What I meant was that consumer protection in general tends to be a lower priority for Republicans, which compounds the problems caused by the corruption.
This isn't really like the coffee case. In that case, the product harmed the actual user of the product. This is more like a home safety system that watches for unknown home invaders, and because of a bug, occasionally shoots random strangers that walk by on the street, incorrectly believing that they are inside the house. The owner of the home safety system is never harmed directly, but the product still causes harm, even when used as intended, even without any negligence on the part of the user.
Additionally, the ruling ignores that the harm caused by false advertising can be indirect. Users were told that these products were more secure than the competition when they were less so. That false advertising harms the free market by unjustly encouraging people to buy one product over another. This unfair competition presumptively causes indirect consumer harm by reducing competition. The onus should be on the false advertiser to prove that their false advertising does not constitute unfair competition, not on the government to prove that consumers were somehow directly harmed.
Lawsuits are for righting wrongs. If you can't show anyone was wronged, then there is nothing to right.
But there's ample proof that people were harmed by the Mirai botnet, and much of that harm was the direct result of D-Link routers getting p0wn3d. What they lacked was proof that the owners of the devices were harmed, and the judge incorrectly jumped from "the owners weren't harmed" to "no one was harmed", when in fact that is clearly not the case.
Why is he wrong if the burden is on the plaintiff to show actual harm, and the plaintiff could not show actual harm?
But there was actual harm. The Mirai botnet attacked other computers on the Internet, and as a part of that botnet, D-Link's routers probably did tens of millions of dollars of economic damage to the Internet as a whole. So there was very clearly harm. It just wasn't directed specifically at the owners of the devices. Rather, the owners of the devices were unknowingly being complicit in that harm to others.
IMO, the judge is wrong in this case. This sort of action shouldn't require showing harm to individuals, because the harm isn't necessarily to the individual device owners. Most of the harm is to the people in aggregate.
Devices with security holes on the public Internet invariably eventually turn into botnets that attack systems in a distributed fashion, which harms the companies being attacked and the users that get locked out of their accounts. The harm to the owners is negligible, because they lose just a tiny bit of bandwidth. But the harm to society is huge.
And even in cases where the harm is to the individual owner, the harm could be impossible to prove, because you could never realistically be certain whether a password shared by several websites got stolen from one of those websites or from the unencrypted copy of the password on the user's router. But that doesn't mean that users weren't harmed. In effect, if this judge's opinion is allowed to stand, the government will be unable to prosecute the vast majority of cases in which consumers are harmed en masse by security-related negligence, and that's a bad thing.
So, you think the Chunnel could now be dug for less than 20 million dollars? Now THAT'S optimism!
A decade from now, given sufficient effort spent on automated boring, not now. And even then, not the Chunnel, because it is underwater.
You forgot the "under a waterway" part. That made construction a lot more complex.
Why would that be more complex?
Because saturated dirt tends to collapse relatively easily, and it only takes a small error to punch through into water and flood the tunnel. Also, you have to build the tunnel walls in such a way that they are completely watertight. Both of those problems basically go away if you're tunneling through solid rock or even compacted, dry dirt.
The real problem is that in order for the algorithm to work, there have to be enough people out there who are in fact buying the items together. Either those items are often bought together for some other non-bomb-making purpose -- people buying plastic buckets, nails, engine oil and high-nitrogen fertilzer because they are DIY gardners who are working on a shed to hold their lawnmower -- or there are enough people out there building bombs with ingredients purchased together on Amazon that their algorithm can identify the pattern and promote it over the alternative patterns of engine oil and lawnmower parts, fertilizer and garden trowels, nails and hammers and saws, plastic buckets and cleaning supplies.
You left out the third (and IMO most likely) possibility: that a bunch of high school kids decided to troll Amazon by ordering those things together (and probably canceling the order afterwards) to trick the algorithm into believing that it is some common combination.
Wasn't the entire point of "wifi assist" to deal with this very issue, phones connected to bad access points that couldn't really pass data?
In theory, yes. In practice, it seems to be frequently broken. I routinely connect to a computer-to-computer network used for controlling some hardware that has no Internet connection. Frequently, either:
I have no networking because Wi-Fi assist failed to kick in (it doesn't tell me that the network has no Internet connection)
I have no connection to devices on the disconnected Wi-Fi network—presumably because Wi-Fi assist was too aggressive—and I have to temporarily put the phone into Airplane mode, turn on Wi-Fi to force it to connect, do something on the offline network, then turn Airplane mode back off.
The latter failure mode seems to be more common lately, but I've seen both behaviors sporadically. My guess would be that Wi-Fi assist fails around 10% of the time. It's better than nothing, but IMO it still doesn't rise to the level of "good", much less "robust" or "reliable".
Much more than that doesn't seem too helpful though, three months is a whole lot of searches, and should give plenty of information about what I'm searching for right now.
Maybe, but maybe not. There are very good reasons for keeping data longer than three months. Not all data will be valuable after three months, but it doesn't take much effort at all to come up with counterexamples in which longer retention could make a significant difference in search quality.
For example, consider people who are in college (and younger). For about three months out of the year, they're doing something entirely different from what they do during the other nine months. If you retained data for only three months, you'd lose their entire history of school-related searches by the time they started back at the end of the summer.
Or consider people who are planning their summer vacations. The search history from when they planned their previous summer vacation is suddenly very relevant, even though it is about a year old.
That historical data starts to become important when words have multiple meanings. For example, when I'm searching for "Apple yield", I'm probably looking for information about the DOA rate for a model of iPhone, whereas a farmer is probably looking for ways to keep the caterpillars off of fruit. Mind you, you're unlikely to need very much history to distinguish between those two particular examples, but there's no reason that semantic information in general can't be intuited based on things that you have searched for less frequently.
I also question the fundamental assumptions upon which these conclusions were based. They appear to be assuming that you can derive search quality based on the number of people coming to a site from a particular search engine over a couple of years' time scale. That's not a given, for several reasons:
It can take some time for changes in search quality to produce meaningful changes in the total number of users using a given service, thanks to a combination of momentum, pre-installation of software on devices, and other factors entirely unrelated to the service itself.
This ignores the critical question of whether Yahoo and Bing had comparable search quality before making those potentially quality-reducing changes. If not, then the factors keeping those users on those search sites were something other than search quality, and thus the numbers would be largely unaffected by additional reductions in search quality.
Changes to downstream sites' rankings for common queries could trivially produce swings in referrer count from a given search engine that would dwarf the swings caused by users switching search engines.
Search quality does not merely affect the number of people sent to a site. It also affects how many of the right people go to a given site versus people who were actually searching for something unrelated that is spelled similarly. So even if everything else is constant, search quality could be changing significantly.
That said, I think perhaps the best refutation involves pointing out that these companies are highly data-driven, constantly using statistics to justify everything they do. They are spending billions of dollars on data centers to store all of that data, and they are constantly running new data processing jobs that expend tremendous computing resources to reprocess that historical data. Don't you think if the benefits of storing and regularly reprocessing all of that extra data were truly minimal, they would have run an experiment on whether processing only the most recent 'n' months gives comparable results, come to that conclusion, and started retaining less data by now?
I mean, I'm not saying that they're definitely wrong, because anything is possible, but I wouldn't take that bet in a million years.
No, seriously. Resetting preferences is a silly concept. It almost never fixes anything. When it does, the misbehavior is still a bug, and will still recur if you somehow manage to get your preferences set back to exactly the way they were. Worse, when it does, you've lost the state information that would help developers figure out what is wrong, which means that if you don't manage to get your preferences set to the same state, the bug will end up biting somebody else down the line.
Why does that button even exist? Nuking your preferences should be a last resort way of getting back up and running, and such last resorts should be as painful as possible to discourage their use as much as possible.
Let's skip right to your Chunnel example. The Chuunel cost 21 BILLION dollars. To go 24 miles.
You forgot the "under a waterway" part. That made construction a lot more complex. Compare that with the Seikan tunnel, which is 40% longer, but has a significantly smaller underwater portion. It cost only $3.6 billion dollars.
Even if Musk makes an order of magnitude improvement to this...
I said I expected an order of magnitude improvement in equipment cost (at least), times an order of magnitude improvement in speed (which translates to an order of magnitude less worker time), times at least an order of magnitude from the use of automated systems to replace human labor. I would expect 3+ orders of magnitude reduction in cost.
I agree that you can make the PER MILE cost lower by digging more due to economies of scale, but it's still going to be stupid expensive to dig a nationwide hyperloop system.
Of course. But most of that cost is the R&D cost up front, and you can spread that cost out over the cost of the entire network (plus networks on other continents).
Put yourself back in 1960. Someone says that they need a megaFLOP of computing power. By your logic, the right response is to say, "Go look up the cost of computing projects. They are CRAZY expensive, and that's for a few FLOPS." Yet now, just a few decades later, many of us have computers with gigaFLOPS on our wrists.
Of course building tunnels is expensive now:
Other than the boring itself, everything is done by hand.
Current-generation boring machines move something on the order of eight or ten feet per day.
There's not a big market for boring machines, so they're very expensive to build and very expensive to rent.
There's not much call for workers trained in operating them, so the operator labor is probably expensive, too.
The tunnel must be continuously safe for occupation by the people working there (even with diesel equipment running inside), because people dying is bad for business.
In the 1870s, they built a railroad tunnel in the greater SF Bay Area that was only a little over a mile long and wide enough for one small passenger train. It took 27 months. Just over a hundred years later, it took 71 months to build the Chunnel, a 24-mile triple tunnel (two rail tunnels, one service tunnel) that can handle double-decker auto trains. That's less than 3x as long to remove on the order of 100x as much material. And that was still largely a one-off project, and still fairly manual in nature.
A hyperloop, by contrast, would not be a one-off project. This would be building something on the same order of magnitude as the Interstate highway system. So ask yourself how many road paving machines are in existence, and you'll start to understand why any assumptions based on current boring technology are about as applicable to a hyperloop as the cost of slaves and cobblestones are to the cost of building an Interstate highway.
There is a ton of data on that available from Crossrail in London. (The money itself is generally measured in metric megatons).
Part of the reason that it costs so much is that there isn't much tunneling being done, which makes it a very niche industry involving huge amounts of manual labor at every step of the process. If the U.S. decided to build the sort of network of underground rail tunnels that Musk proposes, assuming Musk is involved (as opposed to a typical, grossly over-budget government project), the scale would be so huge that it would be completely infeasible using current technology, which is why they're working on the technology to make it possible:
The first key piece is an automated excavator setup wherein the digging and retaining wall construction can happen with minimal human intervention and at an order of magnitude faster speed than today. (Musk's The Boring Company is working on this piece.)
The second key piece is an automated trucking system for moving the dirt. (Tesla is working on this piece, whether they realize it or not.)
The third key piece is mass production of that equipment so that the cost per excavator is reduced from a multi-million-dollar machine to a multi-hundred-thousand-dollar machine.
The final piece is building the HyperLoop tech. (SpaceX is working on this piece.)
Notice that these are all the sorts of R&D projects that Elon Musk's companies have experience at doing. And once those costs are incurred, the main incremental costs per mile will be for powdered concrete, water, excavator fuel, replacement boring bits, and lawyers to handle the environmental impact challenges to save the spotted earthworm.
Imagine a network of ten thousand excavators all digging 100 feet per day each, for a total of 189 miles per day. At that rate, you could bore a tunnel from coast to coast in two weeks, give or take. And then you could do it again. And again. And again. To be fair, it would take longer than that in practice, because the boring machines would be much slower when going through rock, but you get the point. When tunneling at that sort of scale, you don't build an excavator. You built a giant fleet of excavators the likes of which have never been seen on the planet Earth. And that's what brings the cost down.
Oh, is THAT all? How long (and how much money) do you think this underground tunnel will take?
Why do you think Elon Musk started "The Boring Company"? These projects are all tied together. Anybody who thinks Musk's hyperloop will be above ground is kidding him/herself. I mean maybe there might be short stretches somewhere, like when crossing canyons or something, but I think it's safe to say that the intent is for it to be predominantly (if not completely) underground, which makes the terrorism discussion largely moot.
Explosive charges attached to the tube that detonated five seconds before the arrival of a pod would likely kill everyone on board.
First, that's probably not true; there's no reason you can't design the nose to be aerodynamic enough to survive the air impact, and there's a good chance some people would survive the impact from the pod car falling out of the tube, because they would hit the ground at an angle.
Second, even if it were true, explosive charges attached to the tube that detonate as the car is passing overhead would definitely kill everyone on board, and the same would be true for a train, a subway, a bus, a taxi, or a bicycle. If the sole safety argument against it is that the higher speeds allow a terrorist to be less accurate in his/her timing, then that's probably a good indication that there's no meaningful increase in safety risk over existing modes of transportation.
Third, all you have to do to eliminate that added risk entirely is to use an underground tunnel instead of an elevated tube. With that change, the would-be bombers would have to survive in near-vacuum conditions (less atmospheric pressure than on the surface of Mars) if they wanted to plant a bomb in the tube, which, of course, would be rather implausible.
It's only potentially a privacy violation if you can be readily identified anyway. A photo of a random person taken at an unspecified date and time could ostensibly be a photo of any of thousands of people who have, at some point in the past, looked enough like that photo when shot at that angle, so long as those people have been in that city at the right time of year between when the building in the background was built and torn down. If there are no trees, then the time of year doesn't matter. If there are no identifiable buildings, then it could have been taken at any time (ignoring identifiable changes in film chemistry, DPI, etc.) over the course of many decades, anywhere in the world, so the potential list of candidates is huge.
What makes it a privacy problem in the modern age is that the photo has a precise timestamp, and Facebook has so much data about which family members of each possible candidate were in a given city on that date that they can guess who is in the photo with reasonable accuracy, and thus de-anonymize what would otherwise be a photo of an anonymous person. That is the privacy violation—not the picture itself, but rather the giant repository of data that can be trivially used to unmask people who wish to remain anonymous.
This. Trusting even a first-party server is a flaw, much less a third-party server. If it is possible to significantly harm users by replacing official data with malicious data, that data should be signed, and the app should refuse to accept data that is not signed properly.
I am not saying they said the truth. And it probably is a bug that needs to be corrected in the code. There is so much pressure to release products on time that you need to bug-fix soon after release. In this case, who knows what really happened. Maybe they do not even know it yet. But they had to come up with some story for PR purpose.
Yeah, it's called saving face. The only problem is that they saved face afterwards instead of saving everyone's face ahead of time.
As the stage was full of white males. Note how while 50% of the on ground workforce at Apple is Indians the executives have no Indian representation.
First, that's not true. It's nowhere near half. The latest figures from any reputable source were one third, and even that is dubious, because Apple doesn't break down its diversity numbers based on country of origin, which means there's a lot of guessing involved in that number.
But even if it were true, it would only be true because Apple has a huge IT organization in India. Not software; not hardware; IT. You can safely assume that they are not going to fly IT people over from India for a keynote even if somehow there were a big IT component to that keynote, much less a keynote that's mostly about iPhone hardware and software. The people at the keynote are likely a fairly accurate sampling of the diversity of Apple's Cupertino workforce if you ignore the non-technical staff.
- It's not actually very creepy at all. See this article, for example: https://techcrunch.com/2015/09... [techcrunch.com]
It isn't creepy to you. What you're failing to acknowledge is that different people react differently to always-listening tech. For me, personally, it's the first feature I disable when I get new hardware, and there's a certain percentage of users who feel the same way.
- Not sure you have any data to judge how many users have Hey Siri disabled or enabled. I'd be impressed if you did
Of course not, but that number also doesn't really matter, because the total number of units makes even a tiny percentage important. Each iPhone model year is likely to sell well over 200 million units. Even if only 1% of users disable always-listening Siri, that's still 2 million people with the feature disabled. And even if only 1% of those occasionally unlock it with their face while driving, that's still 20,000 people who are now looking at their phone who likely would not have looked at their phone in previous versions. And even if only one tenth of one percent of those people get into fatal car accidents as a result of doing so, that's still 20 fatal car accidents and roughly 22 people dead as a direct result of this unlock mechanism replacing the fingerprint unlock.
And I suspect that those numbers are conservative. You don't get to hide behind statistics when you're talking about such a large pool of users.
- A touch sensor on the back sounds like a more dangerous implementation for driving, not less. Fiddling around on the back of a phone is not a safe thing to do!
I'm in the rare position of regularly using both an iPhone (personal) and a Pixel (work). Because the touch sensor on the back of the Pixel is indented noticeably, it is every bit as easy to use as the one on the front of the iPhone. In fact, if anything, the one on the back of the Pixel might be a little *too* easy to use *accidentally* while picking up the device. The only meaningful difference is which finger you use (because your thumb doesn't typically touch the back of your phone).
Because in previous models, you could trigger Siri eyes-free even with the hotword disabled, and that is no longer possible in the iPhone X. It's a feature regression.
You mean you paid for a 2 year warranty since Apple factored in the EU warranty requirement in their prices.
No, you didn't, because they didn't.
iPhone 7 base model in U.S.: $649
iPhone 7 base model in France: $1022 (€ 909)
To be fair, the EU has a 20% VAT on cellular phones, which brings the real number up to around $772 (I'm subtracting the 1% U.S./China import duty from that $649). And you can safely assume that they provide a 10–15% buffer to avoid having to constantly change the price. That takes it to $849–888. So by my math, you're paying the entire $125 cost of AppleCare for that additional year of warranty, and then some. Now maybe there are some additional duties I'm not aware of that cut into that difference, but the price of an iPhone in Europe has always been higher than in the U.S., and I can only assume that part of the difference covers the cost of the additional mandatory warranty.
Either way, the problem is not really that Apple's warranty is too short, but rather that the cellular companies are allowed to charge for a product over a period that is longer than the warranty period. They should have two options: bill the phone over a year or build the cost of an extended warranty into the purchase price in exchange for the right to bill the phone over a longer period. Either way, IMO, it is the cell phone companies' responsibility to provide that warranty extension, not the manufacturer's responsibility.
Slashdot Mirror
ATT&T-Mobile.
What I meant was that consumer protection in general tends to be a lower priority for Republicans, which compounds the problems caused by the corruption.
Unfortunately, that's quite possible, particularly given the current political climate in Washington.
This isn't really like the coffee case. In that case, the product harmed the actual user of the product. This is more like a home safety system that watches for unknown home invaders, and because of a bug, occasionally shoots random strangers that walk by on the street, incorrectly believing that they are inside the house. The owner of the home safety system is never harmed directly, but the product still causes harm, even when used as intended, even without any negligence on the part of the user.
Additionally, the ruling ignores that the harm caused by false advertising can be indirect. Users were told that these products were more secure than the competition when they were less so. That false advertising harms the free market by unjustly encouraging people to buy one product over another. This unfair competition presumptively causes indirect consumer harm by reducing competition. The onus should be on the false advertiser to prove that their false advertising does not constitute unfair competition, not on the government to prove that consumers were somehow directly harmed.
But there's ample proof that people were harmed by the Mirai botnet, and much of that harm was the direct result of D-Link routers getting p0wn3d. What they lacked was proof that the owners of the devices were harmed, and the judge incorrectly jumped from "the owners weren't harmed" to "no one was harmed", when in fact that is clearly not the case.
But there was actual harm. The Mirai botnet attacked other computers on the Internet, and as a part of that botnet, D-Link's routers probably did tens of millions of dollars of economic damage to the Internet as a whole. So there was very clearly harm. It just wasn't directed specifically at the owners of the devices. Rather, the owners of the devices were unknowingly being complicit in that harm to others.
IMO, the judge is wrong in this case. This sort of action shouldn't require showing harm to individuals, because the harm isn't necessarily to the individual device owners. Most of the harm is to the people in aggregate.
Devices with security holes on the public Internet invariably eventually turn into botnets that attack systems in a distributed fashion, which harms the companies being attacked and the users that get locked out of their accounts. The harm to the owners is negligible, because they lose just a tiny bit of bandwidth. But the harm to society is huge.
And even in cases where the harm is to the individual owner, the harm could be impossible to prove, because you could never realistically be certain whether a password shared by several websites got stolen from one of those websites or from the unencrypted copy of the password on the user's router. But that doesn't mean that users weren't harmed. In effect, if this judge's opinion is allowed to stand, the government will be unable to prosecute the vast majority of cases in which consumers are harmed en masse by security-related negligence, and that's a bad thing.
A decade from now, given sufficient effort spent on automated boring, not now. And even then, not the Chunnel, because it is underwater.
Because saturated dirt tends to collapse relatively easily, and it only takes a small error to punch through into water and flood the tunnel. Also, you have to build the tunnel walls in such a way that they are completely watertight. Both of those problems basically go away if you're tunneling through solid rock or even compacted, dry dirt.
You left out the third (and IMO most likely) possibility: that a bunch of high school kids decided to troll Amazon by ordering those things together (and probably canceling the order afterwards) to trick the algorithm into believing that it is some common combination.
In theory, yes. In practice, it seems to be frequently broken. I routinely connect to a computer-to-computer network used for controlling some hardware that has no Internet connection. Frequently, either:
The latter failure mode seems to be more common lately, but I've seen both behaviors sporadically. My guess would be that Wi-Fi assist fails around 10% of the time. It's better than nothing, but IMO it still doesn't rise to the level of "good", much less "robust" or "reliable".
Maybe, but maybe not. There are very good reasons for keeping data longer than three months. Not all data will be valuable after three months, but it doesn't take much effort at all to come up with counterexamples in which longer retention could make a significant difference in search quality.
For example, consider people who are in college (and younger). For about three months out of the year, they're doing something entirely different from what they do during the other nine months. If you retained data for only three months, you'd lose their entire history of school-related searches by the time they started back at the end of the summer.
Or consider people who are planning their summer vacations. The search history from when they planned their previous summer vacation is suddenly very relevant, even though it is about a year old.
That historical data starts to become important when words have multiple meanings. For example, when I'm searching for "Apple yield", I'm probably looking for information about the DOA rate for a model of iPhone, whereas a farmer is probably looking for ways to keep the caterpillars off of fruit. Mind you, you're unlikely to need very much history to distinguish between those two particular examples, but there's no reason that semantic information in general can't be intuited based on things that you have searched for less frequently.
I also question the fundamental assumptions upon which these conclusions were based. They appear to be assuming that you can derive search quality based on the number of people coming to a site from a particular search engine over a couple of years' time scale. That's not a given, for several reasons:
That said, I think perhaps the best refutation involves pointing out that these companies are highly data-driven, constantly using statistics to justify everything they do. They are spending billions of dollars on data centers to store all of that data, and they are constantly running new data processing jobs that expend tremendous computing resources to reprocess that historical data. Don't you think if the benefits of storing and regularly reprocessing all of that extra data were truly minimal, they would have run an experiment on whether processing only the most recent 'n' months gives comparable results, come to that conclusion, and started retaining less data by now?
I mean, I'm not saying that they're definitely wrong, because anything is possible, but I wouldn't take that bet in a million years.
Why do we even have that button?
No, seriously. Resetting preferences is a silly concept. It almost never fixes anything. When it does, the misbehavior is still a bug, and will still recur if you somehow manage to get your preferences set back to exactly the way they were. Worse, when it does, you've lost the state information that would help developers figure out what is wrong, which means that if you don't manage to get your preferences set to the same state, the bug will end up biting somebody else down the line.
Why does that button even exist? Nuking your preferences should be a last resort way of getting back up and running, and such last resorts should be as painful as possible to discourage their use as much as possible.
You forgot the "under a waterway" part. That made construction a lot more complex. Compare that with the Seikan tunnel, which is 40% longer, but has a significantly smaller underwater portion. It cost only $3.6 billion dollars.
I said I expected an order of magnitude improvement in equipment cost (at least), times an order of magnitude improvement in speed (which translates to an order of magnitude less worker time), times at least an order of magnitude from the use of automated systems to replace human labor. I would expect 3+ orders of magnitude reduction in cost.
Of course. But most of that cost is the R&D cost up front, and you can spread that cost out over the cost of the entire network (plus networks on other continents).
Put yourself back in 1960. Someone says that they need a megaFLOP of computing power. By your logic, the right response is to say, "Go look up the cost of computing projects. They are CRAZY expensive, and that's for a few FLOPS." Yet now, just a few decades later, many of us have computers with gigaFLOPS on our wrists.
Of course building tunnels is expensive now:
In the 1870s, they built a railroad tunnel in the greater SF Bay Area that was only a little over a mile long and wide enough for one small passenger train. It took 27 months. Just over a hundred years later, it took 71 months to build the Chunnel, a 24-mile triple tunnel (two rail tunnels, one service tunnel) that can handle double-decker auto trains. That's less than 3x as long to remove on the order of 100x as much material. And that was still largely a one-off project, and still fairly manual in nature.
A hyperloop, by contrast, would not be a one-off project. This would be building something on the same order of magnitude as the Interstate highway system. So ask yourself how many road paving machines are in existence, and you'll start to understand why any assumptions based on current boring technology are about as applicable to a hyperloop as the cost of slaves and cobblestones are to the cost of building an Interstate highway.
Part of the reason that it costs so much is that there isn't much tunneling being done, which makes it a very niche industry involving huge amounts of manual labor at every step of the process. If the U.S. decided to build the sort of network of underground rail tunnels that Musk proposes, assuming Musk is involved (as opposed to a typical, grossly over-budget government project), the scale would be so huge that it would be completely infeasible using current technology, which is why they're working on the technology to make it possible:
Notice that these are all the sorts of R&D projects that Elon Musk's companies have experience at doing. And once those costs are incurred, the main incremental costs per mile will be for powdered concrete, water, excavator fuel, replacement boring bits, and lawyers to handle the environmental impact challenges to save the spotted earthworm.
Imagine a network of ten thousand excavators all digging 100 feet per day each, for a total of 189 miles per day. At that rate, you could bore a tunnel from coast to coast in two weeks, give or take. And then you could do it again. And again. And again. To be fair, it would take longer than that in practice, because the boring machines would be much slower when going through rock, but you get the point. When tunneling at that sort of scale, you don't build an excavator. You built a giant fleet of excavators the likes of which have never been seen on the planet Earth. And that's what brings the cost down.
Why do you think Elon Musk started "The Boring Company"? These projects are all tied together. Anybody who thinks Musk's hyperloop will be above ground is kidding him/herself. I mean maybe there might be short stretches somewhere, like when crossing canyons or something, but I think it's safe to say that the intent is for it to be predominantly (if not completely) underground, which makes the terrorism discussion largely moot.
Explosive charges attached to the tube that detonated five seconds before the arrival of a pod would likely kill everyone on board.
First, that's probably not true; there's no reason you can't design the nose to be aerodynamic enough to survive the air impact, and there's a good chance some people would survive the impact from the pod car falling out of the tube, because they would hit the ground at an angle.
Second, even if it were true, explosive charges attached to the tube that detonate as the car is passing overhead would definitely kill everyone on board, and the same would be true for a train, a subway, a bus, a taxi, or a bicycle. If the sole safety argument against it is that the higher speeds allow a terrorist to be less accurate in his/her timing, then that's probably a good indication that there's no meaningful increase in safety risk over existing modes of transportation.
Third, all you have to do to eliminate that added risk entirely is to use an underground tunnel instead of an elevated tube. With that change, the would-be bombers would have to survive in near-vacuum conditions (less atmospheric pressure than on the surface of Mars) if they wanted to plant a bomb in the tube, which, of course, would be rather implausible.
It's only potentially a privacy violation if you can be readily identified anyway. A photo of a random person taken at an unspecified date and time could ostensibly be a photo of any of thousands of people who have, at some point in the past, looked enough like that photo when shot at that angle, so long as those people have been in that city at the right time of year between when the building in the background was built and torn down. If there are no trees, then the time of year doesn't matter. If there are no identifiable buildings, then it could have been taken at any time (ignoring identifiable changes in film chemistry, DPI, etc.) over the course of many decades, anywhere in the world, so the potential list of candidates is huge.
What makes it a privacy problem in the modern age is that the photo has a precise timestamp, and Facebook has so much data about which family members of each possible candidate were in a given city on that date that they can guess who is in the photo with reasonable accuracy, and thus de-anonymize what would otherwise be a photo of an anonymous person. That is the privacy violation—not the picture itself, but rather the giant repository of data that can be trivially used to unmask people who wish to remain anonymous.
This. Trusting even a first-party server is a flaw, much less a third-party server. If it is possible to significantly harm users by replacing official data with malicious data, that data should be signed, and the app should refuse to accept data that is not signed properly.
Yeah, it's called saving face. The only problem is that they saved face afterwards instead of saving everyone's face ahead of time.
First, that's not true. It's nowhere near half. The latest figures from any reputable source were one third, and even that is dubious, because Apple doesn't break down its diversity numbers based on country of origin, which means there's a lot of guessing involved in that number.
But even if it were true, it would only be true because Apple has a huge IT organization in India. Not software; not hardware; IT. You can safely assume that they are not going to fly IT people over from India for a keynote even if somehow there were a big IT component to that keynote, much less a keynote that's mostly about iPhone hardware and software. The people at the keynote are likely a fairly accurate sampling of the diversity of Apple's Cupertino workforce if you ignore the non-technical staff.
It isn't creepy to you. What you're failing to acknowledge is that different people react differently to always-listening tech. For me, personally, it's the first feature I disable when I get new hardware, and there's a certain percentage of users who feel the same way.
Of course not, but that number also doesn't really matter, because the total number of units makes even a tiny percentage important. Each iPhone model year is likely to sell well over 200 million units. Even if only 1% of users disable always-listening Siri, that's still 2 million people with the feature disabled. And even if only 1% of those occasionally unlock it with their face while driving, that's still 20,000 people who are now looking at their phone who likely would not have looked at their phone in previous versions. And even if only one tenth of one percent of those people get into fatal car accidents as a result of doing so, that's still 20 fatal car accidents and roughly 22 people dead as a direct result of this unlock mechanism replacing the fingerprint unlock.
And I suspect that those numbers are conservative. You don't get to hide behind statistics when you're talking about such a large pool of users.
I'm in the rare position of regularly using both an iPhone (personal) and a Pixel (work). Because the touch sensor on the back of the Pixel is indented noticeably, it is every bit as easy to use as the one on the front of the iPhone. In fact, if anything, the one on the back of the Pixel might be a little *too* easy to use *accidentally* while picking up the device. The only meaningful difference is which finger you use (because your thumb doesn't typically touch the back of your phone).
Because in previous models, you could trigger Siri eyes-free even with the hotword disabled, and that is no longer possible in the iPhone X. It's a feature regression.
Like many other people, I've also had it turned off for years.
No, you didn't, because they didn't.
iPhone 7 base model in U.S.: $649
iPhone 7 base model in France: $1022 (€ 909)
To be fair, the EU has a 20% VAT on cellular phones, which brings the real number up to around $772 (I'm subtracting the 1% U.S./China import duty from that $649). And you can safely assume that they provide a 10–15% buffer to avoid having to constantly change the price. That takes it to $849–888. So by my math, you're paying the entire $125 cost of AppleCare for that additional year of warranty, and then some. Now maybe there are some additional duties I'm not aware of that cut into that difference, but the price of an iPhone in Europe has always been higher than in the U.S., and I can only assume that part of the difference covers the cost of the additional mandatory warranty.
Either way, the problem is not really that Apple's warranty is too short, but rather that the cellular companies are allowed to charge for a product over a period that is longer than the warranty period. They should have two options: bill the phone over a year or build the cost of an extended warranty into the purchase price in exchange for the right to bill the phone over a longer period. Either way, IMO, it is the cell phone companies' responsibility to provide that warranty extension, not the manufacturer's responsibility.