The law isn't telling people they can't do it, rather it is saying that the rest of society has the right to not be exposed to it involuntarily (as is also the case with regular tobacco smoke). You can still smoke it in your private home, or in your private car, or in other private places.
Maybe we need a "vaping mask" that can be used in public. It would seal around mouth and nose, and filter exhalations. That would address the problem.
Google turns up a few people who have modified gas masks for this purpose, but I think something much more compact could be devised.
I certainly have to grant that the small battery and lack of battery cooling are flaws, though the latter doesn't affect me and the former doesn't affect me much. I'm in line for a Model 3, but fairly far down the list. Probably late 2018 or early 2019. Maybe later if they keep having production ramp-up problems.
People might not realize this, but repair shops will be there, doesn't matter if these laws pass or not.
Not necessarily.
Without the laws, we may move to a world where every component in the device has embedded authentication keys, and all must mutually authenticate to some central component -- or even to a server somewhere -- before they're willing to function. In your washing machine example, suppose the replacement part recognized that it wasn't installed in the machine it's supposed to be installed in and just refused to operate? Only an authorized repair center would be able to rekey the components to make them work together.
That's just one approach, a purely technical one. There are other technical approaches, and other approaches that apply a hybrid of legal and technical constraints to restrict who can make repairs. This is why the laws are needed, because without the laws it's entirely possible that we'll lose the ability to repair devices.
The typical curve is about 4% in the first year, then it greatly slows with time; year five total degradation is about 6-7%.
Assuming reasonable headroom in the capacity and no overheating. Nissan LEAF owners in hot climates have experienced serious battery degradation due to the lack of a battery cooling system. I own a LEAF and have had no problems (50K miles on the clock and negligible degradation), but I don't live in a hot climate.
As horrific as corporate actions have been throughout history, none have even come close to the heinous shit that governments have been able to do when given too much power.
You might want to read up a bit. Start by searching for something called "East India Company." From there you might try "United Fruit Company." There are other examples.
And then you can look up Hitler's Germany, Stalin's Soviet Union, Mao's China, Chiang Kai Shek's China, Leopold II's Congo Free State, the Empire of Japan, Pol Pot's Cambodia, Young Turk's Ottomon Empire... I'll stop now, but I could go on. The total butcher's bill of the list I gave is somewhere north of 100 million people murdered.
I'm not particularly anti-government, but nothing in human history remotely compares to the death toll of government gone bad -- even when you only look at how many of their own people they killed, ignoring war! Religion can't hold a candle to political ideology for murderousness, and corporations aren't even in the same zip code.
Any concentration of power should be watched, and most definitely includes corporations, but government needs the most scrutiny of all because governments are both far larger and wealthier than any corporation and especially because governments are specifically empowered and authorized to use deadly force. Some corporations have gotten away with nasty shit, including pre-meditated murder, but nothing at all like what governments have done.
Explaining to a user why they're wrong for disliking a user interface is a major problem with UI designers. Maybe spend more time listening and you might be able to design a better interface.
I'm not a UI designer, never claimed to be. In fact if my 30-year career as a software developer has a consistent thread running through it, it's staying as far as possible from UI.
But if you actually read my comment, you'll see that much of it was pointing out that his description of how the UI works is factually incorrect. Makes me wonder if he actually uses Gmail.
Delete is a trashcan icon when reading, but different when viewing the inbox.
No, it's a trash can in both cases. When viewing the inbox it's not visible at all unless you select one or more emails.
"Select all" means "select the first 50"
Actually it means select all 50 that are showing.
have to do that numerous times to actually select/delete *all* messages
Well, if "numerous times" means "one more click".
You click "select all" to select all that are visible (50), then you get a little message that says "All 50 conversations on this page are selected", and a link to click to select all conversations, period. So, two clicks to select all if you have more than 50. The alternative is to have "select all" select a bunch of stuff that isn't being displayed, which could well confuse people.
"reply" is down at the bottom of "conversation" mode so that you have to scroll down many pages to reach
But all of the previous messages in the conversation are automatically hidden, so you normally only have to scroll past the part that you should read before replying. Assuming you read your email before replying to it.
Sigh. I typed a screed about the evils of top-posting and the value of proper trimming. But nobody other than oldsters like me even know what any of that means, so I should just give up.
"cc" and "bcc" are hidden *until* you click in the "to" line...
Most people consider that a feature, since it avoids wasting screen space on unneeded fields in the common case. And it's pretty unsurprising that when you want to edit who you're sending the message to, you click in the "to" area, and then the variations in how you're going to include them pop up.
To get a feel for what I mean: deleting an item from the inbox is a very common action, so why is it hidden (until I mark a checkbox)
Well, until Gmail incorporates eye tracking, how is it supposed to know which of the emails in your inbox you want to delete? If you click on one of them, it opens and you can delete it. If you want to delete one without opening it, you mark it (normally this is expected to be used only for bulk actions... where you mark several and then delete them at once).
why does (this really common feature) take several steps?
It takes the minimum possible number of steps: two. This is the minimum possible because you have to tell Gmail which email you want to do something to, and then you have to tell it what to do. The only way to turn that into a single step would be to include a "delete" icon next to every email which would consume a huge amount of screen real-estate to no purpose. The extra click is a better solution.
For all the bad things we say about Apple, at least they know how to make a good interface.
I can think of plenty of bad things to say about Google UIs, and about Gmail in particular, but none of yours make any sense.
BTW, in the spirit of pointing out undiscoverable features, I highly recommend that you type a question mark next time you're looking at your Gmail inbox. A little effort invested in learning the keyboard shortcuts pays huge dividends.
Back in the 80's I worked on a system to analyze customer requirements to verify that there were no conflicting requirements, both internally or legally. This was because the requirements took up a small room full of documents. As part of the analysis process, this program also spit out compile-able and executable code.
Not that revolutionary.
Okay, so use that approach to build a self-driving car.
NOTHING except food and water has any real meaning or value
And clothing. And shelter. And tools used to gather food, make clothing or build shelter. And tools used to make tools. And objects used for entertainment. And...
What has real meaning or value is whatever people want or need. Some things are essential to survival, of course, so those things will have nearly unlimited worth in the eyes of those who lack them... but once people have enough food, more food has little value (though different, tastier food has entertainment value).
What is also true is that the NSA insisted on shortening the key length from 64-bits to 56-bits for no good reason other than weakening the standard to brute force attacks that they alone could afford at the time.
1. No it's not. We're not in the realm of actually verifying that the devices are powered down / disconnected on a modern circuit board. You'll still relying on software to tell you that it's been disabled.
Most of the interesting peripherals in a cellphone are connected via very small ribbon cables, not surface mount. It's easy to wire the hardware switch so it interrupts the power line in one of the cable terminations. Also, given sufficient control over the software on the device, it's also possible to do a verification in software that the hardware switch really works and does what it's supposed to do, and that verification would remain valid even when running the much more complex and difficult to verify production software.
A hardware switch's presence and function is easier to verify.
You have to trust that the vendor doesn't have some way to bypass it, though. That requires a more advanced teardown, and someone has to pay for that, or do the work themselves.
It can be designed so that it's easy to verify, as I pointed out in the post you replied to. It helps if the component has a discrete power line.
Why would you trust the hardware switch anymore than the software switch? In both cases you have to trust the vendor.
Three reasons:
1. A hardware switch's presence and function is easier to verify. A few teardowns of randomly-selected units can verify it to an arbitrarily high degree of confidence. Even better if the device is built so that you can verify the wiring even without destructive teardown, so every sufficiently-paranoid user can check their own device, personally.
2. Even if you verify the presence and functionality of a software switch by carefully analyzing the software, an update can be delivered at any time that disables the switch -- or even reverses it! If I wanted to monitor you, I would be most interested in what you do while the switch is in the off position.
3. Even if the vendor is fully trustworthy and never disables the software switch, malware can still disable it. An attacker who compromises the system can turn it on. To do the same with a hardware switch, the attacker would have to physically steal your phone, modify the internal wiring and return it to you.
There is a lot of value to hardware switches. Note that I don't care that much myself. A software switch from a vendor I trust is good enough for me. But I can understand those who might want more.
Your blog post says security of fingerprint authentication depends on the value of the thing being secured.
It says that as well, but the main thrust of the post is that rotation is useful within the password security model, since passwords require secrecy and rotation restores it, but that rotation does not do anything useful for biometrics, because their security -- to the degree that they have it -- is based on the integrity of the acquisition process, not secrecy.
So statements about fingerprint security being good could be "this crap again" depending on the value of the thing being secured - and inability to change it is definitely not irrelevant.
Sure, the security could be lousy, but the inability to change your fingerprints is irrelevant regardless. They're not secrets, they were never secrets, they never will be secrets, and there's no point in rotating non-secrets.
I'm curious the strategic use of bombers on 24 hour standby, when there are enough ICBMs, including those in nuclear subs which are likely really, really close to North Korea already, to totally decimate that country. North Korea could be a smoldering ruin before the bombers would even leave US airspace (even if they were on standby). So I wonder if the bombers would simply be more "obvious" to Kim Jong or what?
You'll have to narrow it down a bit more, I'm not going to give you a citation for every sentence in my post. If there's something I said that seems particularly unlikely, let me know, and I will either give some evidence to support it, give it a disclaimer, or outright retract it.
There wasn't that much in the post in question. But, here:
But both of them are now process driven companies, primarily focused on not overturning the boat, and the result is code that follows process.
And
As long as process is followed, you don't have to worry about whether you did a good job or not. Just go home at the end of the day.
And
That is the mentality of the vast majority of mediocre programmers at both companies.
I didn't say I disagreed (or agreed), I just asked for substantiation of your claims. Are you speaking from personal experience, having worked at Microsoft and Google? Are you relaying information from friends who work there? Do you have some other sort of basis for your claim?
Bingo! Biometrics suck. How do you change your fingerprints, or your eye's iris?
This is irrelevant. Your erroneous argument arises from the mistaken application of the secrecy-based security model of passwords to the acquisition process-based security model of biometrics. I explained at length here: http://divegeekstuff.blogspot....
When you look at this this way, then you see that things like finger prints or retina have the same problems and worse. they are not mutable, they can be taken from you without you knowing it, and the transmission layer is still vulnerable
Mutability doesn't matter for biometrics, and neither does the fact that copies can be taken without your knowledge, because those don't affect the biometric security model. I wrote a detailed analysis here: http://divegeekstuff.blogspot....
Good has some really good programmers, and so does Microsoft. In the past they were even more impressive.
But both of them are now process driven companies, primarily focused on not overturning the boat, and the result is code that follows process. As long as process is followed, you don't have to worry about whether you did a good job or not. Just go home at the end of the day. That is the mentality of the vast majority of mediocre programmers at both companies.
I personally think the entire concept of bounties and crowd sourcing your QA is utter stupidity and pretty frickin lazy and irresponsible.
I think perhaps you missed the part where Google is offering bounties for vulnerabilities in other companies' apps. Google's QA has no responsibility for these apps, so your argument is off target. Also, your terminology is a little off: QA is usually the organization responsible for functional testing and validation. Vulnerability prevention and discovery usually falls to a dedicated security team. QA and security skills are quite different.
That said, Google absolutely does offer bounties for bugs in its own software, up to $200,000 in the case of Android. Actually, I think the true maximum is a little higher than that, since reporters can get a little more if they provide patches and tests. In addition, Android does have a good-sized security team, including an organization focused on finding vulnerabilities.
Why both? Why do bounties and have an internal team? Because neither approach alone is as effective as both together. Any organization that really cares about the security of its products must do both -- and more; there's a third approach that is also needed. Let me explain why all three are important.
First, the internal team matters because they bring something to the table that no outsider can: long-term focus. Especially with respect to large systems, it's very difficult for someone who researches many different products to develop really deep knowledge of any one of them. Essentially, the internal team provides breadth of focus across all security aspects of the product.
Second, the external vulnerability researchers are important because they provide breadth of focus on attack techniques. Many of the external researchers are academics. Their focus is on devising some clever new way to break systems, or some especially effective way to automate old ways of breaking systems, and so their goal is to apply their technique to a wide variety of products. Bug bounties ensure that they turn their techniques on your product, and that they take the next step to do the work necessary to really prove that the vulnerability they found can be attacked, so you don't waste a lot of time trying to fix theoretical issues.
There's no way to hire all of the world's security researchers, and even if you could, it wouldn't make sense. These guys focus on new techniques, so while you want them to put a little effort into your product, you don't want to pay them full time.
The third group of people you want attacking your product is contract penetration testers. You can (and should!) have your internal team doing penetration testing, but they risk developing tunnel vision. Bringing in outside experts provides an infusion of fresh ideas (like the academic researchers) and the fact that you're paying a nice contract fee provides focus. Thus, they provide a blend of the benefits of external and internal research.
The combination of these three things is dramatically more effective than any one of them.
When discussing a vacuum container at sea level the pressure on the container will be one atmosphere which is 101kPa or 15psi and not equivalent to 10 meters of water.
The pressure differential between a container of perfect vacuum and ambient atmosphere at sea level is about 14.7 psi. The pressure differential between a container of air at one atmosphere and ambient seawater at 10m depth is about 14.7 psi. Each 10m of seawater depth increases ambient pressure by approximately one atmosphere. At 100m, pressure is 10 atmospheres higher than the surface. Freshwater is less dense than seawater, so pressure increases a bit more slowly.
Any SCUBA diver knows this.
Here's a calculator if you'd like to play with it.
Slashdot Mirror
The law isn't telling people they can't do it, rather it is saying that the rest of society has the right to not be exposed to it involuntarily (as is also the case with regular tobacco smoke). You can still smoke it in your private home, or in your private car, or in other private places.
Maybe we need a "vaping mask" that can be used in public. It would seal around mouth and nose, and filter exhalations. That would address the problem.
Google turns up a few people who have modified gas masks for this purpose, but I think something much more compact could be devised.
Yes, I'm assuming a proper EV (aka, not a Leaf) ;)
Hey there! I quite like my LEAF!
I certainly have to grant that the small battery and lack of battery cooling are flaws, though the latter doesn't affect me and the former doesn't affect me much. I'm in line for a Model 3, but fairly far down the list. Probably late 2018 or early 2019. Maybe later if they keep having production ramp-up problems.
People might not realize this, but repair shops will be there, doesn't matter if these laws pass or not.
Not necessarily.
Without the laws, we may move to a world where every component in the device has embedded authentication keys, and all must mutually authenticate to some central component -- or even to a server somewhere -- before they're willing to function. In your washing machine example, suppose the replacement part recognized that it wasn't installed in the machine it's supposed to be installed in and just refused to operate? Only an authorized repair center would be able to rekey the components to make them work together.
That's just one approach, a purely technical one. There are other technical approaches, and other approaches that apply a hybrid of legal and technical constraints to restrict who can make repairs. This is why the laws are needed, because without the laws it's entirely possible that we'll lose the ability to repair devices.
The typical curve is about 4% in the first year, then it greatly slows with time; year five total degradation is about 6-7%.
Assuming reasonable headroom in the capacity and no overheating. Nissan LEAF owners in hot climates have experienced serious battery degradation due to the lack of a battery cooling system. I own a LEAF and have had no problems (50K miles on the clock and negligible degradation), but I don't live in a hot climate.
As horrific as corporate actions have been throughout history, none have even come close to the heinous shit that governments have been able to do when given too much power.
You might want to read up a bit. Start by searching for something called "East India Company." From there you might try "United Fruit Company." There are other examples.
And then you can look up Hitler's Germany, Stalin's Soviet Union, Mao's China, Chiang Kai Shek's China, Leopold II's Congo Free State, the Empire of Japan, Pol Pot's Cambodia, Young Turk's Ottomon Empire... I'll stop now, but I could go on. The total butcher's bill of the list I gave is somewhere north of 100 million people murdered.
I'm not particularly anti-government, but nothing in human history remotely compares to the death toll of government gone bad -- even when you only look at how many of their own people they killed, ignoring war! Religion can't hold a candle to political ideology for murderousness, and corporations aren't even in the same zip code.
Any concentration of power should be watched, and most definitely includes corporations, but government needs the most scrutiny of all because governments are both far larger and wealthier than any corporation and especially because governments are specifically empowered and authorized to use deadly force. Some corporations have gotten away with nasty shit, including pre-meditated murder, but nothing at all like what governments have done.
Explaining to a user why they're wrong for disliking a user interface is a major problem with UI designers. Maybe spend more time listening and you might be able to design a better interface.
I'm not a UI designer, never claimed to be. In fact if my 30-year career as a software developer has a consistent thread running through it, it's staying as far as possible from UI.
But if you actually read my comment, you'll see that much of it was pointing out that his description of how the UI works is factually incorrect. Makes me wonder if he actually uses Gmail.
Delete is a trashcan icon when reading, but different when viewing the inbox.
No, it's a trash can in both cases. When viewing the inbox it's not visible at all unless you select one or more emails.
"Select all" means "select the first 50"
Actually it means select all 50 that are showing.
have to do that numerous times to actually select/delete *all* messages
Well, if "numerous times" means "one more click".
You click "select all" to select all that are visible (50), then you get a little message that says "All 50 conversations on this page are selected", and a link to click to select all conversations, period. So, two clicks to select all if you have more than 50. The alternative is to have "select all" select a bunch of stuff that isn't being displayed, which could well confuse people.
"reply" is down at the bottom of "conversation" mode so that you have to scroll down many pages to reach
But all of the previous messages in the conversation are automatically hidden, so you normally only have to scroll past the part that you should read before replying. Assuming you read your email before replying to it.
Sigh. I typed a screed about the evils of top-posting and the value of proper trimming. But nobody other than oldsters like me even know what any of that means, so I should just give up.
"cc" and "bcc" are hidden *until* you click in the "to" line...
Most people consider that a feature, since it avoids wasting screen space on unneeded fields in the common case. And it's pretty unsurprising that when you want to edit who you're sending the message to, you click in the "to" area, and then the variations in how you're going to include them pop up.
To get a feel for what I mean: deleting an item from the inbox is a very common action, so why is it hidden (until I mark a checkbox)
Well, until Gmail incorporates eye tracking, how is it supposed to know which of the emails in your inbox you want to delete? If you click on one of them, it opens and you can delete it. If you want to delete one without opening it, you mark it (normally this is expected to be used only for bulk actions... where you mark several and then delete them at once).
why does (this really common feature) take several steps?
It takes the minimum possible number of steps: two. This is the minimum possible because you have to tell Gmail which email you want to do something to, and then you have to tell it what to do. The only way to turn that into a single step would be to include a "delete" icon next to every email which would consume a huge amount of screen real-estate to no purpose. The extra click is a better solution.
For all the bad things we say about Apple, at least they know how to make a good interface.
I can think of plenty of bad things to say about Google UIs, and about Gmail in particular, but none of yours make any sense.
BTW, in the spirit of pointing out undiscoverable features, I highly recommend that you type a question mark next time you're looking at your Gmail inbox. A little effort invested in learning the keyboard shortcuts pays huge dividends.
This is revolutionary.
Back in the 80's I worked on a system to analyze customer requirements to verify that there were no conflicting requirements, both internally or legally. This was because the requirements took up a small room full of documents. As part of the analysis process, this program also spit out compile-able and executable code.
Not that revolutionary.
Okay, so use that approach to build a self-driving car.
NOTHING except food and water has any real meaning or value
And clothing. And shelter. And tools used to gather food, make clothing or build shelter. And tools used to make tools. And objects used for entertainment. And...
What has real meaning or value is whatever people want or need. Some things are essential to survival, of course, so those things will have nearly unlimited worth in the eyes of those who lack them... but once people have enough food, more food has little value (though different, tastier food has entertainment value).
What is also true is that the NSA insisted on shortening the key length from 64-bits to 56-bits for no good reason other than weakening the standard to brute force attacks that they alone could afford at the time.
The original key size was 128 bits, not 64.
1. No it's not. We're not in the realm of actually verifying that the devices are powered down / disconnected on a modern circuit board. You'll still relying on software to tell you that it's been disabled.
Most of the interesting peripherals in a cellphone are connected via very small ribbon cables, not surface mount. It's easy to wire the hardware switch so it interrupts the power line in one of the cable terminations. Also, given sufficient control over the software on the device, it's also possible to do a verification in software that the hardware switch really works and does what it's supposed to do, and that verification would remain valid even when running the much more complex and difficult to verify production software.
A hardware switch's presence and function is easier to verify.
You have to trust that the vendor doesn't have some way to bypass it, though. That requires a more advanced teardown, and someone has to pay for that, or do the work themselves.
It can be designed so that it's easy to verify, as I pointed out in the post you replied to. It helps if the component has a discrete power line.
Why would you trust the hardware switch anymore than the software switch? In both cases you have to trust the vendor.
Three reasons:
1. A hardware switch's presence and function is easier to verify. A few teardowns of randomly-selected units can verify it to an arbitrarily high degree of confidence. Even better if the device is built so that you can verify the wiring even without destructive teardown, so every sufficiently-paranoid user can check their own device, personally.
2. Even if you verify the presence and functionality of a software switch by carefully analyzing the software, an update can be delivered at any time that disables the switch -- or even reverses it! If I wanted to monitor you, I would be most interested in what you do while the switch is in the off position.
3. Even if the vendor is fully trustworthy and never disables the software switch, malware can still disable it. An attacker who compromises the system can turn it on. To do the same with a hardware switch, the attacker would have to physically steal your phone, modify the internal wiring and return it to you.
There is a lot of value to hardware switches. Note that I don't care that much myself. A software switch from a vendor I trust is good enough for me. But I can understand those who might want more.
Your blog post says security of fingerprint authentication depends on the value of the thing being secured.
It says that as well, but the main thrust of the post is that rotation is useful within the password security model, since passwords require secrecy and rotation restores it, but that rotation does not do anything useful for biometrics, because their security -- to the degree that they have it -- is based on the integrity of the acquisition process, not secrecy.
So statements about fingerprint security being good could be "this crap again" depending on the value of the thing being secured - and inability to change it is definitely not irrelevant.
Sure, the security could be lousy, but the inability to change your fingerprints is irrelevant regardless. They're not secrets, they were never secrets, they never will be secrets, and there's no point in rotating non-secrets.
I'm curious the strategic use of bombers on 24 hour standby, when there are enough ICBMs, including those in nuclear subs which are likely really, really close to North Korea already, to totally decimate that country. North Korea could be a smoldering ruin before the bombers would even leave US airspace (even if they were on standby). So I wonder if the bombers would simply be more "obvious" to Kim Jong or what?
Fail Safe explains it fairly well https://www.youtube.com/watch?...
A sentence or two summary of the 90-minute movie would be useful, at least to give me a vague idea of whether it's worth that much of my time.
So:
1) this barely covers expenses of a site
2) it doesn't even close to cover what ads from a traditional site could.
3) if in-browser mining makes sense, it makes even more sense to mine and show ads.
You'll have to narrow it down a bit more, I'm not going to give you a citation for every sentence in my post. If there's something I said that seems particularly unlikely, let me know, and I will either give some evidence to support it, give it a disclaimer, or outright retract it.
There wasn't that much in the post in question. But, here:
But both of them are now process driven companies, primarily focused on not overturning the boat, and the result is code that follows process.
And
As long as process is followed, you don't have to worry about whether you did a good job or not. Just go home at the end of the day.
And
That is the mentality of the vast majority of mediocre programmers at both companies.
Which part exactly do you disagree with?
I didn't say I disagreed (or agreed), I just asked for substantiation of your claims. Are you speaking from personal experience, having worked at Microsoft and Google? Are you relaying information from friends who work there? Do you have some other sort of basis for your claim?
Bingo! Biometrics suck. How do you change your fingerprints, or your eye's iris?
This is irrelevant. Your erroneous argument arises from the mistaken application of the secrecy-based security model of passwords to the acquisition process-based security model of biometrics. I explained at length here: http://divegeekstuff.blogspot....
You cannot change your fingerprints or other biometric data so when it's compromised
This is irrelevant. I wrote a detailed explanation here: http://divegeekstuff.blogspot....
Do the people proposing this ever have _ANY_ real world experience at all?
I do, about 30 years' worth, in both physical and information security.
When you look at this this way, then you see that things like finger prints or retina have the same problems and worse. they are not mutable, they can be taken from you without you knowing it, and the transmission layer is still vulnerable
Mutability doesn't matter for biometrics, and neither does the fact that copies can be taken without your knowledge, because those don't affect the biometric security model. I wrote a detailed analysis here: http://divegeekstuff.blogspot....
Good has some really good programmers, and so does Microsoft. In the past they were even more impressive. But both of them are now process driven companies, primarily focused on not overturning the boat, and the result is code that follows process. As long as process is followed, you don't have to worry about whether you did a good job or not. Just go home at the end of the day. That is the mentality of the vast majority of mediocre programmers at both companies.
Cite?
I personally think the entire concept of bounties and crowd sourcing your QA is utter stupidity and pretty frickin lazy and irresponsible.
I think perhaps you missed the part where Google is offering bounties for vulnerabilities in other companies' apps. Google's QA has no responsibility for these apps, so your argument is off target. Also, your terminology is a little off: QA is usually the organization responsible for functional testing and validation. Vulnerability prevention and discovery usually falls to a dedicated security team. QA and security skills are quite different.
That said, Google absolutely does offer bounties for bugs in its own software, up to $200,000 in the case of Android. Actually, I think the true maximum is a little higher than that, since reporters can get a little more if they provide patches and tests. In addition, Android does have a good-sized security team, including an organization focused on finding vulnerabilities.
Why both? Why do bounties and have an internal team? Because neither approach alone is as effective as both together. Any organization that really cares about the security of its products must do both -- and more; there's a third approach that is also needed. Let me explain why all three are important.
First, the internal team matters because they bring something to the table that no outsider can: long-term focus. Especially with respect to large systems, it's very difficult for someone who researches many different products to develop really deep knowledge of any one of them. Essentially, the internal team provides breadth of focus across all security aspects of the product.
Second, the external vulnerability researchers are important because they provide breadth of focus on attack techniques. Many of the external researchers are academics. Their focus is on devising some clever new way to break systems, or some especially effective way to automate old ways of breaking systems, and so their goal is to apply their technique to a wide variety of products. Bug bounties ensure that they turn their techniques on your product, and that they take the next step to do the work necessary to really prove that the vulnerability they found can be attacked, so you don't waste a lot of time trying to fix theoretical issues.
There's no way to hire all of the world's security researchers, and even if you could, it wouldn't make sense. These guys focus on new techniques, so while you want them to put a little effort into your product, you don't want to pay them full time.
The third group of people you want attacking your product is contract penetration testers. You can (and should!) have your internal team doing penetration testing, but they risk developing tunnel vision. Bringing in outside experts provides an infusion of fresh ideas (like the academic researchers) and the fact that you're paying a nice contract fee provides focus. Thus, they provide a blend of the benefits of external and internal research.
The combination of these three things is dramatically more effective than any one of them.
LOL you Elon shills are so funny.
I accept your concession.
When discussing a vacuum container at sea level the pressure on the container will be one atmosphere which is 101kPa or 15psi and not equivalent to 10 meters of water.
The pressure differential between a container of perfect vacuum and ambient atmosphere at sea level is about 14.7 psi. The pressure differential between a container of air at one atmosphere and ambient seawater at 10m depth is about 14.7 psi. Each 10m of seawater depth increases ambient pressure by approximately one atmosphere. At 100m, pressure is 10 atmospheres higher than the surface. Freshwater is less dense than seawater, so pressure increases a bit more slowly.
Any SCUBA diver knows this.
Here's a calculator if you'd like to play with it.