As I understood the plan originally, the code that they give the delivery person to open the door is a one-time code. So, if the would-be thief has no way to get in again, how is this a total failure? I'd also bet that both the usage time of the code and whether the door was left locked are both sent back to Amazon. They obviously have communication with the lock if they can set a one-time code.
Someone comes along, looks at it, decides it's all garbage, decides that they can do better, and rewrites it. The only real difference is that those using it in their own product have the option of locally sticking with and maintaining the old code.
The flip side is that there is a huge profit motive here to actually pursue some of the many ways that energy usage could be reduced. This could create the incentive for someone to take a new approach that could then save energy throughout the computing industry which consumes orders of magnitude more energy than bitcoin mining alone. Many breakthroughs and milder advances spinoff from people madly pursuing wealth. How much of the modern internet would we have without the energy wasted on distributing porn?
In the late 80s - early 90s time frame, Sam's and Pace (owned by KMart) were in competition. There were over a dozen Pace Membership Warehouses in the St. Louis metropolitan area by my recollection. I personally preferred Pace and thus watched what happened with interest.
Walmart decided they wanted the business. They proceeded to build a Sam's within sight of almost every Pace at great expense because they had to get whatever land was there instead of cherry-picking sites. It was so blatant that you knew what was going on from day one. After doing so, they opened the stores, set the prices below Pace (running them all at a loss), and fairly quickly put the Pace stores out of business. They then built some more new stores in the area with a different distribution so that they could cover the area with fewer stores and closed down almost everything they had built to put Pace out of business. When Pace started talking lawsuit, Walmart purchased the corpses to shut them up.
In a remarkably short time, we went from a competitive market to a monopoly market.
I sincerely hope that we're not seeing similar tactics happen here, but now to Walmart. Having Amazon in competition with Walmart helps us. Losing Walmart in that competition would put us right back in the monopoly situation with an even stronger predator.
I can't imagine why the count of reviewers would be spoken about in the same article as the fakes. The reviewers are obviously not there to address the fakes problem. There would have to be a lot more of them than 20,000. Perhaps that's just the number of people tweaking the AI system.
I don't see a link to the law in the article so I have to assume its language is correct. "can" isn't "will". I doubt we'll ever see a company hit with a 4% fine.
Also, since NY settled, I'd guess they didn't take as much as they might have either. I see no indication in the article as to what they could have gotten.
Yep. They likely don't have to start from pixelated photos. Most people have high-quality photos out there somewhere. If not, a determined attacker can take his own as you walk down the street.
Also, you don't need the mind pool of Google or Apple to do this. Check out this publicly available demo of 3D Facial Recognition from a Single Image. You can submit your own photo or someone else's and view a 3D model created from it. The model certainly isn't perfect, but it isn't too bad either.
Anyone who uses Face ID on a device with data that they are concerned about protecting is making a mistake. The same is true of fingerprint scanners or virtually any other biometric identification.
For a very interesting look at all of the types of data being collected today, take a peek at the National Information Exchange Model. or the NIEM on github. The easiest way to look at the data is to download the models and open the niem-????.xlsx spreadsheet (name changes with version). The last time I checked, the rules for adding a schema to this model included a strong requirement that it be in use by two agencies before being eligible because its purpose is "exchange". So, it can be assumed that everything here is in use today.
Spend some time looking and you should find models for storing biometric data ranging from the expected fingerprints, DNA, facial images, scar locations, etc. to other things you may have never thought of such as your gait, lip prints, your lip movement during speech, and your body odor composition. The jxdm models are as or more interesting as the biometrics models and include a lot of biometric model augmentations.
Note that for some of these items such as gait and body odor, you'll need to look back at the 2.1 version of the standard. I don't know when, but at some point I'm guessing they realized this data revealed too much of what they were doing and they pulled some models. The j:PersonAugmentationType entry on the jxdm page was particularly interesting in 2.1.
Facial data flew the coop long ago. There is software available today that can create a 3D facial reconstruction from a single image using a neural network. It's not super accurate, but other software can do it much better with many images. Most people have many images in public whether they know it or not. If you go downtown, how many cameras capture your image? Some cities are now estimated to have an average of three angles on you at any moment.
We shouldn't be concerned about the use of this data by software to deliver us fun, and, down the road a bit, serious features. Unless you want to wear a mask everywhere you go, you can't stop it. Your face, and anything else that can be observed while you're walking down the street, is public data.
We should be concerned about any use of biometric data of any type for authentication on any system that we consider critical or valuable.
True, though most have at some time at least asked for suggestions from friends. I can't remember everything my spouse has said they like in the past year. But an assistant could do so flawlessly and analyze emotional context to give me a summary with predictions of satisfaction. I'd be negligent not to ask it for a ranked summary and take it into consideration.
Would you consult your spouse's sibling, best-friend, parent, your kids?
I'd temper this a bit to say thinking about, consulting, and choosing is important.
I expect assistants to be more like cyber extensions of us, not like a friend. It will know more about me than any person can know - likely even more than I know myself.
We already have much help in our lives with procuring a present. We aren't making the gifts ourselves in most instances. Even if we are, it would be very rare indeed that we aren't using things made by others in doing that making. We aren't even usually buying direct from others who did the making. We are instead benefitting from distribution networks and stores that do the advertising to get the suggestions in our heads.
We always think of these progressions as something that changes everything fundamentally when in fact they are just another step in a process that has been going on since the dawn of man. It's relative, not absolute.
The human brain is a lot of separate little neural nets only a fraction of which get used at a time depending on what is necessary. Granted, the way the neural networks being used for object recognition now work is poor. They need to break up the steps into many little ones like the brain does instead of training one big network to go all the way from pixels to objects. But, AI will appear someday, possibly even unexpectedly, when we start putting thousands of these little neural networks together with other layers of networks acting as crowd supervisors. The tech Apple is using to perform the object recognition is a component of AI.
And, yes, right now it is sort of a phone. Personally, making calls or sending messages is less than 5% of my phone use. And, no, I don't play any games on my phone.
Eventually, I expect smartphones to drop the displays, keep other sensors and a small speaker, and change to using AR glasses or any other display around as the display output as well as automatically using the most appropriate devices to talk to you at any given time. They will stay in your pocket or we'll come up with other more secure ways to carry them since they will never have to be pulled out for use.
Basically, they are going to evolve somewhat full circle to being an even more personal PC. They will be able to charge or receive power without leaving our person. And they will be able to utilize nearby processing resources to do heavier processing when necessary (and trusted). With a vastly lower latency, usually less than a millisecond, your "PC" will utilize processing power in your home more like distributed processing than cloud computing. When you come home, it will join more directly into a distributed intelligence that is your true assistant.
I personally want to see the long-predicted age of AI assistants arrive. I want them to evolve to be as good as the best executive assistants out there. If I ask my assistant to buy a gift for my wife, I expect it to be as good as the best executive assistant would be at picking just the right gift for my wife. In order for that to happen, it needs to know a lot of personal stuff about me (to know what kinds of things I might actually buy someone) and about my wife. Without information, an assistant is worthless to me. We've entrusted human assistants with this kind of information, and, frankly, it is probably easier to hack a human assistant then it is to hack iPhone's current protections.
Anyway, assistants need data, and I want assistants.
If a picture would be compromising or something that you wouldn't let your executive assistant see, don't take it.
Why? If there are 50 standards, it would be reasonable for Verizon to set a single policy that simply meets or beats every single one of them. The only increase in the costs that is necessary is the analysis to create the virtual super-standard that meets them all.
It is their choice to decide whether it is worth it to them to go to the extra expense of splitting their system to take advantage of the lower standards in some states to make more profits off of their user's data. I would hope they'd take a higher road but certainly don't expect them to.
Google does not do this. Google does not use end-to-end encryption in storing your data. This allows them to analyze your photos in the cloud. In return for opening your entire photo world up to them (and anyone who can pry it out of them legally or illegally), they do give you some nice features. I enjoy the assistant-generated collages, animations, and movies.
Apple, on the other hand, stores photos using end-to-end encryption. Their cloud servers cannot analyze them. They are performing this analysis on the device. And the results of the analysis are kept under the same security as the photos.
Google may start doing photo analysis on the Pixel 2 with its new AI chip soon, but they are not now. When they do, it is still unlikely that they will give up on being able to access your photos while they are on their server.
The article points to Apple's support site that states clearly "When you search your photos, all of the face recognition and scene and object detection are done completely on your device."
It is great that the AI is finally moving off the cloud and onto the device. This is where it belongs. We should be highly praising any software that implements personal assistance features like this locally. They have the potential of eventually supporting personal clouds made up of nothing but my devices with cradle to grave encryption.
Apple's provision of end-to-end encryption of user data stored in their cloud forces this architecture. They don't have the ability to decrypt your photos stored in iCloud and analyze them on a cloud server. So they have to do it on the device. Kudos to Apple for taking the high road instead of the easy road.
We should have no concerns whatsoever about what it is capable of recognizing. Ultimately, I want it to be capable of accurately recognizing anything in the pictures. Why not? Both the analysis and the photos are stored with the same security. If an attacker can get the analysis, then they can just skip that and get the photos instead.
The important thing is secure storage of both the picture and any analysis of it and that neither ever leaves my device pool at any point in time without explicit intentional action on my part such as "sharing" a photo. Apple is almost all of the way there. I'd like to have the further option of pulling the encrypted cloud storage out of the picture and just have my devices automatically synchronize to each other.
No good social site has as its purpose taking advantage of its users. The net has been a social media platform since before it was the web. Both newsgroups and BBSs were more effectively social and enjoyable than most social media platforms today. I would call most "social media" sites today anti-social because they've taken the substance out of the social exchange / discourse.
Given the choice of whether Slashdot is more news aggregator or more social media, I have to go with more social media. News aggregators don't typically have such an extremely low amount of aggregated news. Slashdot does because it attempts to post only the news worthy of discussion.
On a tech news aggregation site such as phys.org, arstechnica, cnet, etc. I rarely read the comments. They aren't the point.
The comments / discussions aren't an afterthought here, they are the point.
Both Amazon Alexa and (soon) Google Home support video-capable devices and have opened up for the creation of third-party devices. The speakers aren't great in these devices, but they'd be a very cost-effective means of expanding Alexa and Google Home to other rooms.
Of course, you won't get this from Samsung because it would compete with their own offerings.
then I'm most definitely a bot. Even the 66.6% beats me. Perhaps they mean humans reach 87% after retries. I've only had it stop giving me retries a couple of times.
FaceID uses an infrared blaster that projects a point cloud of over 30,000 infrared dots onto the surface (often a face) being scanned. If, for example, the infrared dots weren't of the expected intensity, sharpness, or spacing precision, I'd imagine performance of the 3D scan would be degraded. Presumably, SW might be able to correct for spacing precision, but not sharpness or intensity issues. This technology is essentially the same as the MS Kinect v1 device and was acquired by Apple when they purchased PrimeSense in 2013.
The HW behind FaceID, essentially an evolution of the original MS Kinect technology that Apple purchased a few years back, is used for more than just FaceID. Any app intending to use the AR capabilities or simply to scan 3D surfaces would be degraded by any reduction in performance of this feature. Also, the new portrait capabilities of the camera would be reduced. So, deviations in quality of the sensor would certainly have an effect on the end user experience beyond FaceID itself.
It's not as common in the last couple of years because so many prices are now available online, but I've talked to people going through the aisles of stores checking prices and availability of items that were working for a competitor. I've also seen jobs posted for doing it.
In a competitive leasing business like this that doesn't rely on some patent or otherwise protected product, you can either compete heads up with all of your "secrets" plainly written in your fliers, likely because you've made better deals on properties and support infrastructure, or you're doomed.
Well, over 130,000 bitcoins were traded yesterday. So, yes, it's near certain that you could get rid of 1,000 in a day without significantly blipping the price. You'd probably want to do it in lots of 100 or so on the biggest exchange if you're paranoid just to be sure.
But, really, who cares? You just buy in with a portion of your portfolio that you're willing to lose, wait till it doubles or triples, take out your principal plus perhaps the profit you might have made with the same investment in the market, and then sit back and occasionally give yourself a gift with the remainder. As long as you make it past that first withdrawal, you'll either be able to gift yourself a new home a few years down the road or will have lost nothing.
Whether you call it a finder's fee or a payment for protection, to go to an independent business that already exists and would normally not be beholden to a larger entity, change the rules that they live under, and say we'll be very, very generous and only take xxx percent of the business we let you keep is a classic mob-style shakedown.
It seems like someone needs to be taking a serious look at the racketeering laws.
In order for crypto-mining or any other micro-payment as you go methodology to reduce advertising, there would have to be browser support and something like community policing.
First, Javascript sucks for mining. The browser needs to have a trusted high-quality mining engine optimized to utilize whatever hardware acceleration is available in your system built-in. There should be a standard API for accessing it. Access should be controlled on a per-site basis by the same permissions system that the browser is using to control access to your camera, microphone, location, etc.
Second, permission needs to be given as a contract - in exchange for x percentage of my CPU cycles while your site is in focus, you'll drop your ads. The browser should have an easy means for users to flag violations of the contract. Violations should feed back to a centralized database. The compliance record could be presented during the permissions process as both a color and a rating or something.
Also, whenever mining is occurring, there should be a clear indication in a status area that allows you to see who is benefitting from it.
Note that this system doesn't have to use mining as the barter. Browsers are already incorporating true payment capabilities. This API would be useful for trading just about anything, even perhaps bartering some work on Amazon Turk-like tasks, as micropayments to support websites that you'd like to support but don't want to make traditional payments to.
For example, a website might stream a movie to you for free in exchange for a few minutes of transcription work. There would also need to be a means of securely depositing credit into your "bank". You could go to one site, build up the transcription credit, go back to the movie site, and use that credit to watch a movie. The cryptocurrency miner would just be a thing that can put credit into the system.
Obviously, Google would resist this, though they might accept it if they could scrape a percentage. This is why we still need Firefox and other community-driven platforms.
Slashdot Mirror
Note, I'd never use this, but...
As I understood the plan originally, the code that they give the delivery person to open the door is a one-time code. So, if the would-be thief has no way to get in again, how is this a total failure? I'd also bet that both the usage time of the code and whether the door was left locked are both sent back to Amazon. They obviously have communication with the lock if they can set a one-time code.
Someone comes along, looks at it, decides it's all garbage, decides that they can do better, and rewrites it. The only real difference is that those using it in their own product have the option of locally sticking with and maintaining the old code.
The flip side is that there is a huge profit motive here to actually pursue some of the many ways that energy usage could be reduced. This could create the incentive for someone to take a new approach that could then save energy throughout the computing industry which consumes orders of magnitude more energy than bitcoin mining alone. Many breakthroughs and milder advances spinoff from people madly pursuing wealth. How much of the modern internet would we have without the energy wasted on distributing porn?
In the late 80s - early 90s time frame, Sam's and Pace (owned by KMart) were in competition. There were over a dozen Pace Membership Warehouses in the St. Louis metropolitan area by my recollection. I personally preferred Pace and thus watched what happened with interest.
Walmart decided they wanted the business. They proceeded to build a Sam's within sight of almost every Pace at great expense because they had to get whatever land was there instead of cherry-picking sites. It was so blatant that you knew what was going on from day one. After doing so, they opened the stores, set the prices below Pace (running them all at a loss), and fairly quickly put the Pace stores out of business. They then built some more new stores in the area with a different distribution so that they could cover the area with fewer stores and closed down almost everything they had built to put Pace out of business. When Pace started talking lawsuit, Walmart purchased the corpses to shut them up.
In a remarkably short time, we went from a competitive market to a monopoly market.
I sincerely hope that we're not seeing similar tactics happen here, but now to Walmart. Having Amazon in competition with Walmart helps us. Losing Walmart in that competition would put us right back in the monopoly situation with an even stronger predator.
I can't imagine why the count of reviewers would be spoken about in the same article as the fakes. The reviewers are obviously not there to address the fakes problem. There would have to be a lot more of them than 20,000. Perhaps that's just the number of people tweaking the AI system.
I don't see a link to the law in the article so I have to assume its language is correct. "can" isn't "will". I doubt we'll ever see a company hit with a 4% fine.
Also, since NY settled, I'd guess they didn't take as much as they might have either. I see no indication in the article as to what they could have gotten.
Yep. They likely don't have to start from pixelated photos. Most people have high-quality photos out there somewhere. If not, a determined attacker can take his own as you walk down the street.
Also, you don't need the mind pool of Google or Apple to do this. Check out this publicly available demo of 3D Facial Recognition from a Single Image. You can submit your own photo or someone else's and view a 3D model created from it. The model certainly isn't perfect, but it isn't too bad either.
Anyone who uses Face ID on a device with data that they are concerned about protecting is making a mistake. The same is true of fingerprint scanners or virtually any other biometric identification.
For a very interesting look at all of the types of data being collected today, take a peek at the National Information Exchange Model. or the NIEM on github. The easiest way to look at the data is to download the models and open the niem-????.xlsx spreadsheet (name changes with version). The last time I checked, the rules for adding a schema to this model included a strong requirement that it be in use by two agencies before being eligible because its purpose is "exchange". So, it can be assumed that everything here is in use today.
Spend some time looking and you should find models for storing biometric data ranging from the expected fingerprints, DNA, facial images, scar locations, etc. to other things you may have never thought of such as your gait, lip prints, your lip movement during speech, and your body odor composition. The jxdm models are as or more interesting as the biometrics models and include a lot of biometric model augmentations.
Note that for some of these items such as gait and body odor, you'll need to look back at the 2.1 version of the standard. I don't know when, but at some point I'm guessing they realized this data revealed too much of what they were doing and they pulled some models. The j:PersonAugmentationType entry on the jxdm page was particularly interesting in 2.1.
Facial data flew the coop long ago. There is software available today that can create a 3D facial reconstruction from a single image using a neural network. It's not super accurate, but other software can do it much better with many images. Most people have many images in public whether they know it or not. If you go downtown, how many cameras capture your image? Some cities are now estimated to have an average of three angles on you at any moment.
We shouldn't be concerned about the use of this data by software to deliver us fun, and, down the road a bit, serious features. Unless you want to wear a mask everywhere you go, you can't stop it. Your face, and anything else that can be observed while you're walking down the street, is public data.
We should be concerned about any use of biometric data of any type for authentication on any system that we consider critical or valuable.
True, though most have at some time at least asked for suggestions from friends. I can't remember everything my spouse has said they like in the past year. But an assistant could do so flawlessly and analyze emotional context to give me a summary with predictions of satisfaction. I'd be negligent not to ask it for a ranked summary and take it into consideration.
Would you consult your spouse's sibling, best-friend, parent, your kids?
I'd temper this a bit to say thinking about, consulting, and choosing is important.
I expect assistants to be more like cyber extensions of us, not like a friend. It will know more about me than any person can know - likely even more than I know myself.
We already have much help in our lives with procuring a present. We aren't making the gifts ourselves in most instances. Even if we are, it would be very rare indeed that we aren't using things made by others in doing that making. We aren't even usually buying direct from others who did the making. We are instead benefitting from distribution networks and stores that do the advertising to get the suggestions in our heads.
We always think of these progressions as something that changes everything fundamentally when in fact they are just another step in a process that has been going on since the dawn of man. It's relative, not absolute.
The human brain is a lot of separate little neural nets only a fraction of which get used at a time depending on what is necessary. Granted, the way the neural networks being used for object recognition now work is poor. They need to break up the steps into many little ones like the brain does instead of training one big network to go all the way from pixels to objects. But, AI will appear someday, possibly even unexpectedly, when we start putting thousands of these little neural networks together with other layers of networks acting as crowd supervisors. The tech Apple is using to perform the object recognition is a component of AI.
And, yes, right now it is sort of a phone. Personally, making calls or sending messages is less than 5% of my phone use. And, no, I don't play any games on my phone.
Eventually, I expect smartphones to drop the displays, keep other sensors and a small speaker, and change to using AR glasses or any other display around as the display output as well as automatically using the most appropriate devices to talk to you at any given time. They will stay in your pocket or we'll come up with other more secure ways to carry them since they will never have to be pulled out for use.
Basically, they are going to evolve somewhat full circle to being an even more personal PC. They will be able to charge or receive power without leaving our person. And they will be able to utilize nearby processing resources to do heavier processing when necessary (and trusted). With a vastly lower latency, usually less than a millisecond, your "PC" will utilize processing power in your home more like distributed processing than cloud computing. When you come home, it will join more directly into a distributed intelligence that is your true assistant.
I personally want to see the long-predicted age of AI assistants arrive. I want them to evolve to be as good as the best executive assistants out there. If I ask my assistant to buy a gift for my wife, I expect it to be as good as the best executive assistant would be at picking just the right gift for my wife. In order for that to happen, it needs to know a lot of personal stuff about me (to know what kinds of things I might actually buy someone) and about my wife. Without information, an assistant is worthless to me. We've entrusted human assistants with this kind of information, and, frankly, it is probably easier to hack a human assistant then it is to hack iPhone's current protections.
Anyway, assistants need data, and I want assistants.
If a picture would be compromising or something that you wouldn't let your executive assistant see, don't take it.
Why? If there are 50 standards, it would be reasonable for Verizon to set a single policy that simply meets or beats every single one of them. The only increase in the costs that is necessary is the analysis to create the virtual super-standard that meets them all.
It is their choice to decide whether it is worth it to them to go to the extra expense of splitting their system to take advantage of the lower standards in some states to make more profits off of their user's data. I would hope they'd take a higher road but certainly don't expect them to.
Google does not do this. Google does not use end-to-end encryption in storing your data. This allows them to analyze your photos in the cloud. In return for opening your entire photo world up to them (and anyone who can pry it out of them legally or illegally), they do give you some nice features. I enjoy the assistant-generated collages, animations, and movies.
Apple, on the other hand, stores photos using end-to-end encryption. Their cloud servers cannot analyze them. They are performing this analysis on the device. And the results of the analysis are kept under the same security as the photos.
Google may start doing photo analysis on the Pixel 2 with its new AI chip soon, but they are not now. When they do, it is still unlikely that they will give up on being able to access your photos while they are on their server.
The article points to Apple's support site that states clearly "When you search your photos, all of the face recognition and scene and object detection are done completely on your device."
It is great that the AI is finally moving off the cloud and onto the device. This is where it belongs. We should be highly praising any software that implements personal assistance features like this locally. They have the potential of eventually supporting personal clouds made up of nothing but my devices with cradle to grave encryption.
Apple's provision of end-to-end encryption of user data stored in their cloud forces this architecture. They don't have the ability to decrypt your photos stored in iCloud and analyze them on a cloud server. So they have to do it on the device. Kudos to Apple for taking the high road instead of the easy road.
We should have no concerns whatsoever about what it is capable of recognizing. Ultimately, I want it to be capable of accurately recognizing anything in the pictures. Why not? Both the analysis and the photos are stored with the same security. If an attacker can get the analysis, then they can just skip that and get the photos instead.
The important thing is secure storage of both the picture and any analysis of it and that neither ever leaves my device pool at any point in time without explicit intentional action on my part such as "sharing" a photo. Apple is almost all of the way there. I'd like to have the further option of pulling the encrypted cloud storage out of the picture and just have my devices automatically synchronize to each other.
No good social site has as its purpose taking advantage of its users. The net has been a social media platform since before it was the web. Both newsgroups and BBSs were more effectively social and enjoyable than most social media platforms today. I would call most "social media" sites today anti-social because they've taken the substance out of the social exchange / discourse.
Given the choice of whether Slashdot is more news aggregator or more social media, I have to go with more social media. News aggregators don't typically have such an extremely low amount of aggregated news. Slashdot does because it attempts to post only the news worthy of discussion.
On a tech news aggregation site such as phys.org, arstechnica, cnet, etc. I rarely read the comments. They aren't the point.
The comments / discussions aren't an afterthought here, they are the point.
Both Amazon Alexa and (soon) Google Home support video-capable devices and have opened up for the creation of third-party devices. The speakers aren't great in these devices, but they'd be a very cost-effective means of expanding Alexa and Google Home to other rooms.
Of course, you won't get this from Samsung because it would compete with their own offerings.
In 2003, Walmart announced plans to RFID tag everything in the store and track it to the shelf it was on 24/7. So, I'll believe it when I see it.
then I'm most definitely a bot. Even the 66.6% beats me. Perhaps they mean humans reach 87% after retries. I've only had it stop giving me retries a couple of times.
FaceID uses an infrared blaster that projects a point cloud of over 30,000 infrared dots onto the surface (often a face) being scanned. If, for example, the infrared dots weren't of the expected intensity, sharpness, or spacing precision, I'd imagine performance of the 3D scan would be degraded. Presumably, SW might be able to correct for spacing precision, but not sharpness or intensity issues. This technology is essentially the same as the MS Kinect v1 device and was acquired by Apple when they purchased PrimeSense in 2013.
https://www.theverge.com/circu...
The HW behind FaceID, essentially an evolution of the original MS Kinect technology that Apple purchased a few years back, is used for more than just FaceID. Any app intending to use the AR capabilities or simply to scan 3D surfaces would be degraded by any reduction in performance of this feature. Also, the new portrait capabilities of the camera would be reduced. So, deviations in quality of the sensor would certainly have an effect on the end user experience beyond FaceID itself.
It's not as common in the last couple of years because so many prices are now available online, but I've talked to people going through the aisles of stores checking prices and availability of items that were working for a competitor. I've also seen jobs posted for doing it.
In a competitive leasing business like this that doesn't rely on some patent or otherwise protected product, you can either compete heads up with all of your "secrets" plainly written in your fliers, likely because you've made better deals on properties and support infrastructure, or you're doomed.
Well, over 130,000 bitcoins were traded yesterday. So, yes, it's near certain that you could get rid of 1,000 in a day without significantly blipping the price. You'd probably want to do it in lots of 100 or so on the biggest exchange if you're paranoid just to be sure.
But, really, who cares? You just buy in with a portion of your portfolio that you're willing to lose, wait till it doubles or triples, take out your principal plus perhaps the profit you might have made with the same investment in the market, and then sit back and occasionally give yourself a gift with the remainder. As long as you make it past that first withdrawal, you'll either be able to gift yourself a new home a few years down the road or will have lost nothing.
Whether you call it a finder's fee or a payment for protection, to go to an independent business that already exists and would normally not be beholden to a larger entity, change the rules that they live under, and say we'll be very, very generous and only take xxx percent of the business we let you keep is a classic mob-style shakedown.
It seems like someone needs to be taking a serious look at the racketeering laws.
In order for crypto-mining or any other micro-payment as you go methodology to reduce advertising, there would have to be browser support and something like community policing.
First, Javascript sucks for mining. The browser needs to have a trusted high-quality mining engine optimized to utilize whatever hardware acceleration is available in your system built-in. There should be a standard API for accessing it. Access should be controlled on a per-site basis by the same permissions system that the browser is using to control access to your camera, microphone, location, etc.
Second, permission needs to be given as a contract - in exchange for x percentage of my CPU cycles while your site is in focus, you'll drop your ads. The browser should have an easy means for users to flag violations of the contract. Violations should feed back to a centralized database. The compliance record could be presented during the permissions process as both a color and a rating or something.
Also, whenever mining is occurring, there should be a clear indication in a status area that allows you to see who is benefitting from it.
Note that this system doesn't have to use mining as the barter. Browsers are already incorporating true payment capabilities. This API would be useful for trading just about anything, even perhaps bartering some work on Amazon Turk-like tasks, as micropayments to support websites that you'd like to support but don't want to make traditional payments to.
For example, a website might stream a movie to you for free in exchange for a few minutes of transcription work. There would also need to be a means of securely depositing credit into your "bank". You could go to one site, build up the transcription credit, go back to the movie site, and use that credit to watch a movie. The cryptocurrency miner would just be a thing that can put credit into the system.
Obviously, Google would resist this, though they might accept it if they could scrape a percentage. This is why we still need Firefox and other community-driven platforms.