You know the auto-industry has huge focus groups (usually done by email) where they present icons to people, like a trunk opening button. But they don't tell the people what it does initially - they ASK them what they think it does.
That's very interesting to hear. A friend called me for help saying "the naked butt light" on her car had turned on. (It's a low tire pressure warning.)
It also took me a while to figure out what the inductor light was. (It has nothing to do with inductors. It's a light to tell you the glow plugs in a diesel engine are still warming up.)
Outside of Apple circles, that's called a SSHD - solid state hard drive. It does not give the best of both types of media.
It's great for speeding up small, frequently-accessed files. Like the files needed to boot the OS or start a program. Hard drives suck at reading or writing these - typically 0.75 to 1.5 MB/s. A small SSD can typically read these at around 10-30 MB/s. This is the only thing it speeds up.
It doesn't do anything for large files because the NAND cache is too small to hold large files, and hard drives are already pretty fast at large files (typically 75-150 MB/s, though newer ones can surpass 200 MB/s).
The NAND size is too small to hit the blazing-fast 500+ MB/s speeds of real SSDs. You need multiple NAND dies operating in parallel for that, and the NAND sizes in SSHDs are too small for that.
It doesn't do anything for infrequently accessed small files because they're not loaded into cache.
It doesn't do anything for writes because that could compromise data integrity if the drive loses power before it's written to the HDD. So writes are sent straight to the HDD portion and proceed at HDD speeds.
Don't get me wrong, it's a substantial speedup in boot times off a slow HDD. I played around with it on a Thinkpad (had a HDD and a 16 SSD configured as cache). Even when I reduced the caching partition down to 8 GB, it was still booting up Windows in about 20 seconds, vs over a minute with the cache disabled. But (1) you don't get the blazing fast SSD speeds (500 MB/s to 2 GB/s) with large files, (2) It doesn't get you these fast speeds with infrequently accessed files, and (3) it doesn't give you these fast speeds with writes. Once you factor in these drawbacks, the performance improvement simply isn't worth the additional cost compared to a SSD + HDD setup.
I only recommend SSHDs if you have a laptop (which typically use 5400 RPM drives), you need the large storage of a HDD, but the laptop can only take a single drive (i.e. doesn't have a separate M.2 slot for a real SSD). On laptop with a M.2 slot you can run with both SSD + HDD. On a desktop a 7200 RPM HDD is tolerable, and you can install a real SSD. In fact on a desktop, you can toss in a small spare used SSD, install Intel Rapid Storage Technology, and configure that SSD to cache all your HDDs.
Apple uses these SSHDs in their all-in-ones (iMac line) and Mac Mini basically because they're too much of a control freak to add a M.2 slot to let you add your own SSD. The iMacs which do allow you to add a smaller M.2-type SSD use a proprietary interface, forcing you to buy the SSD from Apple. The whole thing is a racket designed to extract obnoxious sums of money from their customers for the same features PC users get as standard. And yet somehow Apple users are proud that Apple has the highest profit margin in the PC industry? Talk about Stockholm syndrome.
No, it's worse than password-only. If your account is only protected by a password, then there's no password recovery. You forget your password and you're locked out of the account, permanently. OTOH that means anyone trying to get into your account has to guess/know your password in order to get in.
With this SMS intercept exploit, they can get into your account without knowing your password.
You're thinking of using a SMS in addition to your password in order to login to an account - i.e. 2FA. Yes in that case it's better than password-only (unless it lulls you into picking a poor password because you think you're being protected by the SMS). But that's not what this exploit is about. It's about resetting your password by intercepting a SMS that was supposed to go to your phone. The SMS is used to bypass your password, not to augment it. (In your defense, TFA conflates the two as well, leading to the confusion.)
In other words, it's stupid using 2FA to login, if your password reset procedure is 1FA. Attackers will simply ignore the stronger security to target the weakest link - the 1FA step.
A bacteria species that can metabolize a drug designed to kill cells.
The wiki page on how gemcitabine works is pretty fascinating. The drug as administered doesn't kill cells. It gets modified by enzymes in the cell into a form which interferes with DNA replication and blocks DNA repair. That's what kills the cell.
Presumably the bacteria are just metabolizing it before these enzymes can convert it into its toxic form (or lack one of the needed enzymes). So no, not a superbug which digests a toxic material.
I discovered this when I tried switching from a mouse to an Xbox controller for gaming. With the mouse, you change its position and the direction of the camera changes appropriately. Once you've learned how much mouse motion corresponds to how much angular rotation, you can instantly move the camera from one one direction to another simply by moving the mouse to the appropriate spot on the desk. If you want to rapidly move the camera back and forth between two set directions, it's trivial because you're just moving the mouse between two fixed positions on the desk. You could do it blindfolded.
Not so for the thumbsticks on the XBox-style controller. It doesn't control the direction the camera is pointed. It controls the rate of change of the direction the camera is pointed. You have to push the thumbstick in the direction you want the camera to move, wait for the camera to almost get there, ease up on the thumbstick so the slew rate slows down, ease up some more, ease up some more, then let go when the camera is finally pointed in the desired direction. If you want to rapidly move the camera back and forth between two set directions, it's a lot of work each time, and you need to be watching the camera view to do it.
I emailed one of the developers working on drivers to allow you to use the controller in games which didn't support it, and learned why. The thumbsticks only have 256x256 resolution. That is, there are only 256 discrete measurable directions the stick can be pointing in each axis. This isn't enough resolution for precise aiming, so they have to use the gimpy slew-rate aiming.
For a camera limited to just one degree of motion like it sounds like this one is, you want to use a paddle controller. It's just a potentiometer. You rotate it and the camera rotates along with it. It too has a 1:1 correspondence between direction of the controller and direction of the camera. So you could rapidly move the camera back and forth between two set directions, blindfolded. In particular, if you put a raised ridge on the paddle wheel, the operator can know which direction the camera is pointed by feel, instead of having to read a numerical bearing readout. This is much more intuitive (the operator basically won't need any training) and less prone to error.
We know exactly what will happen. The problem isn't the tube losing its pressure seal. Air is compressible, so it makes a great shock absorber (in fact that is exactly how some shock absorbers are designed - with a gas at one end of a cylinder being pressurized).
The problem is trains moving at high speed tends to do bad things when they hit a stationary solid object. The Eschede derailment probably would've only had a dozen or so fatalities due to losing a wheel at 200 kph. In fact, the wheel failure was in the first car, but the engine and first four cars survived relatively intact, scraping the bridge supports but coasting to a stop or derailing and hitting some trees (the guy sitting above the wheel which failed survived despite being out of seat showing the wheel to the conductor when the accident happened). The bridge collapsed onto the 5th car, causing the rapid deceleration of all subsequent cars. That's where all the fatalities occurred. It was just bad luck the train happened to be passing underneath a bridge just as the accident occurred.
Now, consider that with a hyperloop train, the cars will be traveling at speed a few inches from the stationary wall for the entire length of the track. It's not an air leak you need to worry about. It's an IED-type device placed on the side of the track wall, designed to blow it inwards just before the train arrives. At 700 MPH the explosive only needs enough energy to blow enough of the wall inwards to destroy the first train car causing it to block the subsequent cars. The kinetic energy of the train itself will then be more than sufficient to destroy it. When US Air 427 hit the ground at just 300 MPH, its kinetic energy was enough to shred all the metal into pieces smaller than a sheet of paper. United 93 hit the ground at 563 MPH, and its kinetic energy fragmented the plane into such small pieces that conspiracy theorists (who can't seem to grasp the notion that solid metal will fragment when presented with no other means of shedding kinetic energy) have gone nuts with theories that no plane actually crashed there.
A hyperloop train is going to have more than 4x the kinetic energy per unit mass of US Air 427, 1.5x that of United 93. If one strikes the wall and crashes, it kinetic energy is literally going to turn it (and its occupants) into confetti.
What makes it more dangerous than a plane is that planes fly miles away from the nearest solid object when they're at top speed (mid-air collisions excepted). Even systems designed to cause a deliberate collision (surface to air missiles) have a high failure rate. OTOH Hyperloop is going to be traveling a few inches from the nearest stationary object the entire length of its trip. So you're now faced with the prospect of protecting the entire length of track from vandalism or terrorism.
It is true that compromising the cryptographic proof of blockchain tech may cause identity theft to be far less believed and thus so much more destructive but it is also significantly harder AND the tech is constantly improving.
Yes it makes someone stealing your identity harder. But it makes the proliferation of fake IDs trivial. If there's no central authority, what's to stop someone from flooding the decentralized database with a bunch of fake birth IDs every day. Then in the future if you need a fake birth certificate, you can just pay the guy and he'll send you the private key that corresponds to a birth ID blockchain on the date of your choosing with the proper gender and blood type with a pre-generated name. From TFA:
In the framework that Illinois and Evernym are partnering on, government agencies are expected to verify birth registration information and then âoecryptographically signâ attributes such as legal name, date of birth, sex or blood type.
A rogue employee with access to the system can generate a few extra IDs every day and pocket them. Then sell them on the black market in the future. This is currently prevented by being able to compare a birth certificate against a centralized database at the listed hospital of birth (i.e. the two or more centralized databases at different authorities must agree on the information). But if you're going to eliminate the authoritative centralized databases and rely solely on the blockchain as proof of birth ID...
If you read TFA (I know, I know) they controlled for this. Kids were randomly divided into 3 groups - one was praised for being smart, one praised for behavior, one not praised. The group praised for being smart had a higher incidence of cheating. So the cause and effect is correct.
Summary then does a 180 by linking to a study which speculates praise for being smart reduces motivation to learn. That has cause and effect reversed in my experience. I breezed through high school with little effort, but college actually challenged me so I had a hard time. The study skills most kids had developed in high school to learn stuff which challenged them, I had to develop while in college. So it's not that praise for being smart reduced my motivation to learn. It's that being smart meant I (initially) sucked at learning stuff I found challenging.
The original TFA speculates that praising kids for being smart puts them under the pressure of raised expectations. And the kids do whatever they can to meet those expectations - including cheating.
I'd say they're privacy oriented, like I am. What can you do? The same thing that people have done in the past, refuse to participate in it. Something is only lost when you give up, as it stands there is no "social media" presence for me out there. I don't exist at all among social media networks or anything else. It's not hard to do and still keep a large enough social and work network.
You have a presence in social media even if you've never created an account. You know those little 'f' icons you see on websites which link back to Facebook? They're not a link. They're a script which sets a cookie or examines your cookies to uniquely identify your computer. When you visit slashdot, that 'f' icon in the upper right tells Facebook that user #51853601342 has visited slashdot. And they add it to their database with all other sites user #51853601342.
Then one day a friend of yours sends you a Facebook invite via email. You happen to click on it to delete it (instead of doing a select-delete) which causes it to load in your browser, and now Facebook knows that user #51853601342 is yourname@gmail.com. They start cross-referencing your name with comments, other friend requests, public documents, etc. And now Facebook knows who you are, where you live, who your family is, who your friends are, what you look like (thanks to photo face ID), where you work, who your co-workers are, and how much you make. All without you having an account.
It's not enough to avoid creating social media accounts. You also have to run script blockers specifically targeting these tracking scripts (e.g. Ghostery) and/or browse in incognito mode at all times.
That's kinda the point though. As long as your TV cable company has a monopoly on high-speed Internet in your area, if they start losing money from too many people canceling cable TV, they will just raise the price of cable Internet to compensate. And the cord cutters won't be reaping huge savings.
The price the market will bear for a product, and the cost to supply that product can be very different. Market price falls somewhere in between these two. Competition is what pushes the price close to the cost to supply. Without competition the price gets pushed close to what the market will bear, which is what the cable companies have gotten fat off of for decades thanks to local governments granting them legalized monopolies.
Switching from cable TV to cable internet doesn't change the price the market will bear. So the price won't change as long as these monopolies persist. You're just jumping from the low end of the see-saw to the high end. It raises you up (get more for paying less) only until enough people do it, and the see-saw (cable company) changes the pricing to tip things back to their previous profit margin.
There are no doctors without medical degrees. There are no lawyers without law degrees.
At their core, law and medicine are regurgitation professions. You learn a bunch of facts (a helluva lot of facts), so you can sift through them in your mind, find the ones most relevant to a specific case or patient, and regurgitate them. The licenses needed to practice in those professions are certifications that you've learned enough of those facts to professionally advise people.
While all professions require some degree of regurgitation, they differ in how important it is. At the opposite end of the spectrum are creative professions, which are almost entirely based on skill and talent. It's impossible to certify these professions because often times the talent is in the eye of the beholder (most slashdotters consider most TV programming to be banal, yet it's still incredibly successful).
Working with computers falls somewhere in the middle. Programming falls more towards the creative end, IT more towards the regurgitation end (which is why there are a lot of certifications for IT). But there's a significant amount of crossover (programmers still need to learn algorithms, IT still needs to come up with creative systems or configurations of equipment).
Yet somehow, tech seems to be the one place where a degree is considered near irrelevant
Back during the tech bubble, when programmers with CS degrees were very hard to come by, I was talking with the CEO of a financial analysis company while we walked our dogs in the park. He hired CS grads when he could. His second choice was other STEM graduates (his company made financial simulations, which were very similar to the physical simulations physics majors and engineers were experienced with).
His third choice was a bit surprising - music majors. Music is fundamentally based on mathematical patterns, which apparently makes music majors on average better than the general population at programming and coming up with simulations which try to detect subtle mathematical patterns in finance.
Yet this line of thinking has apparently led to the collapse of the US consumer credit system.
Let's not jump to conclusions. It's still unclear whether or not Mauldin had anything to do with the company being hacked, or even if she wasn't qualified (people are assuming she isn't because of her major). My bet is still on management deciding that money to hire competent IT security professionals was a waste because they'd never suffered a major hack. Some of the other things Equifax did that have come to light (using a timestamp as a PIN, storing passwords as plaintext instead of as hashes, and crucially - not patching their Apache servers immediately after the explain became public and then was fixed) point to the same flawed management decisions when it came to hiring programmers and admins.
That doesn't really make much sense. Technically you're not donating CPU time. You're donating the price of the electricity to run the CPU to perform those calculations. Unless the project is transient (e.g. crack RC5-64, then it's over), the acquisition cost of the CPUs is tiny compared to the operational cost (electricity) to perform the actual calculations. A computer animation company is presumably going to continue to remain in the business of computer animation for decades, so it makes more sense for them to buy their own CPUs/GPUs and pay for their own electricity to run them, rather than use the money on advertising to run the calculations in Javascript much less efficiently.
It's like that stupid idea which comes up every now and then to harvest energy from cars driving over roads. Yes it'll work, but you're not tapping some cheap form of energy. You're stealing energy from all the cars which drive on the road, and you're doing it very inefficiently because of all the mechanical losses involved in transferring energy from the car, to the road, to your energy capture device. The cumulative cost to all the cars (slight decrease in MPG) plus the cost of re-engineering the road to incorporate the energy capture device is several times higher than what it would've cost you to just build a power plant and generate the energy directly.
That's really what the economic viability of this sort of thing hinges on: How energy-efficient is it to perform this sort of activity distributed over millions of people's computers, vs. doing the activity on your own dedicated equipment? In nearly all cases, it's more efficient to do it on dedicated equipment, and this idea is a non-starter.
The malware cryptocurrency mining works because the miners aren't paying for the costs. So even though it's less efficient, by externalizing the electricity cost it becomes a net benefit for the malware authors.
Likewise, Folding@home is less energy- and cost-efficient than if people just sent donations to the project to buy their own equipment and run their own protein folding simulations. The project is successful because people tend to lump their electricity bill into an "I gotta pay it" category, whereas a separate charitable contribution could end up axed when they think "I really don't have the extra money to be donating to this."
Donating CPU time to crack RC5-64 worked because it was a temporary project. Once it was completed, there was no need to do further computations. So if the team had acquired their own computers, they would've had to sell them at the end of the project at a substantial loss (depreciation). Doing it as a distributed project neatly avoided that expense by shifting that cost from the project onto the computer purchase and amortization schedules of computer owners around the world (which is near zero because people buy those computers anyway). But on a commercial scale (not taking charitable donations) this sort of activity is more cheaply performed by temporarily renting servers via a hosting service like Amazon EC2.
Because of the relative inefficiency of Javascript, I can't really think of any activities where distributing it to browsers via ads is really cost- or energy-efficient. Maybe Amazon EC2 might find it useful if there were a sudden temporary spike in CPU demand which exceeded their servers' capability to deliver, and they needed to temporarily find some extra CPU cycles to take the additional load. But in every other case I can think of, it's cheaper to just buy your own computers or to rent time on them directly. The only way selling CPU time via browser ads is competitive is if you're actually stealing CPU time - selling the CPU time for cheaper than the extra electricity cost to the person running the browser. And at that point you've crossed the line from being a legitimate ad to being malware.
Fiscal conservative here. Personally I think these tax break deals for anyone (companies, sports teams, individuals) should be illegal. IMHO they violate the equal protection clause - if I have to pay taxes, other people (and their companies) damn well should have to pay as well.
I am all for competition and market forces in private enterprise. But pitting state or city governments against each other to see who will give you the biggest tax break is just wrong IMHO. The entire reason people create a government is because they want to be treated fairly. Giving tax breaks only to specific individuals or companies defeats government's reason for existing
If you want to create new economic activity by instituting a tax break, give it to all companies, not just one specific company. Government should be in the business of improving society overall. Not in the business of favoring companies and individuals who can leverage what they can offer society into tax breaks. Cities and states should compete with each other to attract business on the basis of offering the lowest overall tax rate, not compete by giving tax breaks to only a select few.
I have a feeling that Republican lawmakers are not quite as sharp as the economists that Foxconn, a $135B company, has on its staff to figure out whether they're getting the better end of the deal...
You're assuming that the economics of this type of deal is zero sum. It's not. It's positive sum. When a new factory is built (and there is market demand for what the factory is producing), everyone wins - the factory, its employees, the surrounding community and government, and the customers of the products the factory produces. Yes someone is probably getting the better end of the deal, but it's not as important if everyone comes out a winner. You are still better off for doing it, than not doing it. (The sports stadium deals where the local government pays to build the stadium are an exception, since it involves actual cash expenditures, not giving up increases in future tax proceeds. The expenditures means the accounting balance for the deal can shift into the red.)
My gripe is with how this type of deal destroys a level playing field between Foxconn and its competitors who don't get the same tax break. You're preemptively shutting out a potential future factory which could have generated even greater benefits than the Foxconn factory. The government shouldn't be placing bets on horses - that's the role of private enterprise (where someone who picks a loser bears the cost of a bad choice themselves, not forces all of society to pay for it).
The current status established by the Supreme Court (in rulings dating back to the 1800s - this isn't some recent legal construct) is that the Constitution only applies to U.S. territory. This is why Bush imprisoned alleged terrorists in Guantanamo Bay - it wasn't U.S. soil, it was Cuban soil, and thus the prisoners there wouldn't have U.S. Constitutional rights. I agree the 100 mile claim by DHS is ridiculous, but as people at border checkpoints who haven't been admitted into the U.S. are technically still outside the country, Constitutional protections do not apply.
Alternative interpretations are not as appealing as one would first think.
If you claim the Constitution applies to U.S. citizens regardless of location, then suddenly non-citizens (both legal and illegal) and even people brought into the country against their will (e.g. alleged terrorists captured in Iraq and Afghanistan) have no basis for claiming Constitutional rights.
If you claim the Constitution applies outside U.S. territory, then you're basically advocating that the U.S. should be allowed to apply its laws to other countries.
Given these three choices, I think limiting the Constitution to U.S. territories is the safest.
Over 100 websites were tracked here during the final weeks of the 2016 U.S. election campaign, producing fake news that mostly favored Republican candidate for President Donald Trump.
They should try self-reflecting on why almost every major news organization and poll showed Clinton winning before the election. Maybe the problem isn't that a handful of websites which favored Trump somehow skewed the election. Maybe it's that the media's expectation for the election was skewed from reality.
The one poll which called the election correctly noticed that Trump supporters were less likely to admit to pollsters that they were going to vote for Trump. When they corrected for this, surprise surprise, their poll showed that Trump would win. That's what happens when you engage in campaigns to shame people with certain political beliefs - they disappear from public view, but still show up in elections (thanks to secret ballots).
Real change comes about from convincing people that your way is right. Not from shaming, ridiculing, and attacking those who disagree with you. Unfortunately, the media and to a lesser extent the Democratic party seems to be trying its hardest not to learn this lesson.
This would all be a lot simpler (and friendlier to the consumer) if the studios offered upgrade pricing like the software industry does. So you could buy the HD version of the movie today. But in the future when you had a 4k projector which fills the entire living room wall and 1080p looks like crap, you could simply pay the price differential between HD and 4k version to upgrade. After all, you've already paid for a license for the movie. All you want to buy are the extra pixels, not an entirely new movie license.
But the movie (and record) industry got used to scamming people by selling them multiple licenses of the same thing during all the format shifts the last couple decades. And they're intent on milking that cow for as long as they can.
I agree most people can't tell the difference between resolutions. (Quiz: Two of the major TV networks in the US use 720p cameras, two use 1080i cameras. Based on the TV you've watched, name each network's resolution.)
However, I also have a projector which throws a 150 inch image (almost floor to ceiling). 1080p looks like crap when enlarged to that size, and I'm anxiously waiting for 4k projectors to come down in price so I can replace my 1080p projector. (Quiz answer: ABC and Fox are 720p, NBC and CBS are 1080i. And if you're a sports aficionado, ESPN is also 720p.)
As the Sci-Fi channel (now SyFy) learned, if you name your company after a word which already has common usage in the field in which you're doing business, you cannot trademark your own company's name. Which causes huge legal and marketing problems.
And the energy does come from the computation. If you were already going to do the computation anyway, then the computation uses energy from the wall socket and produces waste heat. This idea is just finding a use for that waste heat.
Heat pumps are only effective when there's some heat in the outside air to pump. Typically, that's down to around 10 C. Below that, the heat is so scarce it takes more energy to pump it than to generate new heat. If the outside temp drops below that, the typical window-mounted heat pump shuts off and turns on an electric heating element (a resistive heater).
This is why the more ambitious heat pump installations use a geothermal heat sink - loops of cooling (heating) pipes buried underground. The ground stays relatively close to 13 C year-round, providing an good, stable sink to pump heat out of in winter, into in summer.
Anyhow, the PC heater is just finding a use of heat that would be generated anyway. So it's basically free heat.
Crappy oil radiators seem to start at 600 W (about 6 CPUs) and better ones have a power consumption of up to 2500 W (26 CPUs).
Space heaters don't run constantly. They heat up until they (or the room) hits a certain temperature, then shut off. So a typically 1500 W heater (designed not to trip a circuit breaker providing 15A 11V = 1650 W in the U.S.). This electric company estimates a 1500 W heater will use 274 kWh in a month. 1 month is 730 hours, so that's just 375 W of consumption on average. In other words, they estimate it'll be turned on 25% of the time.
So a 95 Watt CPU (125 Watt system) = 500 Watt heater. Just 3 of them will equal a typical 1500 W space heater.
Even if the CPUs reach a rather elevated temperature (1700X maxes out at 95 C), the surface temperature of the rack is only going to be luke warm, so you're not going to get any heat radiated to you. The heat is going to reach you by convection via the fans, which is a crappy way to warm yourself up.
Heat is heat. It's the entropic end-state of energy, so the form of heating doesn't really matter. 100 W of heating is 100 W of heating.
Radiant heat feels better because it triggers the temperature sensing nerves in your skin, but it doesn't really warm you up any more than a convection heater. If you rely on a small radiant heater, you're gonna have to put on a sweater or wrap a blanket around yourself. Because your air temp is going to be a heckuva lot colder than a larger convection (air) heater, and you'll need the insulation to slow down the rate at which your body is losing heat to the air.
The bigger problem with this idea is the relative thermal inefficiency of generating heat from electricity. See, although an electric heater is 100% efficient at converting electricity into heat, its energy source is not. With a localized heating source (oil or gas heat), the process of burning the fuel converts almost 100% of its energy into heat (a little is "lost" as light or vented to the atmosphere to remove exhaust gases). The same is true when generating electricity, except the heat is generated at the plant, not in your house. So if the power generation plant's efficiency is 40% (typical coal plant), then 60% of the energy in the coal becomes waste heat at the plant, and only 40% of the energy becomes electricity to heat your house.
This is why a gas or oil or even a wood heater in your home is preferable to electric heat. And why power plants in cold climates often have co-generation plants which use that waste heat to send steam through pipes for industrial heating. Of course, this is France, which gets about 3/4 of its electricity from nuclear. So they probably don't care about efficiency, just cost.
TFA linked in summary had a lot of scary hype and little info. The vulnerability was found earlier this year and affected companies were notified in April. So they've had several months to work on fixes. The vulnerability was made public recently after giving these companies time to prepare patches.
Microsoft patched it in Windows back in July (Windows Phone was not affected, if you're one of the handful of people still using it).
Apple has fixed it in iOS version 10, but is not patching older version of iOS (they want you to update to version 10).
Google is patching all versions of Android from version 4.4.4 (Kit Kat) and newer. But whether manufacturers and carriers will pass on those patches to end-user devices remains to be seen.
I used to manage an email server and mailing list back in the 1990s. The MBOX format most text-based email programs used for storing mail uses two carriage returns and a "From " at the beginning of the next line as the deliminator for a new mail. That's it.
Occasionally people would send an email to the list which randomly had two blank lines followed by "From " as the first word of the next line. If your (text-based) email client wasn't smart to look at subsequent lines to determine that the person had just randomly typed it in the body of the email rather than the actual start of a new email, it would display it as if it was a new email message.
One day someone sent an email to the mailing list which deliberately abused this. The body was crafted so that the "From" in it and subsequent text was formatted like it was a real, separate email. And the people whose clients interpreted it as a new email got duped into thinking the mail list admin had banned them from the mailing list for inappropriate remarks. The perp was just playing a joke of course, but I shudder to think what modern spammers and phishers would do with that capability.
The professional critics' reviews tell me if the movie is well made, has an intriguing plot, good story development, etc.
The public's reviews tell me if the movie is enjoyable to watch.
Best case, both sets of reviews will be high. But when they diverge, the above is usually what I find.
The bigger problem is ratings inflation. I tend to rate things pretty neutrally. i.e. On a 1-10 scale, the average movie score I give will be a 5.5 (actually it'll be closer to 7 because I use review sites to avoid the really bad movies). But there are lots of people whose average score is a 9 and the lowest score they've ever given a movie is a 7. (There are also people whose average score is a 3, but there tend to be fewer of them.) Review sites try to compensate for this by normalizing scores (basically grade the movie on a curve). But for anonymous reviews this means that my "Ok" rating of 7 gets turned into a "poor" rating. To really normalize properly, you have to give up anonymity and normalize each reviewer's ratings before compiling the scores for all reviews. (Amazon has the same problem. The average review score for all products is about 4.5. So having 4-star reviews doesn't necessarily mean the product is good.)
You do realize Apple doesn't actually make any of this stuff, right? They buy them from the same suppliers who make it for Android phones. They get their flash memory from Toshiba and Samsung (the Samsung memory is slightly faster). RAM is from SK Hynix. They get their LED screens from LG, and will get their OLED screens from Samsung. Their camera is sourced from Sony. The cellular and wireless chipsets are from Qualcomm. The much-hyped headphone jack-less audio is by Cirrus Logic. Same with virtually every component that goes into the iPhone.
The only things which are Apple's are the CPU (which they designed, although they use third party fabs to manufacture it - Samsung and TSMC), the fingerprint scanner (they bought the company which makes them back around 2012), and the software.
So the Apple fans who tell themselves that "Apple makes it best" are deluding themselves as a way to rationalize paying an exorbitant price for the same components which go into Android phones.
Or have people who uploaded those photos to Yelp also uploaded them to Google? I've noticed Google Maps has been pretty aggressive lately asking for photos of places I've visited. Considering Yelp doesn't pay their contributors, it's reasonable to assume contributors are very altruistic people who will honor Google's request and upload their photos there too.
Slashdot Mirror
That's very interesting to hear. A friend called me for help saying "the naked butt light" on her car had turned on. (It's a low tire pressure warning.)
It also took me a while to figure out what the inductor light was. (It has nothing to do with inductors. It's a light to tell you the glow plugs in a diesel engine are still warming up.)
Don't get me wrong, it's a substantial speedup in boot times off a slow HDD. I played around with it on a Thinkpad (had a HDD and a 16 SSD configured as cache). Even when I reduced the caching partition down to 8 GB, it was still booting up Windows in about 20 seconds, vs over a minute with the cache disabled. But (1) you don't get the blazing fast SSD speeds (500 MB/s to 2 GB/s) with large files, (2) It doesn't get you these fast speeds with infrequently accessed files, and (3) it doesn't give you these fast speeds with writes. Once you factor in these drawbacks, the performance improvement simply isn't worth the additional cost compared to a SSD + HDD setup.
I only recommend SSHDs if you have a laptop (which typically use 5400 RPM drives), you need the large storage of a HDD, but the laptop can only take a single drive (i.e. doesn't have a separate M.2 slot for a real SSD). On laptop with a M.2 slot you can run with both SSD + HDD. On a desktop a 7200 RPM HDD is tolerable, and you can install a real SSD. In fact on a desktop, you can toss in a small spare used SSD, install Intel Rapid Storage Technology, and configure that SSD to cache all your HDDs.
Apple uses these SSHDs in their all-in-ones (iMac line) and Mac Mini basically because they're too much of a control freak to add a M.2 slot to let you add your own SSD. The iMacs which do allow you to add a smaller M.2-type SSD use a proprietary interface, forcing you to buy the SSD from Apple. The whole thing is a racket designed to extract obnoxious sums of money from their customers for the same features PC users get as standard. And yet somehow Apple users are proud that Apple has the highest profit margin in the PC industry? Talk about Stockholm syndrome.
No, it's worse than password-only. If your account is only protected by a password, then there's no password recovery. You forget your password and you're locked out of the account, permanently. OTOH that means anyone trying to get into your account has to guess/know your password in order to get in.
With this SMS intercept exploit, they can get into your account without knowing your password.
You're thinking of using a SMS in addition to your password in order to login to an account - i.e. 2FA. Yes in that case it's better than password-only (unless it lulls you into picking a poor password because you think you're being protected by the SMS). But that's not what this exploit is about. It's about resetting your password by intercepting a SMS that was supposed to go to your phone. The SMS is used to bypass your password, not to augment it. (In your defense, TFA conflates the two as well, leading to the confusion.)
In other words, it's stupid using 2FA to login, if your password reset procedure is 1FA. Attackers will simply ignore the stronger security to target the weakest link - the 1FA step.
The wiki page on how gemcitabine works is pretty fascinating. The drug as administered doesn't kill cells. It gets modified by enzymes in the cell into a form which interferes with DNA replication and blocks DNA repair. That's what kills the cell.
Presumably the bacteria are just metabolizing it before these enzymes can convert it into its toxic form (or lack one of the needed enzymes). So no, not a superbug which digests a toxic material.
I discovered this when I tried switching from a mouse to an Xbox controller for gaming. With the mouse, you change its position and the direction of the camera changes appropriately. Once you've learned how much mouse motion corresponds to how much angular rotation, you can instantly move the camera from one one direction to another simply by moving the mouse to the appropriate spot on the desk. If you want to rapidly move the camera back and forth between two set directions, it's trivial because you're just moving the mouse between two fixed positions on the desk. You could do it blindfolded.
Not so for the thumbsticks on the XBox-style controller. It doesn't control the direction the camera is pointed. It controls the rate of change of the direction the camera is pointed. You have to push the thumbstick in the direction you want the camera to move, wait for the camera to almost get there, ease up on the thumbstick so the slew rate slows down, ease up some more, ease up some more, then let go when the camera is finally pointed in the desired direction. If you want to rapidly move the camera back and forth between two set directions, it's a lot of work each time, and you need to be watching the camera view to do it.
I emailed one of the developers working on drivers to allow you to use the controller in games which didn't support it, and learned why. The thumbsticks only have 256x256 resolution. That is, there are only 256 discrete measurable directions the stick can be pointing in each axis. This isn't enough resolution for precise aiming, so they have to use the gimpy slew-rate aiming.
For a camera limited to just one degree of motion like it sounds like this one is, you want to use a paddle controller. It's just a potentiometer. You rotate it and the camera rotates along with it. It too has a 1:1 correspondence between direction of the controller and direction of the camera. So you could rapidly move the camera back and forth between two set directions, blindfolded. In particular, if you put a raised ridge on the paddle wheel, the operator can know which direction the camera is pointed by feel, instead of having to read a numerical bearing readout. This is much more intuitive (the operator basically won't need any training) and less prone to error.
We know exactly what will happen. The problem isn't the tube losing its pressure seal. Air is compressible, so it makes a great shock absorber (in fact that is exactly how some shock absorbers are designed - with a gas at one end of a cylinder being pressurized).
The problem is trains moving at high speed tends to do bad things when they hit a stationary solid object. The Eschede derailment probably would've only had a dozen or so fatalities due to losing a wheel at 200 kph. In fact, the wheel failure was in the first car, but the engine and first four cars survived relatively intact, scraping the bridge supports but coasting to a stop or derailing and hitting some trees (the guy sitting above the wheel which failed survived despite being out of seat showing the wheel to the conductor when the accident happened). The bridge collapsed onto the 5th car, causing the rapid deceleration of all subsequent cars. That's where all the fatalities occurred. It was just bad luck the train happened to be passing underneath a bridge just as the accident occurred.
Now, consider that with a hyperloop train, the cars will be traveling at speed a few inches from the stationary wall for the entire length of the track. It's not an air leak you need to worry about. It's an IED-type device placed on the side of the track wall, designed to blow it inwards just before the train arrives. At 700 MPH the explosive only needs enough energy to blow enough of the wall inwards to destroy the first train car causing it to block the subsequent cars. The kinetic energy of the train itself will then be more than sufficient to destroy it. When US Air 427 hit the ground at just 300 MPH, its kinetic energy was enough to shred all the metal into pieces smaller than a sheet of paper. United 93 hit the ground at 563 MPH, and its kinetic energy fragmented the plane into such small pieces that conspiracy theorists (who can't seem to grasp the notion that solid metal will fragment when presented with no other means of shedding kinetic energy) have gone nuts with theories that no plane actually crashed there.
A hyperloop train is going to have more than 4x the kinetic energy per unit mass of US Air 427, 1.5x that of United 93. If one strikes the wall and crashes, it kinetic energy is literally going to turn it (and its occupants) into confetti.
What makes it more dangerous than a plane is that planes fly miles away from the nearest solid object when they're at top speed (mid-air collisions excepted). Even systems designed to cause a deliberate collision (surface to air missiles) have a high failure rate. OTOH Hyperloop is going to be traveling a few inches from the nearest stationary object the entire length of its trip. So you're now faced with the prospect of protecting the entire length of track from vandalism or terrorism.
Yes it makes someone stealing your identity harder. But it makes the proliferation of fake IDs trivial. If there's no central authority, what's to stop someone from flooding the decentralized database with a bunch of fake birth IDs every day. Then in the future if you need a fake birth certificate, you can just pay the guy and he'll send you the private key that corresponds to a birth ID blockchain on the date of your choosing with the proper gender and blood type with a pre-generated name. From TFA:
A rogue employee with access to the system can generate a few extra IDs every day and pocket them. Then sell them on the black market in the future. This is currently prevented by being able to compare a birth certificate against a centralized database at the listed hospital of birth (i.e. the two or more centralized databases at different authorities must agree on the information). But if you're going to eliminate the authoritative centralized databases and rely solely on the blockchain as proof of birth ID...
If you read TFA (I know, I know) they controlled for this. Kids were randomly divided into 3 groups - one was praised for being smart, one praised for behavior, one not praised. The group praised for being smart had a higher incidence of cheating. So the cause and effect is correct.
Summary then does a 180 by linking to a study which speculates praise for being smart reduces motivation to learn. That has cause and effect reversed in my experience. I breezed through high school with little effort, but college actually challenged me so I had a hard time. The study skills most kids had developed in high school to learn stuff which challenged them, I had to develop while in college. So it's not that praise for being smart reduced my motivation to learn. It's that being smart meant I (initially) sucked at learning stuff I found challenging.
The original TFA speculates that praising kids for being smart puts them under the pressure of raised expectations. And the kids do whatever they can to meet those expectations - including cheating.
You have a presence in social media even if you've never created an account. You know those little 'f' icons you see on websites which link back to Facebook? They're not a link. They're a script which sets a cookie or examines your cookies to uniquely identify your computer. When you visit slashdot, that 'f' icon in the upper right tells Facebook that user #51853601342 has visited slashdot. And they add it to their database with all other sites user #51853601342.
Then one day a friend of yours sends you a Facebook invite via email. You happen to click on it to delete it (instead of doing a select-delete) which causes it to load in your browser, and now Facebook knows that user #51853601342 is yourname@gmail.com. They start cross-referencing your name with comments, other friend requests, public documents, etc. And now Facebook knows who you are, where you live, who your family is, who your friends are, what you look like (thanks to photo face ID), where you work, who your co-workers are, and how much you make. All without you having an account.
It's not enough to avoid creating social media accounts. You also have to run script blockers specifically targeting these tracking scripts (e.g. Ghostery) and/or browse in incognito mode at all times.
That's kinda the point though. As long as your TV cable company has a monopoly on high-speed Internet in your area, if they start losing money from too many people canceling cable TV, they will just raise the price of cable Internet to compensate. And the cord cutters won't be reaping huge savings.
The price the market will bear for a product, and the cost to supply that product can be very different. Market price falls somewhere in between these two. Competition is what pushes the price close to the cost to supply. Without competition the price gets pushed close to what the market will bear, which is what the cable companies have gotten fat off of for decades thanks to local governments granting them legalized monopolies.
Switching from cable TV to cable internet doesn't change the price the market will bear. So the price won't change as long as these monopolies persist. You're just jumping from the low end of the see-saw to the high end. It raises you up (get more for paying less) only until enough people do it, and the see-saw (cable company) changes the pricing to tip things back to their previous profit margin.
At their core, law and medicine are regurgitation professions. You learn a bunch of facts (a helluva lot of facts), so you can sift through them in your mind, find the ones most relevant to a specific case or patient, and regurgitate them. The licenses needed to practice in those professions are certifications that you've learned enough of those facts to professionally advise people.
While all professions require some degree of regurgitation, they differ in how important it is. At the opposite end of the spectrum are creative professions, which are almost entirely based on skill and talent. It's impossible to certify these professions because often times the talent is in the eye of the beholder (most slashdotters consider most TV programming to be banal, yet it's still incredibly successful).
Working with computers falls somewhere in the middle. Programming falls more towards the creative end, IT more towards the regurgitation end (which is why there are a lot of certifications for IT). But there's a significant amount of crossover (programmers still need to learn algorithms, IT still needs to come up with creative systems or configurations of equipment).
Back during the tech bubble, when programmers with CS degrees were very hard to come by, I was talking with the CEO of a financial analysis company while we walked our dogs in the park. He hired CS grads when he could. His second choice was other STEM graduates (his company made financial simulations, which were very similar to the physical simulations physics majors and engineers were experienced with).
His third choice was a bit surprising - music majors. Music is fundamentally based on mathematical patterns, which apparently makes music majors on average better than the general population at programming and coming up with simulations which try to detect subtle mathematical patterns in finance.
Let's not jump to conclusions. It's still unclear whether or not Mauldin had anything to do with the company being hacked, or even if she wasn't qualified (people are assuming she isn't because of her major). My bet is still on management deciding that money to hire competent IT security professionals was a waste because they'd never suffered a major hack. Some of the other things Equifax did that have come to light (using a timestamp as a PIN, storing passwords as plaintext instead of as hashes, and crucially - not patching their Apache servers immediately after the explain became public and then was fixed) point to the same flawed management decisions when it came to hiring programmers and admins.
It's like that stupid idea which comes up every now and then to harvest energy from cars driving over roads. Yes it'll work, but you're not tapping some cheap form of energy. You're stealing energy from all the cars which drive on the road, and you're doing it very inefficiently because of all the mechanical losses involved in transferring energy from the car, to the road, to your energy capture device. The cumulative cost to all the cars (slight decrease in MPG) plus the cost of re-engineering the road to incorporate the energy capture device is several times higher than what it would've cost you to just build a power plant and generate the energy directly.
That's really what the economic viability of this sort of thing hinges on: How energy-efficient is it to perform this sort of activity distributed over millions of people's computers, vs. doing the activity on your own dedicated equipment? In nearly all cases, it's more efficient to do it on dedicated equipment, and this idea is a non-starter.
Because of the relative inefficiency of Javascript, I can't really think of any activities where distributing it to browsers via ads is really cost- or energy-efficient. Maybe Amazon EC2 might find it useful if there were a sudden temporary spike in CPU demand which exceeded their servers' capability to deliver, and they needed to temporarily find some extra CPU cycles to take the additional load. But in every other case I can think of, it's cheaper to just buy your own computers or to rent time on them directly. The only way selling CPU time via browser ads is competitive is if you're actually stealing CPU time - selling the CPU time for cheaper than the extra electricity cost to the person running the browser. And at that point you've crossed the line from being a legitimate ad to being malware.
I am all for competition and market forces in private enterprise. But pitting state or city governments against each other to see who will give you the biggest tax break is just wrong IMHO. The entire reason people create a government is because they want to be treated fairly. Giving tax breaks only to specific individuals or companies defeats government's reason for existing
If you want to create new economic activity by instituting a tax break, give it to all companies, not just one specific company. Government should be in the business of improving society overall. Not in the business of favoring companies and individuals who can leverage what they can offer society into tax breaks. Cities and states should compete with each other to attract business on the basis of offering the lowest overall tax rate, not compete by giving tax breaks to only a select few.
You're assuming that the economics of this type of deal is zero sum. It's not. It's positive sum. When a new factory is built (and there is market demand for what the factory is producing), everyone wins - the factory, its employees, the surrounding community and government, and the customers of the products the factory produces. Yes someone is probably getting the better end of the deal, but it's not as important if everyone comes out a winner. You are still better off for doing it, than not doing it. (The sports stadium deals where the local government pays to build the stadium are an exception, since it involves actual cash expenditures, not giving up increases in future tax proceeds. The expenditures means the accounting balance for the deal can shift into the red.)
My gripe is with how this type of deal destroys a level playing field between Foxconn and its competitors who don't get the same tax break. You're preemptively shutting out a potential future factory which could have generated even greater benefits than the Foxconn factory. The government shouldn't be placing bets on horses - that's the role of private enterprise (where someone who picks a loser bears the cost of a bad choice themselves, not forces all of society to pay for it).
Alternative interpretations are not as appealing as one would first think.
Given these three choices, I think limiting the Constitution to U.S. territories is the safest.
They should try self-reflecting on why almost every major news organization and poll showed Clinton winning before the election. Maybe the problem isn't that a handful of websites which favored Trump somehow skewed the election. Maybe it's that the media's expectation for the election was skewed from reality.
The one poll which called the election correctly noticed that Trump supporters were less likely to admit to pollsters that they were going to vote for Trump. When they corrected for this, surprise surprise, their poll showed that Trump would win. That's what happens when you engage in campaigns to shame people with certain political beliefs - they disappear from public view, but still show up in elections (thanks to secret ballots).
Real change comes about from convincing people that your way is right. Not from shaming, ridiculing, and attacking those who disagree with you. Unfortunately, the media and to a lesser extent the Democratic party seems to be trying its hardest not to learn this lesson.
This would all be a lot simpler (and friendlier to the consumer) if the studios offered upgrade pricing like the software industry does. So you could buy the HD version of the movie today. But in the future when you had a 4k projector which fills the entire living room wall and 1080p looks like crap, you could simply pay the price differential between HD and 4k version to upgrade. After all, you've already paid for a license for the movie. All you want to buy are the extra pixels, not an entirely new movie license.
But the movie (and record) industry got used to scamming people by selling them multiple licenses of the same thing during all the format shifts the last couple decades. And they're intent on milking that cow for as long as they can.
I agree most people can't tell the difference between resolutions. (Quiz: Two of the major TV networks in the US use 720p cameras, two use 1080i cameras. Based on the TV you've watched, name each network's resolution.)
However, I also have a projector which throws a 150 inch image (almost floor to ceiling). 1080p looks like crap when enlarged to that size, and I'm anxiously waiting for 4k projectors to come down in price so I can replace my 1080p projector. (Quiz answer: ABC and Fox are 720p, NBC and CBS are 1080i. And if you're a sports aficionado, ESPN is also 720p.)
As the Sci-Fi channel (now SyFy) learned, if you name your company after a word which already has common usage in the field in which you're doing business, you cannot trademark your own company's name. Which causes huge legal and marketing problems.
And the energy does come from the computation. If you were already going to do the computation anyway, then the computation uses energy from the wall socket and produces waste heat. This idea is just finding a use for that waste heat.
Heat pumps are only effective when there's some heat in the outside air to pump. Typically, that's down to around 10 C. Below that, the heat is so scarce it takes more energy to pump it than to generate new heat. If the outside temp drops below that, the typical window-mounted heat pump shuts off and turns on an electric heating element (a resistive heater).
This is why the more ambitious heat pump installations use a geothermal heat sink - loops of cooling (heating) pipes buried underground. The ground stays relatively close to 13 C year-round, providing an good, stable sink to pump heat out of in winter, into in summer.
Anyhow, the PC heater is just finding a use of heat that would be generated anyway. So it's basically free heat.
Space heaters don't run constantly. They heat up until they (or the room) hits a certain temperature, then shut off. So a typically 1500 W heater (designed not to trip a circuit breaker providing 15A 11V = 1650 W in the U.S.). This electric company estimates a 1500 W heater will use 274 kWh in a month. 1 month is 730 hours, so that's just 375 W of consumption on average. In other words, they estimate it'll be turned on 25% of the time.
So a 95 Watt CPU (125 Watt system) = 500 Watt heater. Just 3 of them will equal a typical 1500 W space heater.
Heat is heat. It's the entropic end-state of energy, so the form of heating doesn't really matter. 100 W of heating is 100 W of heating.
Radiant heat feels better because it triggers the temperature sensing nerves in your skin, but it doesn't really warm you up any more than a convection heater. If you rely on a small radiant heater, you're gonna have to put on a sweater or wrap a blanket around yourself. Because your air temp is going to be a heckuva lot colder than a larger convection (air) heater, and you'll need the insulation to slow down the rate at which your body is losing heat to the air.
The bigger problem with this idea is the relative thermal inefficiency of generating heat from electricity. See, although an electric heater is 100% efficient at converting electricity into heat, its energy source is not. With a localized heating source (oil or gas heat), the process of burning the fuel converts almost 100% of its energy into heat (a little is "lost" as light or vented to the atmosphere to remove exhaust gases). The same is true when generating electricity, except the heat is generated at the plant, not in your house. So if the power generation plant's efficiency is 40% (typical coal plant), then 60% of the energy in the coal becomes waste heat at the plant, and only 40% of the energy becomes electricity to heat your house.
This is why a gas or oil or even a wood heater in your home is preferable to electric heat. And why power plants in cold climates often have co-generation plants which use that waste heat to send steam through pipes for industrial heating. Of course, this is France, which gets about 3/4 of its electricity from nuclear. So they probably don't care about efficiency, just cost.
I used to manage an email server and mailing list back in the 1990s. The MBOX format most text-based email programs used for storing mail uses two carriage returns and a "From " at the beginning of the next line as the deliminator for a new mail. That's it.
Occasionally people would send an email to the list which randomly had two blank lines followed by "From " as the first word of the next line. If your (text-based) email client wasn't smart to look at subsequent lines to determine that the person had just randomly typed it in the body of the email rather than the actual start of a new email, it would display it as if it was a new email message.
One day someone sent an email to the mailing list which deliberately abused this. The body was crafted so that the "From" in it and subsequent text was formatted like it was a real, separate email. And the people whose clients interpreted it as a new email got duped into thinking the mail list admin had banned them from the mailing list for inappropriate remarks. The perp was just playing a joke of course, but I shudder to think what modern spammers and phishers would do with that capability.
The professional critics' reviews tell me if the movie is well made, has an intriguing plot, good story development, etc.
The public's reviews tell me if the movie is enjoyable to watch.
Best case, both sets of reviews will be high. But when they diverge, the above is usually what I find.
The bigger problem is ratings inflation. I tend to rate things pretty neutrally. i.e. On a 1-10 scale, the average movie score I give will be a 5.5 (actually it'll be closer to 7 because I use review sites to avoid the really bad movies). But there are lots of people whose average score is a 9 and the lowest score they've ever given a movie is a 7. (There are also people whose average score is a 3, but there tend to be fewer of them.) Review sites try to compensate for this by normalizing scores (basically grade the movie on a curve). But for anonymous reviews this means that my "Ok" rating of 7 gets turned into a "poor" rating. To really normalize properly, you have to give up anonymity and normalize each reviewer's ratings before compiling the scores for all reviews. (Amazon has the same problem. The average review score for all products is about 4.5. So having 4-star reviews doesn't necessarily mean the product is good.)
You do realize Apple doesn't actually make any of this stuff, right? They buy them from the same suppliers who make it for Android phones. They get their flash memory from Toshiba and Samsung (the Samsung memory is slightly faster). RAM is from SK Hynix. They get their LED screens from LG, and will get their OLED screens from Samsung. Their camera is sourced from Sony. The cellular and wireless chipsets are from Qualcomm. The much-hyped headphone jack-less audio is by Cirrus Logic. Same with virtually every component that goes into the iPhone.
The only things which are Apple's are the CPU (which they designed, although they use third party fabs to manufacture it - Samsung and TSMC), the fingerprint scanner (they bought the company which makes them back around 2012), and the software.
So the Apple fans who tell themselves that "Apple makes it best" are deluding themselves as a way to rationalize paying an exorbitant price for the same components which go into Android phones.
Or have people who uploaded those photos to Yelp also uploaded them to Google? I've noticed Google Maps has been pretty aggressive lately asking for photos of places I've visited. Considering Yelp doesn't pay their contributors, it's reasonable to assume contributors are very altruistic people who will honor Google's request and upload their photos there too.