population of 323.1 million in 2016, that works out to a knife-murder rate of 0.496 per 100,000.
175 Canadians were murdered by knife. With a population of 36.3 million in 2016, that works out to a knife-murder rate of 0.482 per 100,000.
71 Australians were murdered by knife. With a population of 24.1 million, that works out to a knife-murder rate of 0.295 per 100,000.
213 people were murdered by knife in the UK. With a population of 65.6 million, that works out to a knife-murder rate of 0.325 per 100,000.
Despite the widespread availability of guns, Americans killed each other by knife at a higher rate than other countries. So it's not the guns. There's just something about Americans which make them more likely to kill each other, period. In that light, it's not at all surprising that U.S. police response is more aggressive than in other countries.
Um, that makes the class action not make sense. According to your reasoning, the threat of a class action caused the very behavior (Intel not informing vendors) the class action is purportedly trying to discourage.
Anyhow, more than likely nobody was harmed by the flaw, and nobody will be harmed by the flaw (unless they refuse to apply the patches, in which case it's on their own heads). Unfortunately for Intel, the harm will mainly come from reduced performance due to the fix. Still, the early reports I'm seeing say that the fix has little performance impact on everyday computing tasks. The large performance degradation seems to be limited mostly to cloud virtual data centers, so I'm not actually sure a class action (for the general population) is warranted.
This right here is IMHO the problem with society today. Not Trump. It's the asinine media which turns inanity into a superstar simply because they value flamboyance over substance. And the public which just eats this stuff up.
Trump being elected President is a symptom, not the problem. The proper reaction to someone being an idiot is to ignore them. Not to post a video of them on Facebook or YouTube just so you can comment "OMG can you believe this guy?" and be the first one to collect a million likes.
Speculative execution requires the CPU to do operations which it might be called upon to do in the future, then throw away the results if it turns out the code doesn't call for that operation. There's got to be an energy cost associated with that since you're making it do a bunch of operations it doesn't need to do. ARM SoCs are mostly used on mobile devices whose only power source is a battery. So it makes sense that they wouldn't incorporate energy-wasting "features" like speculation.
When Tim Cook became CEO, there was one model iPhone, and one model iPad. Today there are 3 iPhone models and 4 iPad models. Meanwhile, the number of Apple employees (non-sales) has not tripled. So there are fewer designers, artists, engineers assigned to each product. More delays are inevitable.
Personally I think it's a step in the right direction. Apple's product line under Jobs was woefully thin. Even at the height of its success the company was literally a single bad iPhone model away from bankruptcy.. Diversifying their product line was exactly what they needed for stability and expanding their customer base (for example, the many Android phone users held off switching until Apple released an iPhone with a bigger screen).
You're 90% of the way there. The problem here isn't that someone filed a copyright claim against white noise - something that inherently cannot be copyrighted.
The problem is that there is no disincentive, no punishment, for filing bogus copyright claims. The law has put the burden of proof entirely upon the purported infringer to prove he is innocent, none upon the accuser to prove an actual crime was committed.