It isn't "a thing" with generic $30 bluetooth earphones with a wire between them and a control / battery hanging from the wire. They only have 8 hours of listening time, so I just use two sets and swap when the battery is dead. They usually live about a year before the batteries are annoyingly short-lived, and I don't feel terrible about throwing a $30 set of earphones away. That's the cost of cutting the cord.
But, it is a thing when the price is over $100 and they only have 5 hours of listening time when brand new. They are pushing very tiny batteries too far to get even 5 hours and have not been showing the same battery lifetimes as standard bluetooth earphones.
At that point, the more than tripled cost per year is simply too high for most consumers and is "a thing". That cost could be greatly improved if they came up with a way to easily replace the fried batteries - perhaps the battery could just be a screw in plug... unscrew the old one and screw in the new one.
Just provide a convenient and cost-effective means of replacing the battery when it fails. The current suite of ridiculously priced wireless earbuds are all drastically limited in lifetime by how many recharges the batteries can take before they don't even last the evening.
Don't worry, in another ten generations of product releases, the batteries will last at least twice as long as the product ever will;)
Extrapolating the rate of deterioration in consumer appliance and device reliability from the past four decades forward, that will happen even with zero improvement in battery technology.
So does this mean that sites like Twitter or even Slashdot would have to pay news sites when users link to them in posts? It seems like a big article being linked with a few thousand retweets could be a crazy expensive hit. Perhaps links will no longer be allowed.
That is the wrong business decision. It is cheaper for Amazon to just take the loss.
Even retail stores like Walmart routinely write-off over 10% of their goods due to theft and fraud because it is cheaper to do so. If they buckled down, they could stop it, but the cost of doing so plus the fact that far fewer would shop there because of the inconveniences makes it a losing deal. This is why they don't even do simple things like open a box up and see what's in it when you make a return. Legit customers don't like being questioned.
In Amazon's case, if you both increase the delivery cost and make delivery inconvenient for the consumer, the losses would outweigh the gains.
They aren't stupid. They very carefully experiment with and measure these things to determine the sweet spot in the loss prevention cost vs. benefit equation.
It is very unlikely that they have any plans to expand this GPS sting thing. The purpose here is more likely on the prevention side. They are trying to create some publicity to make the least intelligent thieves that are working alone think twice.
We've recently seen stories in which a Tesla had video of a man breaking the window that included his face and license plate and where video doorbells were filming those that stole them. In both of those stories, the police had little to no interest in pursuing the case.
In both cases there was also a high probability that the individuals had committed strings of those crimes. Catching the individuals could prevent a lot of theft and damage. It is very possible they have priors and could get very significant time. If not, given that they know who is doing things, they should be able to do a bit of police work and prove the pattern. Who knows, perhaps they'll have a pile of doorbells in their home ready to sell on E-Bay or a little surveillance on the car could catch them doing other drive-by Tesla break-ins.
In both of those cases, I saw many responses on comment sites with worse things that police didn't care to pursue including grand theft auto and night-time residential B&Es.
I've personally had night-time B&Es twice. In both cases I knew who did them. One was an officer and another was someone who had a restraining order against them. Both managed to leave blood evidence. In both cases, the police didn't feel the case worth the time and cost of pursuit. My interpretation was that I was not in upper class neighborhoods where these things matter.
Yet, Amazon is able to get them to spend time on package theft? Why? Are they also paying them or giving a kickback perhaps? Just because they are Amazon? Citizens don't matter but companies do? What's the deal?
The vast majority of users serviced by AI systems have expectations by which they judge the service. When humans serve those users, good customer service usually dictates meeting those expectations. The expectations are largely driven by the microcultural background of the customer. At other times, they are expressed in the phrasing of the question, especially in context with the microcultural background. When a human has a great sense of a customer's expectations and utilizes it to meet those expectations more than other humans, they are considered talented at serving customers. This talented human service person is not expressing bias, they are responding to suspected bias in order to better meet customers' expectations.
I've seen many cases where people have lambasted a system for bias when they asked a leading question and successfully led the system. That is a user bias, not a system bias, unless the system gives the same answer when the question is asked in a manner or from an individual (assuming the system is good enough to take that into account) that leans the other way - it is good customer service.
There absolutely is bad bias in systems. In general, systems will improve in helping their users if bad bias is removed. So companies usually want to remove it when they recognize it. For example, facial recognition systems are improved when they are able to recognize faces from all racial backgrounds well. They are harder to create, not because of the training set, but because the job is harder to perform. Most people are far more capable of recognizing faces of people from their cultures than others because there are real differences in the parameters that are best used to determine uniqueness.
But AI systems cannot match humans in providing service if they are banned from recognizing human bias and utilizing that to provide expected responses.
Laws tend to be a broad sword. It is very likely that a law of this nature will not walk the fine lines that it would have to in order to serve us well.
If you're in a Comcast area, chances are they are already raping you because they are the only internet offering with decent speed. Youtube only competes against the cable portion of the bill. At $50, I believe that my Internet + Youtube would be greater than the Internet + TV package from my internet provider.
Thankfully, I didn't fall for the Youtube thing. It is just cable TV over internet. Nothing new. I haven't had cable TV since about 2004 and have made do without it.
The whole package would only offer me about eight channels that I'd ever watch. Not worth it. I might have considered it if it had CNN even though the initial channels included a lot of junk. Then, by the time they added CNN, it was obvious where things were headed. This latest addition doesn't include a single channel that I would have ever watched.
If someone wants my business, 10 channels that I can pick for $10 would be a good deal. I would go for that if it offered the 10 I want.
co-written by academics at Microsoft Research Asia in Beijing and researchers with affiliations to China’s National University of Defense Technology
When it comes to AI and a growing number of other areas, the Chinese are no slouches. This is why MS has research groups in China. They are taking advantage of Chinese talent. It is likely that MS gained more knowledge about AI from China in this case than they did from us, though I don't know what firewalls China may have required between MS's research groups in China and those in the US. There may very well be export controls on what that group can transmit to the home company.
the Tesla owner noted that he was simply informed that an officer might look into the incident
The emphasis on "might" is mine.
Given that replacing the rear window on a Model 3 will likely cost quite a bit more than your typical video doorbell, I'd be surprised if you even get a "might" from the officers on the doorbell. In fact, they'll likely be annoyed if you make them do a report.
If you can't agree to replace bad law with good law, there are two approaches to dealing with it.
The one that I would actually prefer assuming it can be done without a ton of extra cost is to fully enforce the bad law. Simply remove all prosecutorial discretion and enforce the hell out of it. This would be great for something like the law in West Virginia that makes it illegal to consent to sex when you're legally intoxicated and the person you're having sex with purchased the drinks. They actually have a few people in prison for 17 year sentences because they bought a few drinks and had sex. This is because their date rape drug law simply states that you have committed rape if you give someone an intoxicating substance and have sex with them and alcohol is an intoxicating substance. Obviously, full prosecution of this law as opposed to just prosecuting the ones you don't like would put a few 100 thousand people behind bars and force them to register for life as sex offenders.
In cases such as immigration laws though, yes, I think the best method of dealing with the bad law is to simply not enforce it most of the time. Illegal immigrants have been shown to be better behaved than American citizens on average. They improve us.
Also, the majority of those they have detained are not actually illegal immigrants. They are legally asking for asylum. If the administration was doing its job in providing for fast and orderly processing of asylum requests, no problem. They are clearly manufacturing the crisis as shown by the order of magnitude increase since taking over in 2016.
Acer Aspire V3-772g. With the single exception of the Linux graphics issues, I've been very happy with the machine. Even at 5 years, the 32GB, i4712 and Nvidia 750M still handle my demands with ease. The only upgrade I've made since purchase was to replace the TB hard drive with SSDs.
I probably confuse / stress the issue somewhat by connecting an external 32" monitor to the laptop and running dual screens. This required more changes to eliminate the severe tearing that you get from Optimus on the external monitor.
It was somewhat stable with bumblebee as you indicate until I hit Ubuntu 18.04. After that update, getting it back to the point where I could switch it on and off again took a lot of web searching and trials for a couple of days.
Regrettably, it never fully worked. For some reason, the NVidia GPU would randomly go out of control for a while, hit about 94 degrees C and then some type of protection mechanism on the motherboard would kill the power.
When 18.10 hit and thoroughly broke the system again, rather than digging through bulletin boards one more time to get the graphics to come up, I just gave up and switched back. I was also getting tired of the other glitches like the update window popping up saying that there were updates and not listing them (requiring a directory to be deleted once every couple of months or so to get it to rebuild its list). I guess it was a bit annoying to that once every few days it would want to reboot for one of the near daily updates.
I haven't had a problem with an update since switching back despite being in the preview program.
This is obviously step one to charging for new features. Next step will probably be to split out some very special feature improvements while continuing to give the bulk for free.
I'm not a total MS fan and ran Linux on my main machine from 2015-2018, but I do like the near zero-hassle updates and feature installations. It is a 32GB, i7 quad-core with SSD laptop that has aged well since purchase in 2014, but has Optimus mobile graphics (NVidia / Intel combo with Intel driving the display) which has very bad Linux support. I switched back to booting the Windows partition most of the time to avoid the constant hassles of straightening out update problems which often meant editing files from the terminal mode when the graphics subsystem failed.
can better gauge room reverberation and filter out echoes
For some reason, I read that statement and immediately felt that the speaker was more likely to be thinking about how much more accurately they can measure the room for many other reasons. I guess I'm getting more cynical.
Rooms have audio signatures. Those signatures are altered by how many people are in the room and where they are at. How much more information will they now be able to gain about the room and its contents?
The government would love it. All communication would of course be via SQL queries using blobs in the world's largest Oracle DB. Everything 100% recorded all of the time.
Wow! You have solved the problem! We can build the wall and it will pay for itself!
All we have to do is lay a very sturdy pipe along the length of the border that is very leakproof but transparent to radiation. Then we fill the pipe with high level radioactive waste and make sure that it is lethal to a distance of 30 feet or so. If we lay it just a little bit underground, it could even take care of all but the deepest tunnels. The really wild plus is that we have many billions of dollars available to pay for the disposal of our nuclear waste that will fully pay for the development of this wall! This strategy of actually making use of the waste instead of just burying it under some mountain is a total win-win!
Finally, after doing that to shut down the few illegal aliens who actually come into the country illegally, we could simply eliminate all tourist visas to get rid of the majority who enter legally and just stay after their visa expires. This will have the added benefit of removing the crowds from American attractions so that we can enjoy our own parks without having to be around all of those other people. Problem solved!
Well, 90% solved perhaps. We would then have to figure out how to get Americans that we approve of to actually take the trouble to produce enough offspring to maintain the population growth our economy requires for stability. That is perhaps harder than getting them to mow their own grass and pick their own fruit.
I get it. Troll. Nobody could seriously suggest that Facebook employ enough video experts to perform live extremism detection on every video posted. Assuming they actually had the money to do so (they don't), and enough capable people exist in the worldwide unemployed pool (they don't), I think Facebook would become the world's largest employer very quickly. I can't find precise statistics, but feel it likely after seeing their totals on posted videos over time that their max simultaneous video posting rate is in the millions and would thus require a number of reviewers in the millions.
If the government can define an algorithm with less than 0.001 percent false positives and 99 percent extremism detection rate that can be used to judge what content is extremist with no input other than the content after the moment the content first appears, go for it. I'm even OK if the algorithm can't be executed by an existing computer as long as the only criteria used are unchanging, concise, provably remove all subjectivity, and can be executed by someone of intelligence higher than about a standard deviation below average.
IOW, if you pick people randomly, train them, and have them fully independently follow the algorithm on 100,000 videos with no replay capability (they need to keep up with a live stream), there should not be more than one instance, (including those instances where there is an error in application of the algorithm) where a video's classification is improperly determined to be extremist. Even then, given the millions of daily livestreams, this algorithm would expose many people every day to the public humiliation of having a video improperly pulled for content that was not truly extremist. But, I guess that's the price of progress.
Your vision of the service would cost at least $1 per minute considering the expertise that would be required of the human reviewers. There is no AI that can reliably detect violence and extremism in a general fashion. Actually, $1 / minute could be really low for humans to agree to a job that guarantees jail time if you screw up or simply don't have the same opinion of what is violent or extremist as every possible randomly chosen jury of peers.
At the least, you'd be turning the clock back to the days when very few could exercise the right to communicate to the masses and making that very few be "those with money".
Most of the articles said something like three eggs a day which was more accurate. Some have said three or four a week. This is just a reflection on the quality of modern journalism, not the study.
This study flies in the face of current wisdom. Dietary cholesterol has a weak link to blood cholesterol. The cholesterol in your blood is mostly manufactured in your liver from sugars. This has been clearly demonstrated by the effectiveness of statins which reduce your liver's production of cholesterol, not absorption of cholesterol from foods. If dietary cholesterol was the source of the problem, statins would be ineffective.
Different people will have different results, but I have been able to reduce my LDL from 140 to below 70 by eliminating most sugars in my diet even though that forces me toward foods that raise dietary cholesterol. I took this approach when the side-effects of statins started accumulating. The results for myself prove that dietary cholesterol is not the problem. Eggs are one of the staples of my diet.
It is remarkable that even supposed scientists continually make the mistake of believing that everything has a first order cause. Most human maladies are not so simple.
True. But, at a 15% lab efficiency level, it is much less efficient at producing hydrogen than current solar to electric technologies. Furthermore, the compression/storage/fuel cell/inverter storage system required to utilize the hydrogen is far less efficient and more expensive than the charger/battery/inverter based storage system required to utilize the output of solar electric cells.
Higher efficiencies can be achieved with fuel cells if you also collect and use the waste heat, but that requires even more changes to the home. And, we could increase the systemic efficiency of batteries and inverters in the same way (collect the waste heat and use it for something like boosting the water heater efficiency), but have not found the cost worth it given how high the system efficiency already is.
In short, this would seem to need much more efficiency throughout the system, not only at production, before beating existing renewable technologies even for utility use. It is even further away from enabling transition to a more reliable fully distributed energy industry. Of course, that is why they continue to look at it. They are desperate to maintain the centralized energy production model.
Slashdot Mirror
It isn't "a thing" with generic $30 bluetooth earphones with a wire between them and a control / battery hanging from the wire. They only have 8 hours of listening time, so I just use two sets and swap when the battery is dead. They usually live about a year before the batteries are annoyingly short-lived, and I don't feel terrible about throwing a $30 set of earphones away. That's the cost of cutting the cord.
But, it is a thing when the price is over $100 and they only have 5 hours of listening time when brand new. They are pushing very tiny batteries too far to get even 5 hours and have not been showing the same battery lifetimes as standard bluetooth earphones.
At that point, the more than tripled cost per year is simply too high for most consumers and is "a thing". That cost could be greatly improved if they came up with a way to easily replace the fried batteries - perhaps the battery could just be a screw in plug... unscrew the old one and screw in the new one.
Just provide a convenient and cost-effective means of replacing the battery when it fails. The current suite of ridiculously priced wireless earbuds are all drastically limited in lifetime by how many recharges the batteries can take before they don't even last the evening.
Don't worry, in another ten generations of product releases, the batteries will last at least twice as long as the product ever will ;)
Extrapolating the rate of deterioration in consumer appliance and device reliability from the past four decades forward, that will happen even with zero improvement in battery technology.
So does this mean that sites like Twitter or even Slashdot would have to pay news sites when users link to them in posts? It seems like a big article being linked with a few thousand retweets could be a crazy expensive hit. Perhaps links will no longer be allowed.
That is the wrong business decision. It is cheaper for Amazon to just take the loss.
Even retail stores like Walmart routinely write-off over 10% of their goods due to theft and fraud because it is cheaper to do so. If they buckled down, they could stop it, but the cost of doing so plus the fact that far fewer would shop there because of the inconveniences makes it a losing deal. This is why they don't even do simple things like open a box up and see what's in it when you make a return. Legit customers don't like being questioned.
In Amazon's case, if you both increase the delivery cost and make delivery inconvenient for the consumer, the losses would outweigh the gains.
They aren't stupid. They very carefully experiment with and measure these things to determine the sweet spot in the loss prevention cost vs. benefit equation.
It is very unlikely that they have any plans to expand this GPS sting thing. The purpose here is more likely on the prevention side. They are trying to create some publicity to make the least intelligent thieves that are working alone think twice.
We've recently seen stories in which a Tesla had video of a man breaking the window that included his face and license plate and where video doorbells were filming those that stole them. In both of those stories, the police had little to no interest in pursuing the case.
In both cases there was also a high probability that the individuals had committed strings of those crimes. Catching the individuals could prevent a lot of theft and damage. It is very possible they have priors and could get very significant time. If not, given that they know who is doing things, they should be able to do a bit of police work and prove the pattern. Who knows, perhaps they'll have a pile of doorbells in their home ready to sell on E-Bay or a little surveillance on the car could catch them doing other drive-by Tesla break-ins.
In both of those cases, I saw many responses on comment sites with worse things that police didn't care to pursue including grand theft auto and night-time residential B&Es.
I've personally had night-time B&Es twice. In both cases I knew who did them. One was an officer and another was someone who had a restraining order against them. Both managed to leave blood evidence. In both cases, the police didn't feel the case worth the time and cost of pursuit. My interpretation was that I was not in upper class neighborhoods where these things matter.
Yet, Amazon is able to get them to spend time on package theft? Why? Are they also paying them or giving a kickback perhaps? Just because they are Amazon? Citizens don't matter but companies do? What's the deal?
The vast majority of users serviced by AI systems have expectations by which they judge the service. When humans serve those users, good customer service usually dictates meeting those expectations. The expectations are largely driven by the microcultural background of the customer. At other times, they are expressed in the phrasing of the question, especially in context with the microcultural background. When a human has a great sense of a customer's expectations and utilizes it to meet those expectations more than other humans, they are considered talented at serving customers. This talented human service person is not expressing bias, they are responding to suspected bias in order to better meet customers' expectations.
I've seen many cases where people have lambasted a system for bias when they asked a leading question and successfully led the system. That is a user bias, not a system bias, unless the system gives the same answer when the question is asked in a manner or from an individual (assuming the system is good enough to take that into account) that leans the other way - it is good customer service.
There absolutely is bad bias in systems. In general, systems will improve in helping their users if bad bias is removed. So companies usually want to remove it when they recognize it. For example, facial recognition systems are improved when they are able to recognize faces from all racial backgrounds well. They are harder to create, not because of the training set, but because the job is harder to perform. Most people are far more capable of recognizing faces of people from their cultures than others because there are real differences in the parameters that are best used to determine uniqueness.
But AI systems cannot match humans in providing service if they are banned from recognizing human bias and utilizing that to provide expected responses.
Laws tend to be a broad sword. It is very likely that a law of this nature will not walk the fine lines that it would have to in order to serve us well.
If you're in a Comcast area, chances are they are already raping you because they are the only internet offering with decent speed. Youtube only competes against the cable portion of the bill. At $50, I believe that my Internet + Youtube would be greater than the Internet + TV package from my internet provider.
Thankfully, I didn't fall for the Youtube thing. It is just cable TV over internet. Nothing new. I haven't had cable TV since about 2004 and have made do without it.
The whole package would only offer me about eight channels that I'd ever watch. Not worth it. I might have considered it if it had CNN even though the initial channels included a lot of junk. Then, by the time they added CNN, it was obvious where things were headed. This latest addition doesn't include a single channel that I would have ever watched.
If someone wants my business, 10 channels that I can pick for $10 would be a good deal. I would go for that if it offered the 10 I want.
co-written by academics at Microsoft Research Asia in Beijing and researchers with affiliations to China’s National University of Defense Technology
When it comes to AI and a growing number of other areas, the Chinese are no slouches. This is why MS has research groups in China. They are taking advantage of Chinese talent. It is likely that MS gained more knowledge about AI from China in this case than they did from us, though I don't know what firewalls China may have required between MS's research groups in China and those in the US. There may very well be export controls on what that group can transmit to the home company.
A couple of days ago, Teslarati reported a case in which a Model 3 caught video of a thief breaking in that showed both his face and his license plate. From the article:
the Tesla owner noted that he was simply informed that an officer might look into the incident
The emphasis on "might" is mine.
Given that replacing the rear window on a Model 3 will likely cost quite a bit more than your typical video doorbell, I'd be surprised if you even get a "might" from the officers on the doorbell. In fact, they'll likely be annoyed if you make them do a report.
From the article:
Just four years after his arrest, Jack is now working at an advanced level, carrying out processes like penetration testing
In other news, prostitutes are now screaming discrimination as another group of criminals get off and actually get paid for penetration testing.
If you can't agree to replace bad law with good law, there are two approaches to dealing with it.
The one that I would actually prefer assuming it can be done without a ton of extra cost is to fully enforce the bad law. Simply remove all prosecutorial discretion and enforce the hell out of it. This would be great for something like the law in West Virginia that makes it illegal to consent to sex when you're legally intoxicated and the person you're having sex with purchased the drinks. They actually have a few people in prison for 17 year sentences because they bought a few drinks and had sex. This is because their date rape drug law simply states that you have committed rape if you give someone an intoxicating substance and have sex with them and alcohol is an intoxicating substance. Obviously, full prosecution of this law as opposed to just prosecuting the ones you don't like would put a few 100 thousand people behind bars and force them to register for life as sex offenders.
In cases such as immigration laws though, yes, I think the best method of dealing with the bad law is to simply not enforce it most of the time. Illegal immigrants have been shown to be better behaved than American citizens on average. They improve us.
Also, the majority of those they have detained are not actually illegal immigrants. They are legally asking for asylum. If the administration was doing its job in providing for fast and orderly processing of asylum requests, no problem. They are clearly manufacturing the crisis as shown by the order of magnitude increase since taking over in 2016.
Acer Aspire V3-772g. With the single exception of the Linux graphics issues, I've been very happy with the machine. Even at 5 years, the 32GB, i4712 and Nvidia 750M still handle my demands with ease. The only upgrade I've made since purchase was to replace the TB hard drive with SSDs.
I probably confuse / stress the issue somewhat by connecting an external 32" monitor to the laptop and running dual screens. This required more changes to eliminate the severe tearing that you get from Optimus on the external monitor.
It was somewhat stable with bumblebee as you indicate until I hit Ubuntu 18.04. After that update, getting it back to the point where I could switch it on and off again took a lot of web searching and trials for a couple of days.
Regrettably, it never fully worked. For some reason, the NVidia GPU would randomly go out of control for a while, hit about 94 degrees C and then some type of protection mechanism on the motherboard would kill the power.
When 18.10 hit and thoroughly broke the system again, rather than digging through bulletin boards one more time to get the graphics to come up, I just gave up and switched back. I was also getting tired of the other glitches like the update window popping up saying that there were updates and not listing them (requiring a directory to be deleted once every couple of months or so to get it to rebuild its list). I guess it was a bit annoying to that once every few days it would want to reboot for one of the near daily updates.
I haven't had a problem with an update since switching back despite being in the preview program.
This is obviously step one to charging for new features. Next step will probably be to split out some very special feature improvements while continuing to give the bulk for free.
I'm not a total MS fan and ran Linux on my main machine from 2015-2018, but I do like the near zero-hassle updates and feature installations. It is a 32GB, i7 quad-core with SSD laptop that has aged well since purchase in 2014, but has Optimus mobile graphics (NVidia / Intel combo with Intel driving the display) which has very bad Linux support. I switched back to booting the Windows partition most of the time to avoid the constant hassles of straightening out update problems which often meant editing files from the terminal mode when the graphics subsystem failed.
I guess nothing is really free.
can better gauge room reverberation and filter out echoes
For some reason, I read that statement and immediately felt that the speaker was more likely to be thinking about how much more accurately they can measure the room for many other reasons. I guess I'm getting more cynical.
Rooms have audio signatures. Those signatures are altered by how many people are in the room and where they are at. How much more information will they now be able to gain about the room and its contents?
The government would love it. All communication would of course be via SQL queries using blobs in the world's largest Oracle DB. Everything 100% recorded all of the time.
Wow! You have solved the problem! We can build the wall and it will pay for itself!
All we have to do is lay a very sturdy pipe along the length of the border that is very leakproof but transparent to radiation. Then we fill the pipe with high level radioactive waste and make sure that it is lethal to a distance of 30 feet or so. If we lay it just a little bit underground, it could even take care of all but the deepest tunnels. The really wild plus is that we have many billions of dollars available to pay for the disposal of our nuclear waste that will fully pay for the development of this wall! This strategy of actually making use of the waste instead of just burying it under some mountain is a total win-win!
Finally, after doing that to shut down the few illegal aliens who actually come into the country illegally, we could simply eliminate all tourist visas to get rid of the majority who enter legally and just stay after their visa expires. This will have the added benefit of removing the crowds from American attractions so that we can enjoy our own parks without having to be around all of those other people. Problem solved!
Well, 90% solved perhaps. We would then have to figure out how to get Americans that we approve of to actually take the trouble to produce enough offspring to maintain the population growth our economy requires for stability. That is perhaps harder than getting them to mow their own grass and pick their own fruit.
/s
I completely retrofitted my house for about $2 per socket a couple of years ago with no failures. Maybe you paid too much.
Sure. But why not do both? My bill in Florida for a 1600 sq foot home is $80 per month on average. If I could make it even less, I still would.
I get it. Troll. Nobody could seriously suggest that Facebook employ enough video experts to perform live extremism detection on every video posted. Assuming they actually had the money to do so (they don't), and enough capable people exist in the worldwide unemployed pool (they don't), I think Facebook would become the world's largest employer very quickly. I can't find precise statistics, but feel it likely after seeing their totals on posted videos over time that their max simultaneous video posting rate is in the millions and would thus require a number of reviewers in the millions.
If the government can define an algorithm with less than 0.001 percent false positives and 99 percent extremism detection rate that can be used to judge what content is extremist with no input other than the content after the moment the content first appears, go for it. I'm even OK if the algorithm can't be executed by an existing computer as long as the only criteria used are unchanging, concise, provably remove all subjectivity, and can be executed by someone of intelligence higher than about a standard deviation below average.
IOW, if you pick people randomly, train them, and have them fully independently follow the algorithm on 100,000 videos with no replay capability (they need to keep up with a live stream), there should not be more than one instance, (including those instances where there is an error in application of the algorithm) where a video's classification is improperly determined to be extremist. Even then, given the millions of daily livestreams, this algorithm would expose many people every day to the public humiliation of having a video improperly pulled for content that was not truly extremist. But, I guess that's the price of progress.
Your vision of the service would cost at least $1 per minute considering the expertise that would be required of the human reviewers. There is no AI that can reliably detect violence and extremism in a general fashion. Actually, $1 / minute could be really low for humans to agree to a job that guarantees jail time if you screw up or simply don't have the same opinion of what is violent or extremist as every possible randomly chosen jury of peers.
At the least, you'd be turning the clock back to the days when very few could exercise the right to communicate to the masses and making that very few be "those with money".
Most of the articles said something like three eggs a day which was more accurate. Some have said three or four a week. This is just a reflection on the quality of modern journalism, not the study.
This study flies in the face of current wisdom. Dietary cholesterol has a weak link to blood cholesterol. The cholesterol in your blood is mostly manufactured in your liver from sugars. This has been clearly demonstrated by the effectiveness of statins which reduce your liver's production of cholesterol, not absorption of cholesterol from foods. If dietary cholesterol was the source of the problem, statins would be ineffective.
Different people will have different results, but I have been able to reduce my LDL from 140 to below 70 by eliminating most sugars in my diet even though that forces me toward foods that raise dietary cholesterol. I took this approach when the side-effects of statins started accumulating. The results for myself prove that dietary cholesterol is not the problem. Eggs are one of the staples of my diet.
It is remarkable that even supposed scientists continually make the mistake of believing that everything has a first order cause. Most human maladies are not so simple.
True. But, at a 15% lab efficiency level, it is much less efficient at producing hydrogen than current solar to electric technologies. Furthermore, the compression/storage/fuel cell/inverter storage system required to utilize the hydrogen is far less efficient and more expensive than the charger/battery/inverter based storage system required to utilize the output of solar electric cells.
Higher efficiencies can be achieved with fuel cells if you also collect and use the waste heat, but that requires even more changes to the home. And, we could increase the systemic efficiency of batteries and inverters in the same way (collect the waste heat and use it for something like boosting the water heater efficiency), but have not found the cost worth it given how high the system efficiency already is.
In short, this would seem to need much more efficiency throughout the system, not only at production, before beating existing renewable technologies even for utility use. It is even further away from enabling transition to a more reliable fully distributed energy industry. Of course, that is why they continue to look at it. They are desperate to maintain the centralized energy production model.