Oh, I'm sure they would. One of the staple comparative benchmarks of navigation tools is to see which one can come up with the shortest/fastest route from A to B, so it's definitely a selling point for them and thus will be a focus of development. However, all the other systems would be doing the same thing - trying to find the most efficient route - and if we're not already at the point where the majority of drivers are using one of these tools then we're surely pretty close to it, and those that remain non-users will most likely either stay put or stick to limited number of well known routes that should be accounted for in the traffic flow monitoring.
It's really a case of diminishing returns; at any given moment in time there has to be one route that should be the most optimal way to get from the current location to a destination, which will be a dynamic route based on changing conditions along the route so may change as you progress towards the destination. Sooner or later all the tools are going to get so close to that ideal that there's not really going to be much to differentiate them, unless one doesn't have information on a new road or hasn't yet accounted for some recent disruptive event. So, getting back to the original problem, even if you've got a number of alternative options that slowly fill up as drivers are routed around the problem, there's going to come a point where you can't physically drive the alternative route faster than just sticking it out on the normal route. Once you get to the point where it's all happening in realtime, you're potentially going to find that sweetspot almost by default; if enough traffic takes the alternative routes, then the original route will recover, reducing the amount of traffic being diverted to the alternative routes, and so on.
I've observed the same thing; people will use their own local knowledge and lack of trust in the system to second guess the satnav all the time (yes, I do it too), and statistically that's got to pay off at least some of the time, but it's often impossible to work out which was the correct call (unless you overtake a vehicle that took an alternate route later on, in which case you know you'd made a poor call) to know how good your second guess was. There's also the psychological angle of it being preferable to at least be in motion instead of stuck stop-start traffic, even when the stop-start flow might work out faster overall.
The trick to efficient routing around traffic problems seems to be to not just blindly place your trust in the navigational app but take the time to understand its limitations due to the inherent latency between traffic flow data being captured and incorporated into the guidance. Quite how fast that happens depends on the available sensors and number of cars that are reporting in, and in my experience that can be anything from a few minutes on smart roads to (effectively) never on quieter rural routes. Monitoring the local traffic reports to get an idea of when an issue might be resolved isn't to be downplayed either; if your satnav is telling you to take the off-ramp and the radio is telling you that the incident half a mile up the highway has now been cleared then you're almost certainly going to be much better off staying put and waiting it out.
It's not "selfish" in the context of human attitudes towards achieving a solution to a problem, it's in the context of game theory - the alternative would be to choose the option that leads to the greater good, even if it means you personally benefit less than the alternative option. Assuming that the paper is correct, then the problem we have at the moment is that the software tools need to evolve to a point where they can dynamically determine the optimum number of cars to direct off a congested highway to achieve the most efficient overall flow. I guess the ideal would be that the navigational systems work in real-time to direct enough traffic off the highway that it increases congestion on the sideroads to the point that all possible routes would take the same amount of time, although that would almost certainly be NP complete, given the number of vehicles and unique destinations to account for.
I didn't say the users were abused, I said their profile data was, although there's a pretty good case for both since what they do with that data is to allow the users to be profiled, filtered, and the more susceptible recipients targetted with information (often fake, or at least misleading) that is designed to push buttons and sway opinions. It also seems highly likely that Cambridge Analytics may have gone a bit further than just accessing the information that Facebook made available to them and also deployed some grey/black hat techniques to acquire private profile data as well, in which case if Facebook really wants to make a statement they should probably be thinking about a legal action in the form of a civil suit and at least make a pretense of standing up for their social media platform users (as distinct from their analytics data users). All in all, what Cambridge Analytics does is exactly the same kind of thing that many people are up in arms about when the Russians (or whoever) do it, in otherwords - look at the bigger picture, not the specific circumstances of the one event.
That's something Facebook seems to agree with since they've suspended Cambridge Analytics' account, which speaks volumes given how little value they place on their social platform user's privacy in general, although perhaps Facebook just didn't get the extra payment necessary for that level of abuse as others have postulated and this is retaliation for that. Between their profiling of users and subsequent trolling of them with information designed to incite a given opinion (regardless of what political position it is, they were also involved in the Pro-Brexit campaign and a few other political elections around the world) and potential borderline/actual hacking of Facebook, I'd say that qualifies them as a bad actor, although it's a matter of opinion as to just how "bad" you might think they are. There are certainly far worse players out there.
No, it's not a breach, Facebook is correct on that point. The real issue here, and one that Facebook seems to be pulling off successfully judging by some of the replies so far, is that Facebook's response to 50m user profiles being harvested and abused is to turn it into a discussion about semantics through misdirection. That's *exactly* what Facebook wants the discussion to be on, because it puts them in a favourable light, rather than the real point of TFS, which is that their business model is not only based almost entirely on sharing user data with third parties, but also has no controls or policies in place to effectively govern what happens when they get a bad actor like Cambridge Analytics in the mix.
tl;dr: it's not just about "All your data belong to Facebook (and the rest)", it's that they'll freely share that data with third parties and don't give a fuck what happens when someone abuses their access to it.
Depends on what level you want to get down to as there's an awful lot of genetic variation within the main ethnic groups we general use - asian, caucasian, etc., and even within individual countries with a large geographic areas like, China, India, Russia, and so on. I don't think it's even remotely possible to have a single "archetype" human, or even two or three of them, and will remain that way unless we somehow become one massive cosmopolitan community and massively blend the various major visual traits together, and even if were to achieve some Star Trek Federation-style utopia tomorrow, that's still going to take many, many, generations to achieve. Until then (e.g. "never", given human nature), I think the best option would probably be a colour image showing a decent cross-section of humanity to indicate the degree of variation in ethnic and physical makeup would probably be the best way to go, possibly alongside some kind of generic androgenous human outline and a physical constant based ruler to help establish the scale.
It's not really a kill-switch in the sense of turning the service off, so much as a temporary reset button. What it does is to send a command to memcached to drop the contents of its cache, at which point the spoofed packets of the DDoS will not actually stop outright but be considerably be reduced in size and thus reduce the effective amplification factor of the attack. Unfortunately, the server would then immediately start to repopulate its cache and the amplification factor would gradually recover back to its former levels, something that could happen quite quickly for some cached databases if the underlying query rate is high enough. To effectively shutdown a server, you'd need to keep sending the flush command at regular intervals - in effect launching a DoS at the server to prevent it launching a DoS at another server, so sinking to the same level as those trying to launch the DDoS.
Similar for me (three years for a BSc Hons.), but it does also depend on the nature of the course. Financial, legal, and medical degrees tend to be much longer than arts, humanities, or science degrees for instance - up to seven years for a bachelors in some cases. Four years at college level for your graduation certificate does seem excessive, but not unreasonable if it's one of the new generation of apprenticeship style courses that includes a mix of work placements and formal classroom based education, especially if the subject matter is involved enough.
It's not about the duration of the education, it's about the level of the subject matter. Highschool Graduation < College Degree < Bachelors < Masters < Doctorate. Think of it as the difference between going on to a college or vocational course after highschool rather than to a university for a graduate programme.
Pretty much my view on how to parse this. There are two ways to interpret the data, both of which are potentially correct until more data becomes available on the capabilities of AI in the workplace:
Many Americans are deluding themselves about the level of the threat AI poses to their jobs.
Most Americans are correct in assuming that AI - at least for now - only threatens a small proportion of American jobs.
I'm more inclined to the latter view (with that caveat about "at least for now") because many jobs do entail a lot of corner cases that might be problematic - if not actually unsafe - to trust to an AI, at least at the current state of AI development. More likely, I think, there will be an interim stage with a progressively smaller number of humans overseeing the AIs and stepping in whenever they encounter one of those corner cases, and fairly obviously some professions seeing AIs supplanting humans much sooner and/or faster than others. Basically, the same model as with automation of previously menial labour, only this time around it's going to be for more cerebral tasks.
There's also the fact that a lot of people can't do math, or can't budget, in the face of aggressive marketing and peer pressure surrounding their posistion as a supposed status symbol to factor into the equation. I mean, does anyone *really* care what phone you have as long as it works? I suspect a lot of users of high-end phones never actually own them, they just lease them from their carrier at a monthly rate that ultimately ends up costing them a lot more than it would to just buy the phone outright - especially if they wait out the early adopter phase. The markup for getting some of the top end phones on a contract is just insane, and what happens at the end of that period? Oh, yes, you get to choose between continuing to pay through the nose for what is now an "outdated" phone, or upgrade to the latest model and start the cycle again. It doesn't take Admiral Ackbar to figure out what that scenario is, yet countless people fall into it over and over again.
I didn't say the options where there now; I posited that they might be when the development of the printer technology is ready for them - crawl first, then walk, then run. Currently we're crawling, but walking - using the printer on a Lunar or Martian construction - might not be so far off, especially if the construction process is handled by robots, which seems a logical first step for the initial setup of a potential inhabited base before the humans might arrive.
Persevered for centuries. It literally took centuries for our technology to get where it is.
For some things, sure, but the overall pace of progress pretty much across the board has sped up immensely over the last hundred years or so. Take nautical ships; centuries of limited change based around wooden keels and hulls, then massive progress since the advent of metal hulls and steam power, or aerospace; decades to go from puddle hopping flights to routine intercontinental and tentative spaceflight, then a similar timescale to get from there to where we are now with probes on the fringes of the solar system, roaming around the surface of other planets, and people actually living in space (only LEO, but still). Perhaps it will take a few centuries to get there but, quite frankly, given the huge potential commerical payoff and drive of companies like SpaceX, I wouldn't be at all surprised to see at least some of this kind of stuff bearing fruit somewhat quicker than that. Even optimistically I do think we're still talking decades though, but I'd say that the walking phase (3D printing for construction on the Moon or Mars) might potentially be achievable within the lifespan of people alive today.
Or they're just working on one part of the overall problem - the printer. How it gets supplied with raw material is another issue for someone else to solve, but there are multiple options there. You could ship the stuff direct from Earth (expensive but doable - crawling), mine it in place (potentially viable for the moon, Mars, or maybe an asteroid - walking), or mine it in space and ship it to the printer (probably a long way off but already being explored - running). As with the printer, the problem of supplying the raw material is a matter of one step at a time. Yes, it'll take a lot of time and expense before we get to the running phase, let alone see the fruits of it, but you could say that about a lot of things we now take for granted because someone persevered with their part of the problem and relied on others to sort out the rest.
There's a lot that can be done by launching components that are then bolted to gether in orbit, or having flat-pack components that are expanded in situ - we built the ISS that way, afterall. I guess what they're trying to get across here is what we might be able to accomplish if the current fairing limitation were to be removed - applications that might *require* a solid 100m long truss rather than an oversized shower curtain rail, for instance. The pictures of Musk's Tesla in the Falcon 9 Heavy showed a lot of room around the car, but it's still a pretty tight space in there compared to some of the current structures we build on Earth as single pieces to achieve the necessary structural strength. Assuming this 3D printer can provide all the necessary strength and durability given suitable input material, perhaps this is how we'll create the components for the kind of space stations and other orbital structures that you typically see in SciFi but would currently be impossible to create, even if we could get that much mass into orbit.
I did acknowledge that rights transferal to enable Twitter to function, and explained why it has to happen. That right is explictly granted to Twitter though, and *should* cover any re-tweets, but it wouldn't cover anyone who copied the image and reposted it to another platform, such as Facebook, SnapChat, or whatever. It absolutely does not transfer copyright away from the owner but is doing exactly what copyright is intended to do; granting a specific set of usage rights - for reproduction within their own platform. There's quite a big difference there.
As I read it, what's muddying the waters here isn't the liability cover / reusage rights provided for by Twitter's ToS, it's that the image hasn't *actually* been copied and pasted to another service; they've just embedded part of Twitter's site within their own - the image being displayed is still being served by Twitter's servers. In my view, that's a significant factor as it has a huge implication of whether the publication of the image has violated the original usage license implicitly granted by the posting of the image to Twitter and is something that the ruling has failed to adequately consider. Twitter has done nothing wrong here; they are entirely within their granted rights to reuse the image - the issue is whether those other services have any rights to reuse the image *specifically* by embedding part of Twitter's site rather than reusing the image itself. Ironically, the biggest loser here is probably Twitter - I suspect those embedded tweets generate quite a bit of traffic towards their site, so I'm fully expecting to see a tweak to their ToS to specifically allow that usage case Real Soon Now.
Disclaimer: IANAL, but I am a photographer so have a pretty good grasp of how copyright applies to my images - or is supposed to at least.
Implications and whether or not current copyright law is fit for purpose aside, legislation under the Berne Convention is pretty clear: unauthorised reproduction that does not comply with any explicitly granted permissions of use is a breach of copyright, and that applies to code under the GPL and Creative Commons (amongst licenses) as well as traditional publications, photographs, audio, video and so on. In the case of a photograph, then the copyright belongs to the photographer in the first case, and their agency/publisher only if they have agreed to transfer the rights - if anyone else re-uses that image without express permission, then they have breached the copyright.
Now, if you want to try and assign responsibility for an image embedded in a tweet (or any other form of online post), it's going to get messy however you slice it. Copyright is all about controling others ability to reproduce the data, and the owner of the copyright putting some data into a public forum like Twitter or whatever does not change that right. Publishing a book puts the text into a public forum, yet you can't arbitrarily decide to OCR it and publish your own version of the book without any fear or legal repercussion, for instance. So, unless there is an explicit granting of a license to retweet an image, by doing so - technically - you are going to be in breach of copyright, not Twitter. At best, Twitter can be accused of helping to facilitiate that infringement and perhaps not taking sufficient steps (quite what those might be, however... asking for an affirmative consent to retweet, blocking all image retweets?) to prevent it, only far too many people turn a blind eye to this because retweets = higher profile = (hopefully!) profit. This is why such services often try to cut through the whole mess by including a clause in their ToS/EULA that grants them - as a platform - the right to reproduce the content at will within the confines of that platform; take an image from Twitter and post it elsewhere, and that is no longer Twitter's problem if the copyright owner decides to sue. And that's before you consider the issue of fair use; a snippet of a larger text is one thing, an image is (usually) the entire work, so where do you draw the line for something in the public interest, just how do you define public interest, and should that be any different to a random picture of whatever?
There's a lot of scary implications for both the operators and users of online services here, whichever way any potential legislative ruling might decide to view this kind of thing. While we tend to prefer things to be nice and clear cut with a clear demarcation line separating black from white, in this instance I think it might actually be better for the majority of people on all sides to just leave things in a grey area and rely on the spirit of the law and likely intent as the decider in any legal action that might be undertaken.
@GP - for the record, I compiled the code for a beta release of Mosaic in early 1993 to supplement (read "replace") Berners Lee's code we'd been using on my university's systems until then. Other than those actually working with Berners-Lee at CERN, I've probably been on the current web about as long as it's possible to be (ignoring all the time I spent on Gopher and other WWW precursors based off Ted Nelson's ideas for Project Xanadu before then - one of which was used for my thesis).
Getting back to the question about when animated ads arrived: 1993. It didn't take long, but that was more a matter of timing, I think - all the necessary pieces came together around the same time. The Web was "born" at CERN in 1990, but even in academia didn't really get much traction for a couple of years and I don't recall much in the way of graphical ads (animated or not) on content available via Gopher except for some basic in-line ASCII art level stuff and some very limited use of GIFs. Lots of spam on Usenet and some IRC channels, of course, but that's purely text and attachments. It wasn't until Mosaic v1.0 in 1993 that the web really entered into the public conciousness and, as the first browser capable of handling in-line graphics, it wasn't really possible to even *have* ads until then (before Mosaic any graphics were viewed in a new, dedicated window), and even then it was limited to locally hosted content with IMG tags embedded in HREFs - there was defintely no ad-revenue model in place; purely attempts to get a click-through. Eternal September was also 1993 though, and the proliferation of ads (still just click-bait) started up pretty soon after, so that would have caught the attention of the marketing types too - there were definitely animated ads on the web before then, but they *really* took off after the unwashed masses at AoL arrived.
There is that assumption, yes, but it also ties into the desire the read the specific content. The scarcer the content is, the more likely it is I'm going to work on getting to see it on the available sites, including potentially temporarily whitelisting some elements of a site with the anti-adblock script. That is an *awfully* rare occurance though, and such content - at least for me - tends to be found on sites that I'm likely to revisit so that too tends to increase the effort I'll put into making it work.
I call bullshit on the "without ads there would be no content" argument though; there was plenty of quality content on the Internet before ads came along, and there's still plenty of quality content out there today on sites that don't use ads or require loading a plethora of third party scripts that aren't as profit driven as the likes of Forbes and the Murdoch family who are pretty much the poster children for anti-adblocking. That ad-supported is the only viable business model is a fallacy peddled mostly by those that want to sell the ads in the first place; other revenue generation models exist, and can work very well, but they do require a little more effort than a copy-pasta of a bunch of ad-scripts and a good understanding of your readership though.
My fourth option is the same as the one I use when I come across a site with an anti-adblock script where the effort to work around it outweighs my desire to read their specific take on the content. See if that content is available elsewhere, with a side-option of probably not returning to that site again if it's a high-enough profile to enter my conciousness or particularly obnoxious about the detection. There are already far too many instance of ad-providers serving up malware to make not blocking ads across the board a remotely sane thing to do from a security perspective, and given the recent abuse of a popular script-hosting site to mine crypto-currency it seems like only whitelisting scripts is a pretty damn good idea too.
Of course, I do have the advantage of knowing what I'm doing with script- and ad-blockers so a whitelisting is a viable option. The average Joe's PC, on the otherhand, is probably going to be spending a lot of its time sending spam, mining crypto currency, or acting a a proxy for random script kiddies to do whatever they want. If Google really wants to make a difference with Chrome's ad-blocker (and AMP, for that matter), they could start right there and insist that all pages and content, including any ads, will still render a usable page without any scripting support. I'll still be blocking the ads though.:)
While there is some massive centralised hydro and thermal power generation in Iceland (they smelt bauxite to aluminium, which uses a lot of power), data centres - even for mining - in Iceland are nothing new; this is just people late to the party catching up with the early adopters. In Iceland's major urban locations, mostly around the Reykjavik/Keflavik area and the handful of other larger towns, the DCs are often connected to the local power grid, but the more serious players in the cottage DC industry they have are actually setting up their own power generation on a convenient thermal outlet or suitable river. Iceland's on-grid electrical supplies may be cheap, but the breakpoint for DIY being even cheaper is also surprisingly low so the risk of Iceland being left with a lot of excess capacity (which could easily be redirected to smelting, or other uses) is probably fairly low.
There's a fairly recently built DC facility like this near Gerdi where the Thorbergssetur "Bookshelf" building is, just up the N1 from Jokulsarlon (the glacial lagoon) towards Hofn, that was put in by an enterprising local farming family, if you're in the area and want to get an idea of what the typical installation looks like. No idea how much, if any, of its capacity is used for mining though.
Most likely reason, given the fixed price? Because they know it's not viable to repair the thing and they're planning to simply replace any faulty units at cost - including any admin/diagnostic/handling fees, etc. - so the price of doing so is pretty much a constant. I'm fully expecting this to set a new low on repairability when iFixit does their inevitable teardown. Ease of repair, or even a reasonable capability to recycle, electronics has taken a back style to looks for a long while now and this is just the next step along that path. Besides, there's still plenty of places to use for landfills for all the faulty electronics that it's not economical to repair, right? Just as long as it's not in *my* backyard, of course.
"Right to Repair" can't come soon enough in my view, but I just don't see much support from Trump's government or the Democrats in the states where Apple, etc. are based, and the lobbying opposing it from the consumer electronics companies is likely to be fierce too. The EU might get something passed, however, at which point it's going to be interesting to see whether Apple et al apply that globally to retain economies of scale, or take the path of EU-specific models that can be repaired.
In some regimes certainly, but in that case they'll have done that anyway - and in all liklihood more severely - if they were able to retrieve unencrypted images of something they took exception to. The aim here isn't to remove all the risk from the profession, it's to reduce it to the point that journalists have at least a fighting chance of using plausible deniability and the implicit threat that the regime's practices will be exposed on the world's stage to enable them to do their jobs with the minimal risk of harm possible. They'll still need to manage their situation and have good OpSec/PerSec, especially if they do have some photographic evidence that a regime would want suppressed, but even so being roughed up and eventually release for lack of actual evidence is still going to preferable to the alternative if it comes to it.
The idea is that you load the public key that will be used to encrypt the images on the camera and you leave the private key back home. This is all fully documented by the camera manufacturer so that if the photographer gets challenged to decrypt the images it's easy to establish that they can't actually do that and there's no point in getting out the rubber hose. Ideally there would also be some options for having unencrypted images on the card and simply hiding the encrypted ones from the image review in the event of a more casual stop and search. If the police (or whoever) want to be a little more thorough, then yes, they can still detain the photographer, delete images/format cards, or destroy cameras/cards, but all that is going to based on no actual evidence and they'll eventually be called on it by the media organizations. Whether they care about that or not is another matter entirely.
Of course, the prudent photographer working in a situation where detention or worse might be a possibility would also provide a copy of the private key to a trusted agent (lawyer, employer, etc.) so that if things really did turn ugly the key could still be provided. That could then be backed up with some specific circumstances upon which to disclose the private key appropriate to the types of images being sought, the country of detention, and the overall severity of the situation.
Sure, but that's with at least some form of court order or warrant, right? Unlike the data on the San Bernadino iPhone there's no way they can claim they don't have access to the data here as they require access to it, so they best users of such devices can hope for is that their vendor of choice will at least put up a token fight and insist on a court order before handing over the data. Apple does have a pretty good track record on that, unless like (for instance) Amazon who doesn't seem to fight *anything* on behalf of their users, let alone go to court for them or the principle of it.
The world isn't black and white: there are also those that are not in the Apple garden, and have no intention of going in there, but see Apple as a useful foil to help keep the likes of Dell, Google, Microsoft, Samsung, etc. at least slightly honest, competetive and innovative. I might not have or want any of Apple's products in my life, but I don't want them to slowly fade slowly back into the relative niche obscurity they had in the late 1990s and early 2000s either. While that's obviously not going to happen any time soon thanks to their huge cash reserves, launching technically lame and overpriced products compared to the competition like this are not exactly going to help.
Slashdot Mirror
Oh, I'm sure they would. One of the staple comparative benchmarks of navigation tools is to see which one can come up with the shortest/fastest route from A to B, so it's definitely a selling point for them and thus will be a focus of development. However, all the other systems would be doing the same thing - trying to find the most efficient route - and if we're not already at the point where the majority of drivers are using one of these tools then we're surely pretty close to it, and those that remain non-users will most likely either stay put or stick to limited number of well known routes that should be accounted for in the traffic flow monitoring.
It's really a case of diminishing returns; at any given moment in time there has to be one route that should be the most optimal way to get from the current location to a destination, which will be a dynamic route based on changing conditions along the route so may change as you progress towards the destination. Sooner or later all the tools are going to get so close to that ideal that there's not really going to be much to differentiate them, unless one doesn't have information on a new road or hasn't yet accounted for some recent disruptive event. So, getting back to the original problem, even if you've got a number of alternative options that slowly fill up as drivers are routed around the problem, there's going to come a point where you can't physically drive the alternative route faster than just sticking it out on the normal route. Once you get to the point where it's all happening in realtime, you're potentially going to find that sweetspot almost by default; if enough traffic takes the alternative routes, then the original route will recover, reducing the amount of traffic being diverted to the alternative routes, and so on.
I've observed the same thing; people will use their own local knowledge and lack of trust in the system to second guess the satnav all the time (yes, I do it too), and statistically that's got to pay off at least some of the time, but it's often impossible to work out which was the correct call (unless you overtake a vehicle that took an alternate route later on, in which case you know you'd made a poor call) to know how good your second guess was. There's also the psychological angle of it being preferable to at least be in motion instead of stuck stop-start traffic, even when the stop-start flow might work out faster overall.
The trick to efficient routing around traffic problems seems to be to not just blindly place your trust in the navigational app but take the time to understand its limitations due to the inherent latency between traffic flow data being captured and incorporated into the guidance. Quite how fast that happens depends on the available sensors and number of cars that are reporting in, and in my experience that can be anything from a few minutes on smart roads to (effectively) never on quieter rural routes. Monitoring the local traffic reports to get an idea of when an issue might be resolved isn't to be downplayed either; if your satnav is telling you to take the off-ramp and the radio is telling you that the incident half a mile up the highway has now been cleared then you're almost certainly going to be much better off staying put and waiting it out.
It's not "selfish" in the context of human attitudes towards achieving a solution to a problem, it's in the context of game theory - the alternative would be to choose the option that leads to the greater good, even if it means you personally benefit less than the alternative option. Assuming that the paper is correct, then the problem we have at the moment is that the software tools need to evolve to a point where they can dynamically determine the optimum number of cars to direct off a congested highway to achieve the most efficient overall flow. I guess the ideal would be that the navigational systems work in real-time to direct enough traffic off the highway that it increases congestion on the sideroads to the point that all possible routes would take the same amount of time, although that would almost certainly be NP complete, given the number of vehicles and unique destinations to account for.
I didn't say the users were abused, I said their profile data was, although there's a pretty good case for both since what they do with that data is to allow the users to be profiled, filtered, and the more susceptible recipients targetted with information (often fake, or at least misleading) that is designed to push buttons and sway opinions. It also seems highly likely that Cambridge Analytics may have gone a bit further than just accessing the information that Facebook made available to them and also deployed some grey/black hat techniques to acquire private profile data as well, in which case if Facebook really wants to make a statement they should probably be thinking about a legal action in the form of a civil suit and at least make a pretense of standing up for their social media platform users (as distinct from their analytics data users). All in all, what Cambridge Analytics does is exactly the same kind of thing that many people are up in arms about when the Russians (or whoever) do it, in otherwords - look at the bigger picture, not the specific circumstances of the one event.
That's something Facebook seems to agree with since they've suspended Cambridge Analytics' account, which speaks volumes given how little value they place on their social platform user's privacy in general, although perhaps Facebook just didn't get the extra payment necessary for that level of abuse as others have postulated and this is retaliation for that. Between their profiling of users and subsequent trolling of them with information designed to incite a given opinion (regardless of what political position it is, they were also involved in the Pro-Brexit campaign and a few other political elections around the world) and potential borderline/actual hacking of Facebook, I'd say that qualifies them as a bad actor, although it's a matter of opinion as to just how "bad" you might think they are. There are certainly far worse players out there.
No, it's not a breach, Facebook is correct on that point. The real issue here, and one that Facebook seems to be pulling off successfully judging by some of the replies so far, is that Facebook's response to 50m user profiles being harvested and abused is to turn it into a discussion about semantics through misdirection. That's *exactly* what Facebook wants the discussion to be on, because it puts them in a favourable light, rather than the real point of TFS, which is that their business model is not only based almost entirely on sharing user data with third parties, but also has no controls or policies in place to effectively govern what happens when they get a bad actor like Cambridge Analytics in the mix.
tl;dr: it's not just about "All your data belong to Facebook (and the rest)", it's that they'll freely share that data with third parties and don't give a fuck what happens when someone abuses their access to it.
Depends on what level you want to get down to as there's an awful lot of genetic variation within the main ethnic groups we general use - asian, caucasian, etc., and even within individual countries with a large geographic areas like, China, India, Russia, and so on. I don't think it's even remotely possible to have a single "archetype" human, or even two or three of them, and will remain that way unless we somehow become one massive cosmopolitan community and massively blend the various major visual traits together, and even if were to achieve some Star Trek Federation-style utopia tomorrow, that's still going to take many, many, generations to achieve. Until then (e.g. "never", given human nature), I think the best option would probably be a colour image showing a decent cross-section of humanity to indicate the degree of variation in ethnic and physical makeup would probably be the best way to go, possibly alongside some kind of generic androgenous human outline and a physical constant based ruler to help establish the scale.
It's not really a kill-switch in the sense of turning the service off, so much as a temporary reset button. What it does is to send a command to memcached to drop the contents of its cache, at which point the spoofed packets of the DDoS will not actually stop outright but be considerably be reduced in size and thus reduce the effective amplification factor of the attack. Unfortunately, the server would then immediately start to repopulate its cache and the amplification factor would gradually recover back to its former levels, something that could happen quite quickly for some cached databases if the underlying query rate is high enough. To effectively shutdown a server, you'd need to keep sending the flush command at regular intervals - in effect launching a DoS at the server to prevent it launching a DoS at another server, so sinking to the same level as those trying to launch the DDoS.
Similar for me (three years for a BSc Hons.), but it does also depend on the nature of the course. Financial, legal, and medical degrees tend to be much longer than arts, humanities, or science degrees for instance - up to seven years for a bachelors in some cases. Four years at college level for your graduation certificate does seem excessive, but not unreasonable if it's one of the new generation of apprenticeship style courses that includes a mix of work placements and formal classroom based education, especially if the subject matter is involved enough.
It's not about the duration of the education, it's about the level of the subject matter. Highschool Graduation < College Degree < Bachelors < Masters < Doctorate. Think of it as the difference between going on to a college or vocational course after highschool rather than to a university for a graduate programme.
I'm more inclined to the latter view (with that caveat about "at least for now") because many jobs do entail a lot of corner cases that might be problematic - if not actually unsafe - to trust to an AI, at least at the current state of AI development. More likely, I think, there will be an interim stage with a progressively smaller number of humans overseeing the AIs and stepping in whenever they encounter one of those corner cases, and fairly obviously some professions seeing AIs supplanting humans much sooner and/or faster than others. Basically, the same model as with automation of previously menial labour, only this time around it's going to be for more cerebral tasks.
There's also the fact that a lot of people can't do math, or can't budget, in the face of aggressive marketing and peer pressure surrounding their posistion as a supposed status symbol to factor into the equation. I mean, does anyone *really* care what phone you have as long as it works? I suspect a lot of users of high-end phones never actually own them, they just lease them from their carrier at a monthly rate that ultimately ends up costing them a lot more than it would to just buy the phone outright - especially if they wait out the early adopter phase. The markup for getting some of the top end phones on a contract is just insane, and what happens at the end of that period? Oh, yes, you get to choose between continuing to pay through the nose for what is now an "outdated" phone, or upgrade to the latest model and start the cycle again. It doesn't take Admiral Ackbar to figure out what that scenario is, yet countless people fall into it over and over again.
For some things, sure, but the overall pace of progress pretty much across the board has sped up immensely over the last hundred years or so. Take nautical ships; centuries of limited change based around wooden keels and hulls, then massive progress since the advent of metal hulls and steam power, or aerospace; decades to go from puddle hopping flights to routine intercontinental and tentative spaceflight, then a similar timescale to get from there to where we are now with probes on the fringes of the solar system, roaming around the surface of other planets, and people actually living in space (only LEO, but still). Perhaps it will take a few centuries to get there but, quite frankly, given the huge potential commerical payoff and drive of companies like SpaceX, I wouldn't be at all surprised to see at least some of this kind of stuff bearing fruit somewhat quicker than that. Even optimistically I do think we're still talking decades though, but I'd say that the walking phase (3D printing for construction on the Moon or Mars) might potentially be achievable within the lifespan of people alive today.
Or they're just working on one part of the overall problem - the printer. How it gets supplied with raw material is another issue for someone else to solve, but there are multiple options there. You could ship the stuff direct from Earth (expensive but doable - crawling), mine it in place (potentially viable for the moon, Mars, or maybe an asteroid - walking), or mine it in space and ship it to the printer (probably a long way off but already being explored - running). As with the printer, the problem of supplying the raw material is a matter of one step at a time. Yes, it'll take a lot of time and expense before we get to the running phase, let alone see the fruits of it, but you could say that about a lot of things we now take for granted because someone persevered with their part of the problem and relied on others to sort out the rest.
There's a lot that can be done by launching components that are then bolted to gether in orbit, or having flat-pack components that are expanded in situ - we built the ISS that way, afterall. I guess what they're trying to get across here is what we might be able to accomplish if the current fairing limitation were to be removed - applications that might *require* a solid 100m long truss rather than an oversized shower curtain rail, for instance. The pictures of Musk's Tesla in the Falcon 9 Heavy showed a lot of room around the car, but it's still a pretty tight space in there compared to some of the current structures we build on Earth as single pieces to achieve the necessary structural strength. Assuming this 3D printer can provide all the necessary strength and durability given suitable input material, perhaps this is how we'll create the components for the kind of space stations and other orbital structures that you typically see in SciFi but would currently be impossible to create, even if we could get that much mass into orbit.
I did acknowledge that rights transferal to enable Twitter to function, and explained why it has to happen. That right is explictly granted to Twitter though, and *should* cover any re-tweets, but it wouldn't cover anyone who copied the image and reposted it to another platform, such as Facebook, SnapChat, or whatever. It absolutely does not transfer copyright away from the owner but is doing exactly what copyright is intended to do; granting a specific set of usage rights - for reproduction within their own platform. There's quite a big difference there.
As I read it, what's muddying the waters here isn't the liability cover / reusage rights provided for by Twitter's ToS, it's that the image hasn't *actually* been copied and pasted to another service; they've just embedded part of Twitter's site within their own - the image being displayed is still being served by Twitter's servers. In my view, that's a significant factor as it has a huge implication of whether the publication of the image has violated the original usage license implicitly granted by the posting of the image to Twitter and is something that the ruling has failed to adequately consider. Twitter has done nothing wrong here; they are entirely within their granted rights to reuse the image - the issue is whether those other services have any rights to reuse the image *specifically* by embedding part of Twitter's site rather than reusing the image itself. Ironically, the biggest loser here is probably Twitter - I suspect those embedded tweets generate quite a bit of traffic towards their site, so I'm fully expecting to see a tweak to their ToS to specifically allow that usage case Real Soon Now.
Disclaimer: IANAL, but I am a photographer so have a pretty good grasp of how copyright applies to my images - or is supposed to at least.
Implications and whether or not current copyright law is fit for purpose aside, legislation under the Berne Convention is pretty clear: unauthorised reproduction that does not comply with any explicitly granted permissions of use is a breach of copyright, and that applies to code under the GPL and Creative Commons (amongst licenses) as well as traditional publications, photographs, audio, video and so on. In the case of a photograph, then the copyright belongs to the photographer in the first case, and their agency/publisher only if they have agreed to transfer the rights - if anyone else re-uses that image without express permission, then they have breached the copyright.
Now, if you want to try and assign responsibility for an image embedded in a tweet (or any other form of online post), it's going to get messy however you slice it. Copyright is all about controling others ability to reproduce the data, and the owner of the copyright putting some data into a public forum like Twitter or whatever does not change that right. Publishing a book puts the text into a public forum, yet you can't arbitrarily decide to OCR it and publish your own version of the book without any fear or legal repercussion, for instance. So, unless there is an explicit granting of a license to retweet an image, by doing so - technically - you are going to be in breach of copyright, not Twitter. At best, Twitter can be accused of helping to facilitiate that infringement and perhaps not taking sufficient steps (quite what those might be, however... asking for an affirmative consent to retweet, blocking all image retweets?) to prevent it, only far too many people turn a blind eye to this because retweets = higher profile = (hopefully!) profit. This is why such services often try to cut through the whole mess by including a clause in their ToS/EULA that grants them - as a platform - the right to reproduce the content at will within the confines of that platform; take an image from Twitter and post it elsewhere, and that is no longer Twitter's problem if the copyright owner decides to sue. And that's before you consider the issue of fair use; a snippet of a larger text is one thing, an image is (usually) the entire work, so where do you draw the line for something in the public interest, just how do you define public interest, and should that be any different to a random picture of whatever?
There's a lot of scary implications for both the operators and users of online services here, whichever way any potential legislative ruling might decide to view this kind of thing. While we tend to prefer things to be nice and clear cut with a clear demarcation line separating black from white, in this instance I think it might actually be better for the majority of people on all sides to just leave things in a grey area and rely on the spirit of the law and likely intent as the decider in any legal action that might be undertaken.
@GP - for the record, I compiled the code for a beta release of Mosaic in early 1993 to supplement (read "replace") Berners Lee's code we'd been using on my university's systems until then. Other than those actually working with Berners-Lee at CERN, I've probably been on the current web about as long as it's possible to be (ignoring all the time I spent on Gopher and other WWW precursors based off Ted Nelson's ideas for Project Xanadu before then - one of which was used for my thesis).
Getting back to the question about when animated ads arrived: 1993. It didn't take long, but that was more a matter of timing, I think - all the necessary pieces came together around the same time. The Web was "born" at CERN in 1990, but even in academia didn't really get much traction for a couple of years and I don't recall much in the way of graphical ads (animated or not) on content available via Gopher except for some basic in-line ASCII art level stuff and some very limited use of GIFs. Lots of spam on Usenet and some IRC channels, of course, but that's purely text and attachments. It wasn't until Mosaic v1.0 in 1993 that the web really entered into the public conciousness and, as the first browser capable of handling in-line graphics, it wasn't really possible to even *have* ads until then (before Mosaic any graphics were viewed in a new, dedicated window), and even then it was limited to locally hosted content with IMG tags embedded in HREFs - there was defintely no ad-revenue model in place; purely attempts to get a click-through. Eternal September was also 1993 though, and the proliferation of ads (still just click-bait) started up pretty soon after, so that would have caught the attention of the marketing types too - there were definitely animated ads on the web before then, but they *really* took off after the unwashed masses at AoL arrived.
There is that assumption, yes, but it also ties into the desire the read the specific content. The scarcer the content is, the more likely it is I'm going to work on getting to see it on the available sites, including potentially temporarily whitelisting some elements of a site with the anti-adblock script. That is an *awfully* rare occurance though, and such content - at least for me - tends to be found on sites that I'm likely to revisit so that too tends to increase the effort I'll put into making it work.
I call bullshit on the "without ads there would be no content" argument though; there was plenty of quality content on the Internet before ads came along, and there's still plenty of quality content out there today on sites that don't use ads or require loading a plethora of third party scripts that aren't as profit driven as the likes of Forbes and the Murdoch family who are pretty much the poster children for anti-adblocking. That ad-supported is the only viable business model is a fallacy peddled mostly by those that want to sell the ads in the first place; other revenue generation models exist, and can work very well, but they do require a little more effort than a copy-pasta of a bunch of ad-scripts and a good understanding of your readership though.
My fourth option is the same as the one I use when I come across a site with an anti-adblock script where the effort to work around it outweighs my desire to read their specific take on the content. See if that content is available elsewhere, with a side-option of probably not returning to that site again if it's a high-enough profile to enter my conciousness or particularly obnoxious about the detection. There are already far too many instance of ad-providers serving up malware to make not blocking ads across the board a remotely sane thing to do from a security perspective, and given the recent abuse of a popular script-hosting site to mine crypto-currency it seems like only whitelisting scripts is a pretty damn good idea too.
:)
Of course, I do have the advantage of knowing what I'm doing with script- and ad-blockers so a whitelisting is a viable option. The average Joe's PC, on the otherhand, is probably going to be spending a lot of its time sending spam, mining crypto currency, or acting a a proxy for random script kiddies to do whatever they want. If Google really wants to make a difference with Chrome's ad-blocker (and AMP, for that matter), they could start right there and insist that all pages and content, including any ads, will still render a usable page without any scripting support. I'll still be blocking the ads though.
While there is some massive centralised hydro and thermal power generation in Iceland (they smelt bauxite to aluminium, which uses a lot of power), data centres - even for mining - in Iceland are nothing new; this is just people late to the party catching up with the early adopters. In Iceland's major urban locations, mostly around the Reykjavik/Keflavik area and the handful of other larger towns, the DCs are often connected to the local power grid, but the more serious players in the cottage DC industry they have are actually setting up their own power generation on a convenient thermal outlet or suitable river. Iceland's on-grid electrical supplies may be cheap, but the breakpoint for DIY being even cheaper is also surprisingly low so the risk of Iceland being left with a lot of excess capacity (which could easily be redirected to smelting, or other uses) is probably fairly low.
There's a fairly recently built DC facility like this near Gerdi where the Thorbergssetur "Bookshelf" building is, just up the N1 from Jokulsarlon (the glacial lagoon) towards Hofn, that was put in by an enterprising local farming family, if you're in the area and want to get an idea of what the typical installation looks like. No idea how much, if any, of its capacity is used for mining though.
Most likely reason, given the fixed price? Because they know it's not viable to repair the thing and they're planning to simply replace any faulty units at cost - including any admin/diagnostic/handling fees, etc. - so the price of doing so is pretty much a constant. I'm fully expecting this to set a new low on repairability when iFixit does their inevitable teardown. Ease of repair, or even a reasonable capability to recycle, electronics has taken a back style to looks for a long while now and this is just the next step along that path. Besides, there's still plenty of places to use for landfills for all the faulty electronics that it's not economical to repair, right? Just as long as it's not in *my* backyard, of course.
"Right to Repair" can't come soon enough in my view, but I just don't see much support from Trump's government or the Democrats in the states where Apple, etc. are based, and the lobbying opposing it from the consumer electronics companies is likely to be fierce too. The EU might get something passed, however, at which point it's going to be interesting to see whether Apple et al apply that globally to retain economies of scale, or take the path of EU-specific models that can be repaired.
In some regimes certainly, but in that case they'll have done that anyway - and in all liklihood more severely - if they were able to retrieve unencrypted images of something they took exception to. The aim here isn't to remove all the risk from the profession, it's to reduce it to the point that journalists have at least a fighting chance of using plausible deniability and the implicit threat that the regime's practices will be exposed on the world's stage to enable them to do their jobs with the minimal risk of harm possible. They'll still need to manage their situation and have good OpSec/PerSec, especially if they do have some photographic evidence that a regime would want suppressed, but even so being roughed up and eventually release for lack of actual evidence is still going to preferable to the alternative if it comes to it.
The idea is that you load the public key that will be used to encrypt the images on the camera and you leave the private key back home. This is all fully documented by the camera manufacturer so that if the photographer gets challenged to decrypt the images it's easy to establish that they can't actually do that and there's no point in getting out the rubber hose. Ideally there would also be some options for having unencrypted images on the card and simply hiding the encrypted ones from the image review in the event of a more casual stop and search. If the police (or whoever) want to be a little more thorough, then yes, they can still detain the photographer, delete images/format cards, or destroy cameras/cards, but all that is going to based on no actual evidence and they'll eventually be called on it by the media organizations. Whether they care about that or not is another matter entirely.
Of course, the prudent photographer working in a situation where detention or worse might be a possibility would also provide a copy of the private key to a trusted agent (lawyer, employer, etc.) so that if things really did turn ugly the key could still be provided. That could then be backed up with some specific circumstances upon which to disclose the private key appropriate to the types of images being sought, the country of detention, and the overall severity of the situation.
Sure, but that's with at least some form of court order or warrant, right? Unlike the data on the San Bernadino iPhone there's no way they can claim they don't have access to the data here as they require access to it, so they best users of such devices can hope for is that their vendor of choice will at least put up a token fight and insist on a court order before handing over the data. Apple does have a pretty good track record on that, unless like (for instance) Amazon who doesn't seem to fight *anything* on behalf of their users, let alone go to court for them or the principle of it.
The world isn't black and white: there are also those that are not in the Apple garden, and have no intention of going in there, but see Apple as a useful foil to help keep the likes of Dell, Google, Microsoft, Samsung, etc. at least slightly honest, competetive and innovative. I might not have or want any of Apple's products in my life, but I don't want them to slowly fade slowly back into the relative niche obscurity they had in the late 1990s and early 2000s either. While that's obviously not going to happen any time soon thanks to their huge cash reserves, launching technically lame and overpriced products compared to the competition like this are not exactly going to help.