I get it that solar is a good idea to be considered for each project. On the other hand, diesel also need to be considered as an option. In this case, it seems like reliable high power-to-weight engines would be the better fit.
And given that, if it worked, this thing would have a massive positive environmental impact, I can't see why the insistence on not using the right tool.
There are, but sadly the market for good new on the internet is like "Little boy surprised by neighbors with boy" or "firefighter rescues women from fire, finds out she was nurse at his childbirth".
Good news on the scale of civilization-level progress over decades, not so much. . .
There's an interesting quirk in human psychology that makes negative facts and news seem more salient than positive ones. For media that thrives on reader attention (and that's both new and old media), this naturally leads to more emphasis on the negative.
I think this is a bias worth noting and pushing back on. The world is pretty far from perfect, but there's also huge helpings of good news all around us.
Extreme global poverty continues its decline, although it's getting harder to make progress on that front
The US death rate from cancer continues its steady yearly drop. Cumulatively, this has prevented hundreds of thousands of deaths
The pack of criminals who made a wholesale business of taking sex slaves in war lost their last city
The world continues its steady march towards universal literacy. You can't embed pictures in/. (for reaaaalllly goatse reasons) but the figures here are really striking
Most of these (Daesh not withstanding, but threw them in just because they were really vile) follow the same pattern: slow but steady progress. It's hardly clickbait -- in fact these are not even specific events you can point to, they are trends seen on the scale of decades. And on the scale of decades, the world is consistently becoming a less-bad place.
That you cannot see a use-case for the aircraft says more about your imagination or experience of the world than it does about the actual utility of the vehicle.
It also says a lot about the current market and regulatory structures and their impact on scaling down. On the US side of the pond (YMMV elsewhere), short trips like this would be quite expensive at the endpoints, even using small regional airports. The regulatory costs would likewise be very hard to recoup at small scale.
Bringing a product to market is (sadly) a lot harder than just making the technology work.
Nobody blames the 18-wheeler itself if the driver is too incompetent to load or drive it properly under most conditions
WTF are you talking about? We've long since mandated all kinds of safety features and inspections on 18-wheelers. It's not lawful to load an 18 wheeler above the Federal maximum gross vehicle weight of 80,000lbs, no matter how skilled the driver is. You can't drive without ABS either, even if your driver is so good they don't need it. There are mandatory rest times and maximum duty hours.
So yeah, you're pedantically right that we don't "blame the 18-wheeler" in a crash, in the sense that inanimate objects are really bad targets of blame since they have no agency. Of course the driver is to blame. But we sure as hell go about figuring out how the next generation of 18 wheelers and associated regulation can reduce either the number of mistakes that drivers make or mitigate the negative consequences of those mistakes.
Same thing ought to be in software -- we should work towards memory & type safety that reduces the number of mistakes that programmers make and we should improve isolation and hardening so that the negative consequences of those mistakes are contained. Even in C/C++ there are lots of modern safety features being built in, better verifiers like ASAN and UBSAN. And of course the development of Rust/Go/Java/C# and all the other safe languages.
This isn't really a novel concept. Ever since we decided cars should have crumple zones and safety belts, the road fatalities per mile driven has gone down. Now we have blind spot detection and auto-braking in accidents coming online. In a decade or two we might have self-driving cars. No one is blaming a 1953 Ford Thunderbird for crashing in the rain, but I sure as hell wouldn't drive one in a storm today.
The reason for this state of affairs, according to Gemini, is the lack of U.S. merchant complianceâ"too many of them still use the mag-stripe function at PoS terminals.... If the EMV functionalities are not fully deployed, the track 1 and track 2 data stolen from the chip transaction can be easily encoded by the fraudster onto any magnetic strip.
So to get this straight, you get a plastic card, it supports both the newfangled way and the old-and-busted way (or else people would be up in arms that it wasn't compatible with 100% of readers). By the way, the new hotness is just the old version plus a transaction-unique cryptographic token. Now, when this is deployed, people figure out -- they skim the new way and then use it to create mag-stripe cards that can be used only at places that don't require a chip. But somehow this is a problem with the chip cards?
Nooooo, it's a problem with places that don't require a chip. We've known since the 80s that you can copy a magnetic strip with a 2-tape boombox (seriously, it will work).
TLDR: There's nothing wrong with the chip cards themselves. But there is something wrong with merchants that haven't upgraded to EMV, and definitely something wrong with/. editors that write a completely ass-backwards headline.
Non-removable storage is a deal killer. I don't know what gets stored on local computers so it MUST to be wiped before it goes off site for repair.
Wiping data sounds good, but what does that have to do with removable storage? You can boot from USB (or to the recovery partition) and do an erase right there.
Don't you have to comply with any sort of security polices at Apple?
I imagine yes, but that has nothing to do with the old style of punching holes through magnetic media.
I've worked with more than one company now whose architecture is such that even if a band of thieves broke into the DC and stole the entire rack, they'd have nothing of value on the disks. All the data were encrypted at rest and the keys held only in RAM. As such, they reviewed and removed the 'erase drives' provision entirely. Decommissioning a machine just means ensuring it powercycles.
[ Now, in theory, if they could steal them and keep them powered, then break in, sure. Or if they could pwn the machines before stealing them, they could have it write the keys persistently -- but in that case, why bother physically stealing them at all? ]
I am perfectly fine with the Sun's zenith being at 1:30pm, as it is during the Summer where I live.
In the Northern US, there's a bit more than 9 hours of daylight at the winter solstice. Centering it around 1PM would make it light from 8:30AM to 5:30PM, which would make morning commutes a slog.
Yeah, and change when people normally show up to work. And move lunch too. And store openings and TV schedules too.
At some point modifying our arbitrary mapping of the passage of time into arabic numerals is easier than changing everything else that we pick arbitrary numbers for.
I'm a consumer in love with it, and I understand it's never going to be as fast as wired. I don't use it exclusively, but it's really convenient in cases where I just don't need that kind of speed.
For example, in the car, I have both a 2A wired charger and a 500mA inductive. If my battery is in reasonable shape, I just pop the phone in the inductive charger/holder thingy and off I go. It's easily enough to run Waze and stream bluetooth music without draining the battery at all, and if I really need 2A (or it's going to be a long trip) I can always just plug it in.
Second example, overnight, I have an inductive charger in a little divot on the nightstand. Even from ~0%, 500mA over a night's sleep is enough to wake up at 100%.
Finally, your claim about the power bill going up is absolutely absurd because devices are basically zero fraction of anyone's electricity bill. The average consumption is 900 KWH/month. A phone is ~3000mAH. Even if you charged a phone once a day on an inductive charger at 10% efficiency, that would be (rounding up) ~1KWH = (3000mAH * 10x loss * 30 days/month) or ~0.1% of your total bill. You'd be paying like $0.20/month to charge the phone, of which 18Â was wasted.
Modern inductive chargers don't have excellent efficiency by the way, but they're better than 10%:-)
That is incorrect. "Their" is plural of his/her/its. We know his sex. Their is NOTHING wrong with using the correct pronoun that corresponds with his known nature - It is the suppression of doing so that is becoming the insane norm.
First of all, singular they has a long history within the English language dating back at least as far the Bishops (1568) and King James (1611) Bibles, Shakespeare, Emily Dickinson, Jane Austen and many other notable authors.
By all means, avoid using they with singular antecedents in your own writing and speaking if you feel you cannot bear it. Language Log is not here to tell you how to write or speak. But don't try to tell us that it's grammatically incorrect. Because when a construction is clearly present several times in Shakespeare's rightly admired plays and poems, and occurs in the carefully prepared published work of just about all major writers down the centuries, and is systematically present in the unreflecting conversational usage of just about everyone including Sean Lennon, then the claim that it is ungrammatical begins to look utterly unsustainable to us here at Language Log Plaza. This use of they isn't ungrammatical, it isn't a mistake, it's a feature of ordinary English syntax that for some reason attracts the ire of particularly puristic pusillanimous pontificators, and we don't buy what they're selling.
Second, the sneering and incorrect hyper-grammar-policing of a historically acceptable construction is bad enough, but did you really have to do it in a post mistaking "there" and "their" in the second sentence? Because that's not some marginal or debatable rule of grammar, that's actually two different words with totally different meanings. Even Safari's god-awful grammar checker flags that one as questionable . . .
The funny thing is that Oracle could get back into many peoples' good graces. If they offered ZFS under the GPL and allowed it to become part of the default Linux kernel, this would be one of the biggest enterprise issues that would get solved.
It's too late. If they had done this before BTRFS became production-worthy, it would have taken the air out of BTRFS. Now it's got momentum.
I don't have many hard and fast office rules, but here's a soft one -- if your keyboard is loud enough to bother others, you will politely be asked to select a quieter, which will gladly be provided free of charge.
By all mean, please do enjoy your unholy racket with its excellent tactile feel and whatever, wherever you like except not here.
User's preferences can never be the problem. This is a contradiction in terms. Things are measured by what people think is better. The sooner you lose this attitude that there's any other measure, the sooner you can build things that people actually want to use.
This is not to say that there isn't a dialogue between the person building widgets and the users consuming them. A lot of building and iterating involves not just blindly giving people what they want, but in exploring what their underlying need is and trying to fulfill it. It's not just 'a faster horse'.
But still, in the final analysis, the builder is successful if and only if the user thinks the widget is good. That's the metric.
Fun story time related by a colleague. A pretty common piece of software (hint: there's probably one running within a few hundred yards of you) had an elusive bug. But as the parent noted, printf caused the problem to go away, and it was suspected because it caused synchronization on stdout. Unlike the parent, the developers didn't have time to actually implement a buffered-log solution to figure this out, so they the obviously-logical thing -- they replaced all the printf calls with barrier() and shipped it. It's still running like this today.
Another good one, I worked with someone who would log everything all the time by fprintfing to a high-numbered pipe. When I asked him, he gave a few advantages that still ring partially true (depends on context): first, he said, I can get the log from any running instance without even stopping by d-tracing the system call. But most critically, he said, all the formatting happens in userland and only after the syscall does the kernel actually realize that there's nothing on the other end of the pipe and drop the write. That means, he reasoned, that the release/debug versions would always have very close behavior and would avoid the class of 'bugs that don't reproduce in debug build'. As with the other story, to this day, there's a slew of machines out there, formatting and writing log messages to a pipe that's never open.
"Imagine an artificial-intelligence-driven military drone capable of autonomously patrolling the perimeter of a country or region and deciding who lives and who dies, without a human operator. Now do the same with tanks, helicopters and biped/quadruped robots." A United Nations conference recently decided not to ban these weapons systems outright
Forgetting for a moment the logistics of actually enforcing such a ban, what would such a ban actually entail?
Presumably the ban would not apply to the mere act of developing a drone/robot or the AI to make it move around and perceive the environment. That is, it's not a ban on any of the fundamental building blocks of robotics or AI. Nor would it apply to actually building a search-and-resce bot or teaching it to navigate unknown or difficult terrain. So countries interested in such things would continue to build them just short of actually being weapons (with laughable charades)
After that, what's left? Attaching a gun to it? IFF? It seems fairly clear that this is not the technical challenge here. So even if there's a 'ban', countries will be free to develop technologies and, if they master those problems, will credibly be a screwdriver away" from having full on AI weapons.
Not tax evasion but high taxes that push people to gray market. Especailly that people don't see a point to contribute their tax share since they don't get much out of it. Most of money disapears with bureaucracy and goverment. This is pattern well established around post socialistic countries.
Tax evasion means that, for a given level of government revenue, nominal rates must be higher. Higher nominal rates means, for a given level of risk of getting caught, tax evasion is more rewarding.
It's a circular problem, one in which unfortunately Greece is stuck in a bad attractor. There is also a virtuous version of the cycle, where effectively cracking down on evasion allows lower nominal rates, which in turn lowers the incentive to evade. It's debatable whether it's actually possible to drag Greece out of the bad attractor and into the good one, especially given that some people believe that it's become embedded in Greek culture.
And yes, no one is claiming that Greece's government has been a paragon of efficiency and integrity, or that its spending choices have been ideal. But that's kind of apart from how revenue is generated. You don't have to agree with any of that (either entirely or on the margin) to see that the tax collection problem is really hurting them.
I've also seen as well as heard of job interviews where the person didn't get the job because the candidate actually looked at their watches... in the interview. That's a huge "OH NO HE DIDN'T".
Boy am I glad I don't work for any of those places.
but that's not the same as saying we're born to be couch potatoes; if that were true then how do you explain the existence of marathon runners?
I'm not saying I agree with TFA's conclusion, but you can easily explain the existence of marathon runners as being a completely insignificant fraction of the population.
It's not reasonable to say that researchers cannot claim things about the majority of humans just because you can dredge up a counter-example. It's implied they are saying that this is true on average and not for literally 100% of people.
No one is disagreeing with that or saying there shouldn't be an API that allows you to knowingly chose to share all your email data with whoever you want. But just like the grocery store can't sell a bag chips without disclosing how many grams of fat are in it, so too should the person offering the API disclose clearly what data are shared, whether they are shared-onwards to third parties, whether they are persisted and how to remove yourself.
Maybe another way to put it is that you can't decide on what level of privacy you are sacrificing if the service doesn't clearly explain what you are singing up for, or worse, claims one thing but then does another. Uninformed consent is not valid consent.
If Google is enforcing those disclosure, opt-out and accuracy requirements, great. No problem. But they don't necessarily have an interest to do so too thoroughly, not least because it requires lots of human intervention, which isn't their strong suit.
Slashdot Mirror
I get it that solar is a good idea to be considered for each project. On the other hand, diesel also need to be considered as an option. In this case, it seems like reliable high power-to-weight engines would be the better fit.
And given that, if it worked, this thing would have a massive positive environmental impact, I can't see why the insistence on not using the right tool.
There are, but sadly the market for good new on the internet is like "Little boy surprised by neighbors with boy" or "firefighter rescues women from fire, finds out she was nurse at his childbirth".
Good news on the scale of civilization-level progress over decades, not so much. . .
There's an interesting quirk in human psychology that makes negative facts and news seem more salient than positive ones. For media that thrives on reader attention (and that's both new and old media), this naturally leads to more emphasis on the negative.
I think this is a bias worth noting and pushing back on. The world is pretty far from perfect, but there's also huge helpings of good news all around us.
Most of these (Daesh not withstanding, but threw them in just because they were really vile) follow the same pattern: slow but steady progress. It's hardly clickbait -- in fact these are not even specific events you can point to, they are trends seen on the scale of decades. And on the scale of decades, the world is consistently becoming a less-bad place.
If this were universally accepted, I would gladly remove one more piece of plastic from my wallet.
It also says a lot about the current market and regulatory structures and their impact on scaling down. On the US side of the pond (YMMV elsewhere), short trips like this would be quite expensive at the endpoints, even using small regional airports. The regulatory costs would likewise be very hard to recoup at small scale.
Bringing a product to market is (sadly) a lot harder than just making the technology work.
Indeed. And our SOP for compliance states that we will physically destroy the encryption keys and that completes the procedure of destroying the data.
YMMV because you might have different sit policies, but that's ours :-)
WTF are you talking about? We've long since mandated all kinds of safety features and inspections on 18-wheelers. It's not lawful to load an 18 wheeler above the Federal maximum gross vehicle weight of 80,000lbs, no matter how skilled the driver is. You can't drive without ABS either, even if your driver is so good they don't need it. There are mandatory rest times and maximum duty hours.
So yeah, you're pedantically right that we don't "blame the 18-wheeler" in a crash, in the sense that inanimate objects are really bad targets of blame since they have no agency. Of course the driver is to blame. But we sure as hell go about figuring out how the next generation of 18 wheelers and associated regulation can reduce either the number of mistakes that drivers make or mitigate the negative consequences of those mistakes.
Same thing ought to be in software -- we should work towards memory & type safety that reduces the number of mistakes that programmers make and we should improve isolation and hardening so that the negative consequences of those mistakes are contained. Even in C/C++ there are lots of modern safety features being built in, better verifiers like ASAN and UBSAN. And of course the development of Rust/Go/Java/C# and all the other safe languages.
This isn't really a novel concept. Ever since we decided cars should have crumple zones and safety belts, the road fatalities per mile driven has gone down. Now we have blind spot detection and auto-braking in accidents coming online. In a decade or two we might have self-driving cars. No one is blaming a 1953 Ford Thunderbird for crashing in the rain, but I sure as hell wouldn't drive one in a storm today.
But you can't clone an EMV card anyway. So the best a waiter can do with a clone is use it at a merchant that still accepts magstripe.
Once that goes away, the problem goes away.
So to get this straight, you get a plastic card, it supports both the newfangled way and the old-and-busted way (or else people would be up in arms that it wasn't compatible with 100% of readers). By the way, the new hotness is just the old version plus a transaction-unique cryptographic token. Now, when this is deployed, people figure out -- they skim the new way and then use it to create mag-stripe cards that can be used only at places that don't require a chip. But somehow this is a problem with the chip cards?
Nooooo, it's a problem with places that don't require a chip. We've known since the 80s that you can copy a magnetic strip with a 2-tape boombox (seriously, it will work).
TLDR: There's nothing wrong with the chip cards themselves. But there is something wrong with merchants that haven't upgraded to EMV, and definitely something wrong with /. editors that write a completely ass-backwards headline.
Wiping data sounds good, but what does that have to do with removable storage? You can boot from USB (or to the recovery partition) and do an erase right there.
I imagine yes, but that has nothing to do with the old style of punching holes through magnetic media.
I've worked with more than one company now whose architecture is such that even if a band of thieves broke into the DC and stole the entire rack, they'd have nothing of value on the disks. All the data were encrypted at rest and the keys held only in RAM. As such, they reviewed and removed the 'erase drives' provision entirely. Decommissioning a machine just means ensuring it powercycles.
[ Now, in theory, if they could steal them and keep them powered, then break in, sure. Or if they could pwn the machines before stealing them, they could have it write the keys persistently -- but in that case, why bother physically stealing them at all? ]
In the Northern US, there's a bit more than 9 hours of daylight at the winter solstice. Centering it around 1PM would make it light from 8:30AM to 5:30PM, which would make morning commutes a slog.
7:30AM to 4:30PM seems better in that regards.
Yeah, and change when people normally show up to work. And move lunch too. And store openings and TV schedules too.
At some point modifying our arbitrary mapping of the passage of time into arabic numerals is easier than changing everything else that we pick arbitrary numbers for.
Just checked with a Kill A Watt, the power draw when there are no devices handshaked is 1W.
Maybe yours were from before the handshake?
I'm a consumer in love with it, and I understand it's never going to be as fast as wired. I don't use it exclusively, but it's really convenient in cases where I just don't need that kind of speed.
For example, in the car, I have both a 2A wired charger and a 500mA inductive. If my battery is in reasonable shape, I just pop the phone in the inductive charger/holder thingy and off I go. It's easily enough to run Waze and stream bluetooth music without draining the battery at all, and if I really need 2A (or it's going to be a long trip) I can always just plug it in.
Second example, overnight, I have an inductive charger in a little divot on the nightstand. Even from ~0%, 500mA over a night's sleep is enough to wake up at 100%.
Finally, your claim about the power bill going up is absolutely absurd because devices are basically zero fraction of anyone's electricity bill. The average consumption is 900 KWH/month. A phone is ~3000mAH. Even if you charged a phone once a day on an inductive charger at 10% efficiency, that would be (rounding up) ~1KWH = (3000mAH * 10x loss * 30 days/month) or ~0.1% of your total bill. You'd be paying like $0.20/month to charge the phone, of which 18Â was wasted.
Modern inductive chargers don't have excellent efficiency by the way, but they're better than 10% :-)
First of all, singular they has a long history within the English language dating back at least as far the Bishops (1568) and King James (1611) Bibles, Shakespeare, Emily Dickinson, Jane Austen and many other notable authors.
Or, as Language Log put it:
By all means, avoid using they with singular antecedents in your own writing and speaking if you feel you cannot bear it. Language Log is not here to tell you how to write or speak. But don't try to tell us that it's grammatically incorrect. Because when a construction is clearly present several times in Shakespeare's rightly admired plays and poems, and occurs in the carefully prepared published work of just about all major writers down the centuries, and is systematically present in the unreflecting conversational usage of just about everyone including Sean Lennon, then the claim that it is ungrammatical begins to look utterly unsustainable to us here at Language Log Plaza. This use of they isn't ungrammatical, it isn't a mistake, it's a feature of ordinary English syntax that for some reason attracts the ire of particularly puristic pusillanimous pontificators, and we don't buy what they're selling.
Second, the sneering and incorrect hyper-grammar-policing of a historically acceptable construction is bad enough, but did you really have to do it in a post mistaking "there" and "their" in the second sentence? Because that's not some marginal or debatable rule of grammar, that's actually two different words with totally different meanings. Even Safari's god-awful grammar checker flags that one as questionable . . .
It's too late. If they had done this before BTRFS became production-worthy, it would have taken the air out of BTRFS. Now it's got momentum.
I don't have many hard and fast office rules, but here's a soft one -- if your keyboard is loud enough to bother others, you will politely be asked to select a quieter, which will gladly be provided free of charge.
By all mean, please do enjoy your unholy racket with its excellent tactile feel and whatever, wherever you like except not here.
User's preferences can never be the problem. This is a contradiction in terms. Things are measured by what people think is better. The sooner you lose this attitude that there's any other measure, the sooner you can build things that people actually want to use.
This is not to say that there isn't a dialogue between the person building widgets and the users consuming them. A lot of building and iterating involves not just blindly giving people what they want, but in exploring what their underlying need is and trying to fulfill it. It's not just 'a faster horse'.
But still, in the final analysis, the builder is successful if and only if the user thinks the widget is good. That's the metric.
Fun story time related by a colleague. A pretty common piece of software (hint: there's probably one running within a few hundred yards of you) had an elusive bug. But as the parent noted, printf caused the problem to go away, and it was suspected because it caused synchronization on stdout. Unlike the parent, the developers didn't have time to actually implement a buffered-log solution to figure this out, so they the obviously-logical thing -- they replaced all the printf calls with barrier() and shipped it. It's still running like this today.
Another good one, I worked with someone who would log everything all the time by fprintfing to a high-numbered pipe. When I asked him, he gave a few advantages that still ring partially true (depends on context): first, he said, I can get the log from any running instance without even stopping by d-tracing the system call. But most critically, he said, all the formatting happens in userland and only after the syscall does the kernel actually realize that there's nothing on the other end of the pipe and drop the write. That means, he reasoned, that the release/debug versions would always have very close behavior and would avoid the class of 'bugs that don't reproduce in debug build'. As with the other story, to this day, there's a slew of machines out there, formatting and writing log messages to a pipe that's never open.
But in practice the inability to apply security patches to a work such as Linux (which presumably creates a 'new version') is a death sentence.
Forgetting for a moment the logistics of actually enforcing such a ban, what would such a ban actually entail?
Presumably the ban would not apply to the mere act of developing a drone/robot or the AI to make it move around and perceive the environment. That is, it's not a ban on any of the fundamental building blocks of robotics or AI. Nor would it apply to actually building a search-and-resce bot or teaching it to navigate unknown or difficult terrain. So countries interested in such things would continue to build them just short of actually being weapons (with laughable charades)
After that, what's left? Attaching a gun to it? IFF? It seems fairly clear that this is not the technical challenge here. So even if there's a 'ban', countries will be free to develop technologies and, if they master those problems, will credibly be a screwdriver away" from having full on AI weapons.
Tax evasion means that, for a given level of government revenue, nominal rates must be higher.
Higher nominal rates means, for a given level of risk of getting caught, tax evasion is more rewarding.
It's a circular problem, one in which unfortunately Greece is stuck in a bad attractor. There is also a virtuous version of the cycle, where effectively cracking down on evasion allows lower nominal rates, which in turn lowers the incentive to evade. It's debatable whether it's actually possible to drag Greece out of the bad attractor and into the good one, especially given that some people believe that it's become embedded in Greek culture.
And yes, no one is claiming that Greece's government has been a paragon of efficiency and integrity, or that its spending choices have been ideal. But that's kind of apart from how revenue is generated. You don't have to agree with any of that (either entirely or on the margin) to see that the tax collection problem is really hurting them.
Boy am I glad I don't work for any of those places.
I'm not saying I agree with TFA's conclusion, but you can easily explain the existence of marathon runners as being a completely insignificant fraction of the population.
It's not reasonable to say that researchers cannot claim things about the majority of humans just because you can dredge up a counter-example. It's implied they are saying that this is true on average and not for literally 100% of people.
No one is disagreeing with that or saying there shouldn't be an API that allows you to knowingly chose to share all your email data with whoever you want. But just like the grocery store can't sell a bag chips without disclosing how many grams of fat are in it, so too should the person offering the API disclose clearly what data are shared, whether they are shared-onwards to third parties, whether they are persisted and how to remove yourself.
Maybe another way to put it is that you can't decide on what level of privacy you are sacrificing if the service doesn't clearly explain what you are singing up for, or worse, claims one thing but then does another. Uninformed consent is not valid consent.
If Google is enforcing those disclosure, opt-out and accuracy requirements, great. No problem. But they don't necessarily have an interest to do so too thoroughly, not least because it requires lots of human intervention, which isn't their strong suit.