Is there hard evidence supporting the rumors that there has been a reduction in the accuracy of the Face ID modules? * No. There isn't any hard evidence. Why? Because the iPhone X Face ID cannot be tested by users and labs. Why is that? Because the phone has not yet been released.
* What we are reading in this article (and others about this topic) are unsubstantiated rumors. That's it.
What is interesting is that if it’s true that testing has been loosened, it’s for production that’s coming on line, but not available yet. So the first few million will have the more tightly tested projectors, and the rest will have the more loosely tested projectors.
so if that is true, then an early production phone can be tested against a later production phone. But to be a useful test, rather than just a test of production variations, a number of phones need to be tested. The difference has to be a statistical significance. It needs to be greater than normal production variances.
Sure, once the iPhone X sells in the millions, then labs can start getting statistical data on the accuracy of the sensors. If there is a clear difference in sensor components or sensor settings, then that would be evidence that some change took place in the manufacturing process. (I could then imagine the resulting lawsuits.)
* But the reality is that all of this is speculation based on rumors from business websites. We know nothing for certain about this right now.
Which is why I say “...if that’s true”. It would be a pretty boring discussion if all we said was...”we don’t know if it’s true”
I agree to this extent; beyond stating that "we don't know if it's true", I'm not very interested in speculating about possible Apple failures. But I do understand that a lot of people enjoy discussing some disaster that might happen to Apple. There is a niche in tech journalism devoted just to those possible Apple failure tropes (which is followed on the MacWorld website by the Macalope column).
Just send an iPhone X my way and I'll test the accuracy a million times.
In a random test wouldn't you just have to get past 500k to have a greater than 50% chance?
But it's probably less than that since the closest physical match to your facial features are likely the people closest to you, which is why Face ID needs to be 20x that of Touch ID since fingerprint patterns are more unique between siblings, parents, and offspring, even though finger prints are heavily influenced by genetic factors (like basic size, shape, and spacing of dermatoglyphs).
Its both more and less secure depending on who you think you should protect against. Like you said, a family member is probably one of those one in millions, so a random person's chance of matching all those features are a lot lower. The nice thing is you can actually test to see if your family member are unlocking your phones, unlike with touch ID (just by exposing them to your phone).
In the future, they could just ban those people specifically (at the risk of increasing false negatives for yourself), cause there is undoubtably a margin of error in the mapping and thus they accept matches when it is close enough, says 99.97% of points match you and 99.95% matches your sibling, maybe they would normally consider this acceptable but knowing this variant is not merely an error but a different person it could rule it out.
Is there hard evidence supporting the rumors that there has been a reduction in the accuracy of the Face ID modules? * No. There isn't any hard evidence. Why? Because the iPhone X Face ID cannot be tested by users and labs. Why is that? Because the phone has not yet been released.
* What we are reading in this article (and others about this topic) are unsubstantiated rumors. That's it.
What is interesting is that if it’s true that testing has been loosened, it’s for production that’s coming on line, but not available yet. So the first few million will have the more tightly tested projectors, and the rest will have the more loosely tested projectors.
so if that is true, then an early production phone can be tested against a later production phone. But to be a useful test, rather than just a test of production variations, a number of phones need to be tested. The difference has to be a statistical significance. It needs to be greater than normal production variances.
Sure, once the iPhone X sells in the millions, then labs can start getting statistical data on the accuracy of the sensors. If there is a clear difference in sensor components or sensor settings, then that would be evidence that some change took place in the manufacturing process. (I could then imagine the resulting lawsuits.)
* But the reality is that all of this is speculation based on rumors from business websites. We know nothing for certain about this right now.
Which is why I say “...if that’s true”. It would be a pretty boring discussion if all we said was...”we don’t know if it’s true”
I agree to this extent; beyond stating that "we don't know if it's true", I'm not very interested in speculating about possible Apple failures. But I do understand that a lot of people enjoy discussing some disaster that might happen to Apple. There is a niche in tech journalism devoted just to those possible Apple failure tropes (which is followed on the MacWorld website by the Macalope column).
Sometimes that interesting too. If we don’t discuss failures, they will be repeated again. We can be sure it’s discussed at Apple.
Just send an iPhone X my way and I'll test the accuracy a million times.
In a random test wouldn't you just have to get past 500k to have a greater than 50% chance?
But it's probably less than that since the closest physical match to your facial features are likely the people closest to you, which is why Face ID needs to be 20x that of Touch ID since fingerprint patterns are more unique between siblings, parents, and offspring, even though finger prints are heavily influenced by genetic factors (like basic size, shape, and spacing of dermatoglyphs).
Its both more and less secure depending on who you think you should protect against. Like you said, a family member is probably one of those one in millions, so a random person's chance of matching all those features are a lot lower. The nice thing is you can actually test to see if your family member are unlocking your phones, unlike with touch ID (just by exposing them to your phone).
In the future, they could just ban those people specifically (at the risk of increasing false negatives for yourself), cause there is undoubtably a margin of error in the mapping and thus they accept matches when it is close enough, says 99.97% of points match you and 99.95% matches your sibling, maybe they would normally consider this acceptable but knowing this variant is not merely an error but a different person it could rule it out.
The accuracy of FaceID is not the question we should be asking, I wonder why we need this technology on our phones to start with. Is this the start of a huge change in how we will be identified in the future for everything like passport control, age verification for purchasing alcohol, etc... Is the this the beginning of the end to the last shred of privacy some of us our holding onto? Will this facial information find its way to a huge database shared by corporations, will insurance companies raise rates on faces that appear over weight, car insurance companies could decide you drink to much and raise your rates, will we all be tracked by readers on every city street?
I really wish Apple had not gone there with this technology and it is the single reason I will not buy an iPhone or any cell phone that uses this as the only way to unlock that device.
The accuracy of FaceID is not the question we should be asking, I wonder why we need this technology on our phones to start with. Is this the start of a huge change in how we will be identified in the future for everything like passport control, age verification for purchasing alcohol, etc... Is the this the beginning of the end to the last shred of privacy some of us our holding onto? Will this facial information find its way to a huge database shared by corporations, will insurance companies raise rates on faces that appear over weight, car insurance companies could decide you drink to much and raise your rates, will we all be tracked by readers on every city street?
I really wish Apple had not gone there with this technology and it is the single reason I will not buy an iPhone or any cell phone that uses this as the only way to unlock that device.
I guess you should cue up to buy older phones in a few years, because if you want a flagship model from anywhere, they will all be using this. And when these models drop in price over the next few years, other than for the cheapest phones, there won’t be one without this. Apple is trying very hard to protect your privacy with this. When Google supports it, things will be different, you can be sure.
The accuracy of FaceID is not the question we should be asking, I wonder why we need this technology on our phones to start with. Is this the start of a huge change in how we will be identified in the future for everything like passport control, age verification for purchasing alcohol, etc... Is the this the beginning of the end to the last shred of privacy some of us our holding onto? Will this facial information find its way to a huge database shared by corporations, will insurance companies raise rates on faces that appear over weight, car insurance companies could decide you drink to much and raise your rates, will we all be tracked by readers on every city street?
I really wish Apple had not gone there with this technology and it is the single reason I will not buy an iPhone or any cell phone that uses this as the only way to unlock that device.
Why do you think that facial recognition databases aren't a thing outside of Apple now?
Should I believe Apple that has over 10 years experience in making mobile phones and has the trust of millions of people or should I believe failing Bloomberg that has a consistent record of lying to the public. It’s not that difficult a choice in my opinion.
Failing Bloomberg? I don't think so. They make a fortune selling those terminals.
If the rumor had come out of Gizmodo or CNET, it would be easy to dismiss. But because Bloomberg is highly credible and non-biased, this report has impacted AAPL stock valuation, meaning people are betting millions of dollars on the report's veracity. So now that Apple has made a comment (however contrived "and we can't wait..."), maybe look at how the stock does to decide who to believe.
We should base who to believe on..stock price? Are you shitting me?
I think he meant it the other way around. From what he said, it looks that he’s saying that people are selling off because of the report, which indeed does happen all the time.
but it’s true ( as I think he also believes) that some people will think that big investors will sell because they believe the reports themselves, and so some people will look at the stock and buy or sell because they think the price reflects the truth of the report if big investors are selling. After all, many people believe that big investors have inside information, and trade on that.
its convoluted reasoning. But many people feel powerless, and so will follow the market because of that reasoning.
Apple is up 2.07% on a down market day as of 2:02PM EDT, so people don't seem to be selling off. It's probably up because of reported increased sales of the iPhone8 in China. The stock is just 1.6 points below its 52-week high.
Once the iPhoneX is in users hands I'm sure a whole lot of people will try and break the Face ID with lookalikes, masks, photographs, CGI, etc. And the opposite - that if someone changes their look, shaves their beard, cuts and dyes their hair, shaves off their eyebrows, that it might no longer let one in. We'll know soon enough, but it's hard for me to believe that Apple would release it if it wasn't working well (for obvious reasons).
Comments
But I do understand that a lot of people enjoy discussing some disaster that might happen to Apple. There is a niche in tech journalism devoted just to those possible Apple failure tropes (which is followed on the MacWorld website by the Macalope column).
In the future, they could just ban those people specifically (at the risk of increasing false negatives for yourself), cause there is undoubtably a margin of error in the mapping and thus they accept matches when it is close enough, says 99.97% of points match you and 99.95% matches your sibling, maybe they would normally consider this acceptable but knowing this variant is not merely an error but a different person it could rule it out.
I really wish Apple had not gone there with this technology and it is the single reason I will not buy an iPhone or any cell phone that uses this as the only way to unlock that device.
Once the iPhoneX is in users hands I'm sure a whole lot of people will try and break the Face ID with lookalikes, masks, photographs, CGI, etc. And the opposite - that if someone changes their look, shaves their beard, cuts and dyes their hair, shaves off their eyebrows, that it might no longer let one in. We'll know soon enough, but it's hard for me to believe that Apple would release it if it wasn't working well (for obvious reasons).