We legally enforce censorship all the time post a threat to kill the president of the United States and see how long until you get a visit from Uncle Sam and his G-Men.
Do they actually force you to remove the threat? It was my understanding that they don't. The Secret Service just wants to know if you're a credible threat or not. Your speech isn't censored. It just has consequences. And if you're not actually a credible threat, the consequences amount to momentary inconvenience. And your name on a list. I assume it's difficult to get a Top Secret clearance if you've posted a threat to kill the president, though I don't actually know. It just seems like something the DoD would frown upon.
This is the part conspiracy theorists never grasp. There isn't any benefit to scientists for promoting a massive global conspiracy, but the benefits from puncturing such a conspiracy would be enormous.
Every adjunct professor in the world is waiting for an opportunity like this.
No they're not. Guess who sits on the grant committees? Old scientists, with a vested interest in defending the status quo. And defend it they will, by denying grants to any such upstart adjunct professor. There's a very good reason why Max Planck said science advances one funeral at a time, and it's far more relevant now than it ever was when he first wrote words to that affect in 1906. Adjunct professors don't have tenure. They do not rock the research boat. Not ever. Not anymore, if indeed they ever did.
Mathematics is the only academic pursuit immune from this problem. Math is math. Once a proof has been checked over thoroughly, it's unassailable. This is because math is not science. Math is an abstraction, with perfect internal consistency, because it is designed that way. Science deals with reality, which is not nearly as convenient.
It will take at least several hundred years for thawed out permafrost to become suitable for agriculture.
It will take one season, growing some non-food crop to establish soil bacteria, and then one afternoon spreading petroleum-based fertilizer the following season.
You didn't think natural soil was relevant to modern agriculture, did you?
No, but the only way around the problem is to develop tricky automated mining equipment and make all that stuff once you get there. I work with mining equipment. Maintenance intervals (oils/filters/etc) are every 50 hours of operation, machine-stopping breakdowns occur every few hundred hours, large component changeout (pumps, hydraulic cylinders, etc) is 4000 hours. 4000 hours is a year of operation at a 50% duty cycle. So you're going to ship all this stuff to Mars, and then expect it to run, continuously digging stuff up and crushing it and heating it and so on and so forth, for a couple of years? In a cold, dusty, zero-maintenance environment?
That sounds like dumptrucks and the like. Air-breathing diesel-powered vehicles, which are obviously irrelevant to the surface of Mars, or basically anywhere else in the solar system.
What's the maintenance like on a bucket wheel excavator? Purely electrically powered, it's a much closer match to what mining on Mars might be like.
And then let's embrace decentralized AI and decentralized search indexes to reduce reliance on centralized providers like Google. I want my personal AI assistants to exist on local hardware without spying on me and relaying that information to big corporations and the government. And gigabytes of search indexes can be seeded, shared and synced by your personal preference without reliance on a centralized repository.
Finally, something to do with all the horsepower of a modern desktop.
It's also a nonstarter. With the exception of the Apple captives, no one is spending the big bucks on desktop hardware anymore. Hell, even gamers are getting by with slightly older, slightly lower spec machines. Worse, it takes a damn lot of bandwidth to run a web spider, and sharing all those indexes with the system means vast upstream bandwidth is required, and that's simply not available. I have a 400 megabit connection. Downstream only. Upstream is 20 megabit. Bittorrent works with asymmetrical connections because requests for content generally aren't real time. Start a download, walk away, it's done when it's done. Fetching a few gigabytes worth of an index before your web search will work well isn't reasonable though. And if you're not fetching the index, you're waiting for dozens, hundreds, or thousands of high ping systems with shitty upstream bandwidth to respond. Joe User is not going to put up with that degradation of service.
It sounds like a fascinating software problem, but replicating the Googleplex with unreliable, slow, asymmetrical connections will never be able to compete with the Googleplex's in-datacenter linkages. Sure you can use all the same techniques to deal with the commodity hardware, but the bandwidth will always be a problem.
If all of the consumer Internet connections in the world were symmetrical and gigabit and unlimited, we could talk about decentralizing the behemoths. They're not, and they won't be, and with net neutrality gone in the US, they're going to get worse, not better, with artificial asymmetries induced depending on destination. And no, Sweden can't supply everybody else.
No comma was placed after "handlers," a grammatical nuance a native English speaker would be more likely to add.
Not anymore. Modern English teaching says to omit most commas traditionally used, especially for prepositional phrases in prefix and suffix positions and for objects of command sentences like that one.
Intel's Broadwell Xeons with 24 cores have 7.2 billion transistors and a thermal design power of 300 watts. On a die 456 square millimeters. I'm having trouble finding a physical spec, so let's just say the chip is 4mm thick with package. 1824 cubic millimeters. Let's pack the entire cabinet with just compute silicon and nothing else, as an upper bound. 600mm x 2000mm x 1000mm interior dimensions is 1200000000 cubic millimeters. 657894 chips @ 300 watts is 197368421 watts.
Your standard size rack of silicon will require 197 megawatts to run. Of course it will melt before it finishes startup...
You're going to have to wait until somebody creates computronium before the density you're describing is physically possible. It will require reversible computing and on-board non-volatile memory.
Personally I'd be satisfied with a 100mm cube of the stuff, once it exists. It'd take 164 kilowatts to run if it were silicon. And it'd be capable of several teraFLOPS.
How quickly you forget that IBM, Red Hat, and Novell teamed to soundly defeat the much hated copyright troll SCO over Unix/Linux Copyright Disputes
Copyrights are not patents. IBM files scads of useless patents. They also do some genuine R&D, but its buried under mounds of useless bullshit these days.
Why would you fire an employee that produced systemd, pulseaudio, and contributed to many other projects that were widely successful and adopted by the majority of Linux distributions?
Because his code is shit. Pulseaudio is laughably broken on 90% of the distros that include it, and can not be fixed.
You don't get rid of successful productive employees.
You do get rid of drooling morons whose total lifetime code output couldn't make up for the tens of thousands of hours wasted by sysadmins trying to deal with his shit.
For a while it was all ipads ipads ipads, every student gets an ipad, and schools couldn't buy enough ipads, and then the schools discovered they weren't really all that great for education after all. And now home users are finding between their smartphone and their laptop the tablet isn't that useful there either, and the next great thing is now becoming a niche -- still useful and definitely has a place but we didn't get rid of all our computers for them in the end.
Chromebooks are the new tablets which were the new netbooks... maybe they'll take hold... or maybe they'll be ultimately found to be too limiting too. The jury's still out.
The iPad craze was simply Apple's tremendously capable marketing machine in action. Once people got them in their hands they discovered they were useless.
There is one very big difference between iPads and Chromebooks: Chromebooks have physical keyboards. Tablets are passive consumption devices. They have no other purpose. Even tablet games are practically passive experiences because there's neither the screen real estate nor the input devices available to manage a complex UI. Not even Apple's marketing could paper over the reality that tablet-based educational software is little more than a glorified PowerPoint, no matter how glitzy it can be made to look.
Chromebooks have keyboards. And despite all the nonsense coming out of bad educational studies, the one thing that has survived is the truth that children learn by interacting with words. And that distinction is important. Just tapping the screen to Go Next is a useless interaction. What's required is composing their own words. In traditional classrooms, that meant verbally. In Chromebook classrooms, that means written. And for that, you need a physical keyboard.
Sure a Chromebook isn't a "real" computer. Or is it... Did you know you can program an Arduino from a Chromebook? You can. You can also edit images, audio, and video...and run any Linux application.
The jury is still out indeed, but Chromebooks really do have the manageability needed while still presenting the one required tool to enable composition, which is all-important. This doesn't smell like overhyped marketing to me. This smells like a need fulfilled, with hints of actual classroom success.
Here is a crazy thought - maybe because someone differs with you on their views related to net neutrality; which it self is tied up in big untestable economic theory, it does not automatically have to mark them as evil or your sworn enemy.
it's not an "economic theory" you blithering idiot. It's demonstrated fact. When ISPs are allowed to be shitheads by Ajit Shithead, they ARE shitheads. We've already seen it happen. This is not some theory. This is not guesswork. This is not wishes and dreams and idealism. This is fucking reality. Anti-neutral networks already exist and are already unfair, and are already distorting the economy compared to the previously de facto neutral reality.
This sounds more like it would be the jurisdiction of the SEC, unless they found potential criminal activity beyond just allegedly misleading investors.
The big automakers were waiting for Tesla to die so that can continue business as usual. They terribly underestimated the potential of the electric car.
The big automakers were actively working to kill Tesla, paying for FUD in the press at least as much as the shortsellers and WERE some of the shortsellers themselves, since they're so old and so big they've been characterized as "a finance company with an auto business on the side". They didn't underestimate the potential of the electric car. They just failed to react appropriately.
They reacted with fear and tried to quash it, instead of embracing a new product category. If they had seen the writing on the wall, they actually could have crushed Tesla as the FUD has been saying for years. But they didn't. They didn't budge, and now the FUD has collapsed, with current estimates that they're at least 3 years and as many as 7 years behind Tesla in implementing their own electric cars. Now that Tesla is profitable, with a 3-7 year lead, they will establish an unassailable position for themselves in the automotive industry. It's too late for the big automakers to eliminate them, by FUD or by competing.
They saw it coming. They just thought they could make Tesla fail.
We have the grid capacity to charge even if ALL the cars become electric tomorrow.
Citation required.
He showed his work. Here's (some of) the basis for the 33% number. The link is for New England only, in 2011, but the numbers only look better for the rest of the country which uses more peak power for air conditioning than New England does, not less.
The numbers are actually getting dramatically better for available daytime capacity thanks to the ongoing installation of solar panels. The grid is so overpowered now that in two regions (California and New England), wholesale electricity prices go negative during the day. -$2.65 in the cited article is a vast oversupply. It is not a daily occurrence. Yet. It will become one, and sooner than you'd think. Seven years ago, it never happened in New England. Now it does occasionally. As time goes by, and solar installations continue, it will become more and more frequent.
Considering the grid is already heavily oversupplied today, during the day, in some parts of the country, the idea that if all cars magically become electric overnight and could still be recharged is not very farfetched. It was possible with only night time oversupply. Now that there is also daytime oversupply, it's down right easy.
I suspect relatively few people would have heard of this if Samsung hadn't proposed suing.
I suspect Samsung actually checked and found out. Were people posting the video of her trying to hide her iPhone under a piece of paper on camera? Where people reposting the video? Voting it up? Yes? Then we're suing, and suing big.
Now maybe we wouldn't have heard about it at all if Samsung hadn't sued, since it was all on Russian social media, which is almost as insular as Chinese social media, but then again we might have. And still, she negatively affected their brand image in an entire country. Letting that slide sounds like a bad idea, if only because they have other brand ambassadors around the world.
Americans are the most notoriously friendly people on Earth. "Notorious", because most people from other countries find Americans' tendency to strike up conversations with strangers off-putting.
That's all small talk is, it's establishing a safe space between strangers.
Considering America has more overseas military bases than any other nation on Earth, maybe random citizens wandering around overseas had better keep up their "notorious" signaling.
Moreover, Apple and Amazon allegedly knew about these boards back in 2015, yet Apple didn't dump SuperMicro until 2016...
There seems to be a persistent misunderstanding of the timeline. The initial detection wasn't, "Hey, I found a board with an amazing spy chip in it!" The initial detection was, " That's funny..."
I could easily believe that it took a year of painstaking labor for the alleged Canadian security company to track down the source of the rogue packets on the boards they were sent. There are a lot more likely things in the system to be generating the traffic than a chip that shouldn't be there. I'm sure it took quite a long time to verify all the various pieces of the system, from the apps to the OS to all the various firmwares.
...and Amazon was still using SuperMicro boards as of just a few months ago. Are you telling me that they kept using boards from SuperMicro for a year or three after finding out about this issue?
First year accounted for. Longer, if there was less concern initially, or a less vigorous investigation. After that, somebody has to decide what to do about it. You know damn well that nobody shifts tens of millions of dollars in business from one vendor to another overnight. There will be endless meetings, arguments, complaints, and bickerings, especially if the vendor's customer rep is a hot chick with a low cut blouse who somehow always has court-side seats for "her favorite VP". (Sounds sexist? You know it fucking happens.) Then once the decision is made, another vendor must be found, timelines must be established, samples produced, and contracts signed. That could easily take a year. Hell, getting samples from a new vendor could take a year all by itself. Companies wish they were more agile than that, but once they reach a certain size they aren't, because "due diligence" always transitions to "ass covering" and that takes time.
If it later came out that Bloomberg was right, but that Apple and Amazon had chosen to make categorical denials despite knowing better, we'd lose count at the number of lawsuits and criminal charges filed against them. They'd have knowingly misled their shareholders, repeatedly engaged in fraud in public statements, and lied to Congress, among other crimes and illicit activities.
When it comes to matters of national security, prosecutorial discretion kicks in hard. You can bet that any executive with material knowledge of the matter who is issuing categorical denials has a signed get-out-of-jail-free card squirreled away in a safe, so even if prosecution materializes, they're in the clear. But it won't. When the FBI tells the US Attorney "thou shalt not prosecute, oh, and turn over all notes and other material", the US Attorney hands over the file box and washes their hands of the matter.
Personally I expect the denials to stand unchallenged and the whole thing to go away, with a black eye for Bloomberg. Unless they got their hands on an actual chip or board with the chip and are sitting on it, plus a plausible chain of custody, they won't be able to prove it, regardless of how true the story is.
Is the story true? I bet it is. Just read about all the things the USSR pulled with the US embassy in Moscow. Governments will go to any lengths, no matter how foolish or ultimately useless, in the pursuit of advantage over other governments.
So a fairly large 4x4 meter solar panel (that would cost around $5000 to install)
Uhm. What? 325W solar panel, 1.9 square meters per panel, $225 plus shipping. 16 square meters costs $2025, if you insist on exceeding 16 square meters, requiring 9 panels. For 8 panels, it's $1800.
This is not some theoretical nebulous future price, either. This is the "add-to-cart button" price, including tariffs, available this afternoon.
Polycrystalline, so the efficiency isn't great and the warranty is indeed just 10 years, though they will work at some level of efficiency for decades longer than that. The warranty sheet says they'll provide 80% of their original power rating in full sun for 25 years.
You know I miss the days when stories like this would pop up and the first thing everyone would do is produce actual proof. The story literally says that China planted chips in their servers, but since the planted would have happened before the actual knowing where the board was going, they would have had to planted thousands of chips into boards in hopes of hitting a good target.
You have no understanding of the scale at which the cloud providers operate. Google, Facebook, Amazon, Apple, even Yahoo buy so many machines that they're ordering literally thousands at a time. Huge orders that the manufacturer damn well knows are going to one and only one customer, because they don't have thousands of boards just sitting on a shelf waiting for orders (it's called Just In Time Inventory management).
Further, Google and Facebook, at least, and probably all of them are so big that they're getting custom-designed boards specifically for themselves, which are not available to the general public at all. These customers are so big that the major manufacturers will happily do bespoke manufacturing (and charge concomitantly).
So a compromised board or two in a shipment of a thousand is quite easy to place at a single customer, and then no, we on the outside have no chance of seeing the evidence, if any, because there would be no misplaced compromising chips. The ones shipped, if there was more than one, will all be under the control of an entity with an overwhelming incentive, both financial and legal, to deny, deny, deny.
No I want to see this duplicitous capacitor or resistor looking chip that's somehow so well made that you can't tell the difference between it and an actual cap/resistor and somehow invades the board enough to leak useful info or make susceptible to an outside actor in a way that's undetectable.
I can believe the story is true, and that the named companies were victims. I see no reason to believe the technique was ever all that widespread, specifically because it is detectable. Bloomberg doesn't claim that it was all that widespread. They claim that Apple, out of an abundance of caution after finding one compromised board, removed all boards from that manufacturer. The Bloomberg article made it perfectly clear that the reason they have a story to write at all is because the major cloud providers have extremely good network traffic monitoring tools, saw the rogue network traffic, and started investigating. As far as Chinese Intelligence is concerned, this was an expensive failure. It only works in the bubble of the stereotypical lazy, sloppy American that a good deal of the rest of the world believes. As it turns out, not all Americans are sloppy.
We need to transform Internet-based services, chat, blogs, news (aka facebook, whatsapp, instagram, twitter etc.) into decentralized peer to peer services. Key for this transition is IPv6 because it enables us to give every human a set of personal IP addresses. For anonymization, we can use tor like overlays. Of course we need IPv6 routing without NAT.
You also need high speed symmetrical connections. Given that the majority of the money spent on the Internet is spent by people with radically asymmetrical connections under the thumb of companies with zero effective competition and a natural monopoly, this isn't going to change, and none of those services will ever be decentralized.
Quite aside from the fact that running Internet-facing services is technically difficult without getting pwned. How many consumer machines are part of botnets? To a first approximation, all of them.
Even if you provide a turnkey solution, a hardware box with custom software like those people trying to sell the personal ARM-based email server, your software has to be trivially easy to use, from all platforms, while simultaneously being perfectly secure. It will be neither. This is one of those problems where the 80% solution really doesn't cut it, simply because the competition is a 99% solution already. This is the fundamental mechanism for how monopolies arise.
If they cannot be broken up under tge current laws then I am of the opinion thye must then be changed,for Google has a monopoly of search engines and FB on the social network scene. They have done more harm than good and are already far too embedded into the life of ther average individual. This isnt about laws,but public health.
If we're redefining "monopoly" to mean >60% of a market, then the anti-trust lawyers are going to be busy for the next century.
By that definition:
Raytheon is too big. Lockheed Martin is too big. Boeing is too big. Kraft Heinz is too big. General Mills is too big. Procter & Gamble is too big. Dow Du Pont is too big. GlaxoSmithKline is too big. Walmart is too big. News Corp is too big. Walt Disney is too big. NBC Universal is too big. Comcast is too big. General Electric is too big. Bank of America is too big. Adobe is too big. NVIDIA is too big. BNSF is too big. Visa is too big. AT&T is too big.
Each and every one of those companies has a >60% marketshare in something, and has leveraged that marketshare to gain something else, somewhere, the criteria for violating the Sherman Anti-Trust Act.
You know what? I agree. Let's break them up. Let's break them all up.
Oh wait. You only wanted to break up the ones with politics you don't like. Right.
Selective enforcement is indistinguishable from fascism.
Slashdot Mirror
We legally enforce censorship all the time post a threat to kill the president of the United States and see how long until you get a visit from Uncle Sam and his G-Men.
Do they actually force you to remove the threat? It was my understanding that they don't. The Secret Service just wants to know if you're a credible threat or not. Your speech isn't censored. It just has consequences. And if you're not actually a credible threat, the consequences amount to momentary inconvenience. And your name on a list. I assume it's difficult to get a Top Secret clearance if you've posted a threat to kill the president, though I don't actually know. It just seems like something the DoD would frown upon.
"Super computers are mostly useless" gets a score of 2?
Logged in users with positive enough karma get a default 2 to start all posts. There have been no other mods to the GP post as of this post.
This is the part conspiracy theorists never grasp. There isn't any benefit to scientists for promoting a massive global conspiracy, but the benefits from puncturing such a conspiracy would be enormous.
Every adjunct professor in the world is waiting for an opportunity like this.
No they're not. Guess who sits on the grant committees? Old scientists, with a vested interest in defending the status quo. And defend it they will, by denying grants to any such upstart adjunct professor. There's a very good reason why Max Planck said science advances one funeral at a time, and it's far more relevant now than it ever was when he first wrote words to that affect in 1906. Adjunct professors don't have tenure. They do not rock the research boat. Not ever. Not anymore, if indeed they ever did.
Mathematics is the only academic pursuit immune from this problem. Math is math. Once a proof has been checked over thoroughly, it's unassailable. This is because math is not science. Math is an abstraction, with perfect internal consistency, because it is designed that way. Science deals with reality, which is not nearly as convenient.
It will take at least several hundred years for thawed out permafrost to become suitable for agriculture.
It will take one season, growing some non-food crop to establish soil bacteria, and then one afternoon spreading petroleum-based fertilizer the following season.
You didn't think natural soil was relevant to modern agriculture, did you?
No, but the only way around the problem is to develop tricky automated mining equipment and make all that stuff once you get there. I work with mining equipment. Maintenance intervals (oils/filters/etc) are every 50 hours of operation, machine-stopping breakdowns occur every few hundred hours, large component changeout (pumps, hydraulic cylinders, etc) is 4000 hours. 4000 hours is a year of operation at a 50% duty cycle. So you're going to ship all this stuff to Mars, and then expect it to run, continuously digging stuff up and crushing it and heating it and so on and so forth, for a couple of years? In a cold, dusty, zero-maintenance environment?
That sounds like dumptrucks and the like. Air-breathing diesel-powered vehicles, which are obviously irrelevant to the surface of Mars, or basically anywhere else in the solar system.
What's the maintenance like on a bucket wheel excavator? Purely electrically powered, it's a much closer match to what mining on Mars might be like.
And then let's embrace decentralized AI and decentralized search indexes to reduce reliance on centralized providers like Google. I want my personal AI assistants to exist on local hardware without spying on me and relaying that information to big corporations and the government. And gigabytes of search indexes can be seeded, shared and synced by your personal preference without reliance on a centralized repository.
Finally, something to do with all the horsepower of a modern desktop.
It's also a nonstarter. With the exception of the Apple captives, no one is spending the big bucks on desktop hardware anymore. Hell, even gamers are getting by with slightly older, slightly lower spec machines. Worse, it takes a damn lot of bandwidth to run a web spider, and sharing all those indexes with the system means vast upstream bandwidth is required, and that's simply not available. I have a 400 megabit connection. Downstream only. Upstream is 20 megabit. Bittorrent works with asymmetrical connections because requests for content generally aren't real time. Start a download, walk away, it's done when it's done. Fetching a few gigabytes worth of an index before your web search will work well isn't reasonable though. And if you're not fetching the index, you're waiting for dozens, hundreds, or thousands of high ping systems with shitty upstream bandwidth to respond. Joe User is not going to put up with that degradation of service.
It sounds like a fascinating software problem, but replicating the Googleplex with unreliable, slow, asymmetrical connections will never be able to compete with the Googleplex's in-datacenter linkages. Sure you can use all the same techniques to deal with the commodity hardware, but the bandwidth will always be a problem.
If all of the consumer Internet connections in the world were symmetrical and gigabit and unlimited, we could talk about decentralizing the behemoths. They're not, and they won't be, and with net neutrality gone in the US, they're going to get worse, not better, with artificial asymmetries induced depending on destination. And no, Sweden can't supply everybody else.
No comma was placed after "handlers," a grammatical nuance a native English speaker would be more likely to add.
Not anymore. Modern English teaching says to omit most commas traditionally used, especially for prepositional phrases in prefix and suffix positions and for objects of command sentences like that one.
Intel's Broadwell Xeons with 24 cores have 7.2 billion transistors and a thermal design power of 300 watts. On a die 456 square millimeters. I'm having trouble finding a physical spec, so let's just say the chip is 4mm thick with package. 1824 cubic millimeters. Let's pack the entire cabinet with just compute silicon and nothing else, as an upper bound. 600mm x 2000mm x 1000mm interior dimensions is 1200000000 cubic millimeters. 657894 chips @ 300 watts is 197368421 watts.
Your standard size rack of silicon will require 197 megawatts to run. Of course it will melt before it finishes startup...
You're going to have to wait until somebody creates computronium before the density you're describing is physically possible. It will require reversible computing and on-board non-volatile memory.
Personally I'd be satisfied with a 100mm cube of the stuff, once it exists. It'd take 164 kilowatts to run if it were silicon. And it'd be capable of several teraFLOPS.
How quickly you forget that IBM, Red Hat, and Novell teamed to soundly defeat the much hated copyright troll SCO over Unix/Linux Copyright Disputes
Copyrights are not patents. IBM files scads of useless patents. They also do some genuine R&D, but its buried under mounds of useless bullshit these days.
Why would you fire an employee that produced systemd, pulseaudio, and contributed to many other projects that were widely successful and adopted by the majority of Linux distributions?
Because his code is shit. Pulseaudio is laughably broken on 90% of the distros that include it, and can not be fixed.
You don't get rid of successful productive employees.
You do get rid of drooling morons whose total lifetime code output couldn't make up for the tens of thousands of hours wasted by sysadmins trying to deal with his shit.
For a while it was all ipads ipads ipads, every student gets an ipad, and schools couldn't buy enough ipads, and then the schools discovered they weren't really all that great for education after all. And now home users are finding between their smartphone and their laptop the tablet isn't that useful there either, and the next great thing is now becoming a niche -- still useful and definitely has a place but we didn't get rid of all our computers for them in the end.
Chromebooks are the new tablets which were the new netbooks... maybe they'll take hold... or maybe they'll be ultimately found to be too limiting too. The jury's still out.
The iPad craze was simply Apple's tremendously capable marketing machine in action. Once people got them in their hands they discovered they were useless.
There is one very big difference between iPads and Chromebooks: Chromebooks have physical keyboards. Tablets are passive consumption devices. They have no other purpose. Even tablet games are practically passive experiences because there's neither the screen real estate nor the input devices available to manage a complex UI. Not even Apple's marketing could paper over the reality that tablet-based educational software is little more than a glorified PowerPoint, no matter how glitzy it can be made to look.
Chromebooks have keyboards. And despite all the nonsense coming out of bad educational studies, the one thing that has survived is the truth that children learn by interacting with words. And that distinction is important. Just tapping the screen to Go Next is a useless interaction. What's required is composing their own words. In traditional classrooms, that meant verbally. In Chromebook classrooms, that means written. And for that, you need a physical keyboard.
Sure a Chromebook isn't a "real" computer. Or is it... Did you know you can program an Arduino from a Chromebook? You can. You can also edit images, audio, and video...and run any Linux application.
The jury is still out indeed, but Chromebooks really do have the manageability needed while still presenting the one required tool to enable composition, which is all-important. This doesn't smell like overhyped marketing to me. This smells like a need fulfilled, with hints of actual classroom success.
Here is a crazy thought - maybe because someone differs with you on their views related to net neutrality; which it self is tied up in big untestable economic theory, it does not automatically have to mark them as evil or your sworn enemy.
it's not an "economic theory" you blithering idiot. It's demonstrated fact. When ISPs are allowed to be shitheads by Ajit Shithead, they ARE shitheads. We've already seen it happen. This is not some theory. This is not guesswork. This is not wishes and dreams and idealism. This is fucking reality. Anti-neutral networks already exist and are already unfair, and are already distorting the economy compared to the previously de facto neutral reality.
This sounds more like it would be the jurisdiction of the SEC, unless they found potential criminal activity beyond just allegedly misleading investors.
It pays to have friends in low places.
The big automakers were waiting for Tesla to die so that can continue business as usual. They terribly underestimated the potential of the electric car.
The big automakers were actively working to kill Tesla, paying for FUD in the press at least as much as the shortsellers and WERE some of the shortsellers themselves, since they're so old and so big they've been characterized as "a finance company with an auto business on the side". They didn't underestimate the potential of the electric car. They just failed to react appropriately.
They reacted with fear and tried to quash it, instead of embracing a new product category. If they had seen the writing on the wall, they actually could have crushed Tesla as the FUD has been saying for years. But they didn't. They didn't budge, and now the FUD has collapsed, with current estimates that they're at least 3 years and as many as 7 years behind Tesla in implementing their own electric cars. Now that Tesla is profitable, with a 3-7 year lead, they will establish an unassailable position for themselves in the automotive industry. It's too late for the big automakers to eliminate them, by FUD or by competing.
They saw it coming. They just thought they could make Tesla fail.
We have the grid capacity to charge even if ALL the cars become electric tomorrow.
Citation required.
He showed his work. Here's (some of) the basis for the 33% number. The link is for New England only, in 2011, but the numbers only look better for the rest of the country which uses more peak power for air conditioning than New England does, not less.
The numbers are actually getting dramatically better for available daytime capacity thanks to the ongoing installation of solar panels. The grid is so overpowered now that in two regions (California and New England), wholesale electricity prices go negative during the day. -$2.65 in the cited article is a vast oversupply. It is not a daily occurrence. Yet. It will become one, and sooner than you'd think. Seven years ago, it never happened in New England. Now it does occasionally. As time goes by, and solar installations continue, it will become more and more frequent.
Considering the grid is already heavily oversupplied today, during the day, in some parts of the country, the idea that if all cars magically become electric overnight and could still be recharged is not very farfetched. It was possible with only night time oversupply. Now that there is also daytime oversupply, it's down right easy.
No, Tesla is not making a profit. The financials are public, look at them.
2018 Q2, $520 million loss on revenue of 4 billion.
You Musk/Tesla shills are unbelievable, denying reality.
Yes, Tesla is making a profit. The financials are public, look at them.
2018 Q3, $312 million profit on revenue of $6.8 billion.
You gas guzzler shills are unbelievable, denying reality.
I suspect relatively few people would have heard of this if Samsung hadn't proposed suing.
I suspect Samsung actually checked and found out. Were people posting the video of her trying to hide her iPhone under a piece of paper on camera? Where people reposting the video? Voting it up? Yes? Then we're suing, and suing big.
Now maybe we wouldn't have heard about it at all if Samsung hadn't sued, since it was all on Russian social media, which is almost as insular as Chinese social media, but then again we might have. And still, she negatively affected their brand image in an entire country. Letting that slide sounds like a bad idea, if only because they have other brand ambassadors around the world.
Americans are the most notoriously friendly people on Earth. "Notorious", because most people from other countries find Americans' tendency to strike up conversations with strangers off-putting.
That's all small talk is, it's establishing a safe space between strangers.
Considering America has more overseas military bases than any other nation on Earth, maybe random citizens wandering around overseas had better keep up their "notorious" signaling.
Moreover, Apple and Amazon allegedly knew about these boards back in 2015, yet Apple didn't dump SuperMicro until 2016...
There seems to be a persistent misunderstanding of the timeline. The initial detection wasn't, "Hey, I found a board with an amazing spy chip in it!" The initial detection was, " That's funny..."
I could easily believe that it took a year of painstaking labor for the alleged Canadian security company to track down the source of the rogue packets on the boards they were sent. There are a lot more likely things in the system to be generating the traffic than a chip that shouldn't be there. I'm sure it took quite a long time to verify all the various pieces of the system, from the apps to the OS to all the various firmwares.
...and Amazon was still using SuperMicro boards as of just a few months ago. Are you telling me that they kept using boards from SuperMicro for a year or three after finding out about this issue?
First year accounted for. Longer, if there was less concern initially, or a less vigorous investigation. After that, somebody has to decide what to do about it. You know damn well that nobody shifts tens of millions of dollars in business from one vendor to another overnight. There will be endless meetings, arguments, complaints, and bickerings, especially if the vendor's customer rep is a hot chick with a low cut blouse who somehow always has court-side seats for "her favorite VP". (Sounds sexist? You know it fucking happens.) Then once the decision is made, another vendor must be found, timelines must be established, samples produced, and contracts signed. That could easily take a year. Hell, getting samples from a new vendor could take a year all by itself. Companies wish they were more agile than that, but once they reach a certain size they aren't, because "due diligence" always transitions to "ass covering" and that takes time.
If it later came out that Bloomberg was right, but that Apple and Amazon had chosen to make categorical denials despite knowing better, we'd lose count at the number of lawsuits and criminal charges filed against them. They'd have knowingly misled their shareholders, repeatedly engaged in fraud in public statements, and lied to Congress, among other crimes and illicit activities.
When it comes to matters of national security, prosecutorial discretion kicks in hard. You can bet that any executive with material knowledge of the matter who is issuing categorical denials has a signed get-out-of-jail-free card squirreled away in a safe, so even if prosecution materializes, they're in the clear. But it won't. When the FBI tells the US Attorney "thou shalt not prosecute, oh, and turn over all notes and other material", the US Attorney hands over the file box and washes their hands of the matter.
Personally I expect the denials to stand unchallenged and the whole thing to go away, with a black eye for Bloomberg. Unless they got their hands on an actual chip or board with the chip and are sitting on it, plus a plausible chain of custody, they won't be able to prove it, regardless of how true the story is.
Is the story true? I bet it is. Just read about all the things the USSR pulled with the US embassy in Moscow. Governments will go to any lengths, no matter how foolish or ultimately useless, in the pursuit of advantage over other governments.
When our army can't defend us we are gone.
I hope we will be remembered.
Crocodile Dundee will never be forgotten.
Asked cat. Cat said ...
So a fairly large 4x4 meter solar panel (that would cost around $5000 to install)
Uhm. What? 325W solar panel, 1.9 square meters per panel, $225 plus shipping. 16 square meters costs $2025, if you insist on exceeding 16 square meters, requiring 9 panels. For 8 panels, it's $1800.
This is not some theoretical nebulous future price, either. This is the "add-to-cart button" price, including tariffs, available this afternoon.
Polycrystalline, so the efficiency isn't great and the warranty is indeed just 10 years, though they will work at some level of efficiency for decades longer than that. The warranty sheet says they'll provide 80% of their original power rating in full sun for 25 years.
You know I miss the days when stories like this would pop up and the first thing everyone would do is produce actual proof. The story literally says that China planted chips in their servers, but since the planted would have happened before the actual knowing where the board was going, they would have had to planted thousands of chips into boards in hopes of hitting a good target.
You have no understanding of the scale at which the cloud providers operate. Google, Facebook, Amazon, Apple, even Yahoo buy so many machines that they're ordering literally thousands at a time. Huge orders that the manufacturer damn well knows are going to one and only one customer, because they don't have thousands of boards just sitting on a shelf waiting for orders (it's called Just In Time Inventory management).
Further, Google and Facebook, at least, and probably all of them are so big that they're getting custom-designed boards specifically for themselves, which are not available to the general public at all. These customers are so big that the major manufacturers will happily do bespoke manufacturing (and charge concomitantly).
So a compromised board or two in a shipment of a thousand is quite easy to place at a single customer, and then no, we on the outside have no chance of seeing the evidence, if any, because there would be no misplaced compromising chips. The ones shipped, if there was more than one, will all be under the control of an entity with an overwhelming incentive, both financial and legal, to deny, deny, deny.
No I want to see this duplicitous capacitor or resistor looking chip that's somehow so well made that you can't tell the difference between it and an actual cap/resistor and somehow invades the board enough to leak useful info or make susceptible to an outside actor in a way that's undetectable.
I can believe the story is true, and that the named companies were victims. I see no reason to believe the technique was ever all that widespread, specifically because it is detectable. Bloomberg doesn't claim that it was all that widespread. They claim that Apple, out of an abundance of caution after finding one compromised board, removed all boards from that manufacturer. The Bloomberg article made it perfectly clear that the reason they have a story to write at all is because the major cloud providers have extremely good network traffic monitoring tools, saw the rogue network traffic, and started investigating. As far as Chinese Intelligence is concerned, this was an expensive failure. It only works in the bubble of the stereotypical lazy, sloppy American that a good deal of the rest of the world believes. As it turns out, not all Americans are sloppy.
We need to transform Internet-based services, chat, blogs, news (aka facebook, whatsapp, instagram, twitter etc.) into decentralized peer to peer services. Key for this transition is IPv6 because it enables us to give every human a set of personal IP addresses. For anonymization, we can use tor like overlays. Of course we need IPv6 routing without NAT.
You also need high speed symmetrical connections. Given that the majority of the money spent on the Internet is spent by people with radically asymmetrical connections under the thumb of companies with zero effective competition and a natural monopoly, this isn't going to change, and none of those services will ever be decentralized.
Quite aside from the fact that running Internet-facing services is technically difficult without getting pwned. How many consumer machines are part of botnets? To a first approximation, all of them.
Even if you provide a turnkey solution, a hardware box with custom software like those people trying to sell the personal ARM-based email server, your software has to be trivially easy to use, from all platforms, while simultaneously being perfectly secure. It will be neither. This is one of those problems where the 80% solution really doesn't cut it, simply because the competition is a 99% solution already. This is the fundamental mechanism for how monopolies arise.
If they cannot be broken up under tge current laws then I am of the opinion thye must then be changed,for Google has a monopoly of search engines and FB on the social network scene. They have done more harm than good and are already far too embedded into the life of ther average individual. This isnt about laws,but public health.
If we're redefining "monopoly" to mean >60% of a market, then the anti-trust lawyers are going to be busy for the next century.
By that definition:
Raytheon is too big.
Lockheed Martin is too big.
Boeing is too big.
Kraft Heinz is too big.
General Mills is too big.
Procter & Gamble is too big.
Dow Du Pont is too big.
GlaxoSmithKline is too big.
Walmart is too big.
News Corp is too big.
Walt Disney is too big.
NBC Universal is too big.
Comcast is too big.
General Electric is too big.
Bank of America is too big.
Adobe is too big.
NVIDIA is too big.
BNSF is too big.
Visa is too big.
AT&T is too big.
Each and every one of those companies has a >60% marketshare in something, and has leveraged that marketshare to gain something else, somewhere, the criteria for violating the Sherman Anti-Trust Act.
You know what? I agree. Let's break them up. Let's break them all up.
Oh wait. You only wanted to break up the ones with politics you don't like. Right.
Selective enforcement is indistinguishable from fascism.