Theoretical answer: Recode your algorithm in SystemC (a c++ library that can be used to implement a register transfer language representation of your algorithm) and synthesize it with one of the available tools (e.g., Accelera, Synopsys, Calypto, etc) targeting a typical library (e.g, 28nm TSMC), at a particular clock frequency.
Practical answer: Ask someone with hw design experience to estimate it for you...
FWIW, nobody wants an "exact" size in logic gates, all they want an idea in complexity. The big ticket items people care about are the size in bits of RAMs (and how many simultaneous read/write ports it might need) and complicated math that is likely to take more than 1 clock cycle to complete (e.g., like a floating point math operation) and the data-width of the main data path at the throughput that you want to have. Simply multiplying the data path width by the estimated number of pipeline cycles is generally proportional to the eventual area minus the RAMS and special math ops (which is why you need to identify those parts separately).
Generally, I've found that naïve "software" algorithms have not been very amenable to HW implementation without some amount of rework and the fact that you do not have an answer to the posed question would likely lead me (and probably your potential customer), that your algorithm is half-baked from a HW implementation point of view... Just food for thought...
I would disagree with this. Although most of the CPU instruction set designs and programmer paradigms are directly descended from microcontroller architectures in the 60's and 70's, but most of the lineage of CPUs today is from the early super-scalar designs from the '80s...
Some might credit the CDC6600 (a early generation super-computer, not a microcontroller) as the first super scalar design that inspired this, but it was more akin to a unified instruction stream with several non-pipelined functional units (kind of like how early CPU's dispatched instructions to an FPU or other co-processor device).
Generic malware that mimic alerts to fool computer users to click to download an exploit might be largely ineffective and often ignored. Researchers, however, have proposed a change to the status quo, believing instead that malware should be re-architected to include the same concrete, specific warnings that will be used in the future to maintain the status quo.
The more things change, the more they remain the same...
Binary thinking
on
Losing Aaron
·
· Score: 4, Insightful
Binary thinking is for the young. The dichotomy of the idealist and the pragmatist is really a failing of idealism that pragmatism is thought to be unprincipled. In most cases, there is no true understanding of an ideal, it is really just an unembodied idea with unforeseen consequences. The truism is that world is changed by people that can change the world, not people that only wish it. Of course one would hope to change the world for the better rather than the worse, but it's the height of arrogance that a single person always know which direction is better.
It's not easy to raise a pragmatist that realizes every hand is a winner and a loser, and not to count their money when it's sitting at the table... It's much easier to raise an idealist that ignores these facts of life and laments the world for not seeing things the way they do. That's what my 3-yo son does now. He doesn't know the rules to all the games that we play together and the surprising (to him) strategies that seem to work yet, so he makes some assumptions which sometimes don't turn out to be true and that disappoints him (which is reflected in different severity of negative behaviors depending on his energy status / mood).
I'm hoping to prepare him to experience his own chosen path in life by pointing him in what I think is a good direction and giving him some tools to continue to make his way as he discovers the true rules of various "games" he will experience. I'm sure I too will regret not giving my son enough tools someday, but he owns his own life path...
On principle (one of the few that I have), my son doesn't have to change the world to be better for me (or the rest of society). Actually, I think that deliberately sending my son on such a mission would be sociopathic behavior on *my* part (might be akin to sending him on a suicide mission if I can be momentarily extra insensitive to the above mentioned situation). However, if he choses to walk that path, I wish him all the luck.
Has it actually been proven that it is mathematically impossible for a quantum algorithm to exist capable of defeating this system? I'm sure you could prove that any particular known algorithm wouldn't work, but the only system resistant to unknown algorithms that I'm aware of is the one-time pad.
If this has been proven I'm genuinely interested. I will confess I'm not a cryptographer.
I don't know about ring-learning-with-errors, but if it indeed reduces to an integer lattice problem, I suspect it would eventually prove to be vulnerable to some sort of attack that could be executed by a quantum computer.
As a silly example, here's a proposed attack on lattices that employs a quantum computer implementing a partial Grover's algorithm to speed up looking for solutions...
As with many things, I doubt there is a negative proof. There's much about quantum computability that we do not understand yet (of course there is much about regular computability that we don't understand either, starting with P ?= NP). When people usually say it's resistant to quantum computers, they actually are implying that it's resistant to a quantum computer employing Shor's algorithm (and similar quantum fourier transform techiques) to factor large numbers and compute discrete logarithms (the basis behind the RSA and DH public key cryptosystems). There are other algorithms that quantum computer can run, most of which people have not even discovered yet.
They are not an engineering organization and they will not be at the forefront of qubit manufacturing.
How do you know this? The NSA purchased an old abandoned Sony chip fab in San Antonio and started to re-commission it back in 2006, who knows what they are doing with it for the last 6 years? One of the promising target architectures for a large scale qubit is a cryogenically-cooled silicon double quantum-dot scheme. They might have more going on in this area than you might guess...
Although under maritime conventions, the cost of rescuing people entails no obligation of reimbursement, freeing boat from ice might be considered a salvage operation. If so, the costs for freeing the boat from the ice may mean that the owners of the boat might be liable for the cost of any salvage operation (if successful) maybe even up to 50% of the value of the boat.
Coca-Cola is typically produced not by the main company, but a local distributor/bottler who has a license from the main company.
FWIW: I think fountains (including the new "freestyle" machine) are leased directly from the Coca-Cola company (as opposed to the stuff in cans and bottles in vending machines which are produced and sold by distributors).
As I remember hearing in a journalism conference once upon a time: Noteworthy-Entertainment-With-Substance This has now apparently morphed into: Narcissistic-Enterprise-With-Sarcasm
Sad, as journalism used to be a fun endeavor. You had to juggle 3 things, your sponsors, your subscribers, and your conscience to produce the best product you can. Now there are no subscribers and the advertising revenue are supplied by 3rd parties in bulk (e.g., adwords, and similar platforms) so you are left with only the conscience of the publisher which is the most susceptible to human failing...
Note that in the news there was never objective truth (or so journalism professors tell us). The point of news was always entertainment. The goal of a journalist is to produce information of substance. Of course if the facts are wrong, that isn't substance, so fact checking is always important, but it's important to remember that news being limited in time and duration is always a lie-by-omission.
When people talk about News they often confuse it with Features. Features are the heavily researched reporting that is done in a longer time cycle. News has always been on a short cycle.
Nearly all the stuff they are talking about requires your machine to be compromised somehow (virus/malware allowing malicious code execution). Once that happens, it's game over. It's just that the NSA has considerable resources to both attack a OS network access (looking for zero-day equivalent exploits to do that), and open source probably just makes it easier to look for likely problems in the source code (instead of trying to steal the source code from some company or disassembling the code).
The way to protect yourself is to make your own distribution that doesn't show up on their radar (they don't have infinite resources to attack everything, just the most common stuff).
The thing that they are actually talking about in this article is once the OS is compromised, the NSA has developed attacks on the firmware of many commonly available peripherals (e.g., routers, drives, etc) to insert compromises that live underneath the OS (and thus undetectable by the OS). The other articles talk about intercepting packages and actually physically compromising the firmware of these peripherals in transit.
The expressed protein is a Bt toxin, which is approved for and used on organic farms as a natural pesticide.
Although the expressed protein is a Bt toxin, and application of Bt is considered as a organic natural pesticide spray, they aren't quite the same. In the GM variant, it is actually produced in the corn itself, where in the organic farming case, an inactivated Bt bacteria solution is sprayed on corn and you are only consuming the pesticide residue. The difference is in the quantity that you might ingest.
Personally, I don't think GM modifications like MIR162 as a major consumption hazard (I'm sure I've eaten some already as I live in the USA), but I'm a believer in full disclosure and letting people decide. On the other hand, I think of cultivation of stuff like MIR162 as a latent environmental hazard not unlike anti-bacterial soap or people overusing anti-biotics. Low-level doses that permeate the environment seem like they will inevitably breed resistance in the pests. Indeed this is happening with Bt toxin as some pests have already developed resistance to GM-corn with Bt genes.
In case anyone was wondering, the way they find out if there is GM contamination, is that the GM variant has a special marker gene inserted in it. The tests are for the presence of the marker, not for the toxin itself. http://www.syngentabiotech.com/biotech_licensing/pmi_technology.aspx
FYI, MIR162 was approved for use by the EU. Of course, this might be because it was developed by a Swiss company (Syngenta)...
The modification made to MIR162 (insertion of a gene from Bacillus thuringiensis aka Bt which creates the Vip3Aa protein) is in some ways the complement to what was known as BtCorn (which is the generic moniker for many varieties of corn which inserted one of the other genes from Bt and created a different protein Cry1Ac and was developed by the non-European company Monsanto). Apparently, many pest have developed a resistance to the Cry1Ac expression BtCorn.
There is another GM-corn which was called Starlink (owned by Bayer-Aventis, a French company) which used yet another protein from Bt (Cry9C), and there were problems in testing (the protein is apparently not broken down as easily and testing got prohibitively expensive as it showed some possibility for human allergic response) and it was eventually withdrawn from the market.
The deviations they are talking about aren't things like mountains or bumps, but a systematic non-spherical bias.
For example, the earth isn't spherical either, it's basically a bit fatter around the equator pretty close to an oblate spheroid (e.g., an M&M is a more exaggerated oblate spheroid). Like a baseball, if the electron isn't totally spherical, you can detect a systematic bias as it's being thrown around (you can think of the LHC as throwing an electron spit-ball or a knuckle-ball).
Although even in the standard model, the electron at some energy level probably has a detectable dipole moment (e.g., the charge wouldn't be uniformly spherically distributed in the electron), it is my understanding that it is predicted to be too small to be validated by current experiments. However, some versions of super-symmetry apparently would predict that the electron at some energy levels would have a larger detectable dipole moment . I guess these super-symmetry predictions didn't pan out.
This attack was made possible because of a specific optimization made in the GnuPG implementation. Because modular exponentiation of big numbers is time consuming, a general purpose optimization in question isn't always the fastest (specifically karatsuba multiplication technique) and branch is taken when the library heuristically thinks a naïve algorithm with some special case partial operand code would be faster. The branch is attacked by using a carefully chosen sequence of chosen ciphertexts to attack the branch that tickles the special-case code.
In contrast, most block ciphers do not have interior branches, but are repetitive data operations of substitution and permutation that would not exhibit this type of vulnerability (even more so with AES as there are specific CPU instructions that implement the inner loops of the algorithm making them highly resistant to side channel attacks like this).
And of course counter mode is not a csprng, but this has nothing to do with this. DRBGs, however can be based on block ciphers and they appear to be state of the art (unlike the well known backdoor-ed Dual-EC-DRBG).
Yours is the most useful comment on this fascinating story. Since you RTFA (sucka), what is the source of the CPU noise? What are its characteristics (frequency range, dynamic range) such that it can cut through other noises to register onto an at-distance audio mic?
BTW, I would have thought it bunk but for the name Adi Shamir in the author list, you know, the S of RSA.
The authors speculate that this is electro/magnetic-mechanical noise coming from the voltage regulators near the CPU (e.g., inductor "hum"). They note that it is hard to localize because the whole motherboard assembly is of course mechanically coupled together. The audio spectrum was characterized in a small range (35-39kHz) which is something that you can pick up with a standard microphone pickup. The dynamic range varied on different platforms, but in one case 0.5Hz separated the '0' and the '1' case.
So should academics use the same research and release standards with security bugs/exploits as government "hacks"?
The general consensus is that researchers that find vulnerabilities in commercial software at least notify the vendor and give them a chance to mitigate the problem before releasing a potential attack blue-print to the universe. Should researchers notify government entities before releasing their research on their "hacks", or are they to treat the government as a black-hat?
It's an interesting choice either way... One way you are colluding the other way you apparently might be considered a traitor (or hero depending on your politics).
> Spend the same $4 on fresh foods and they will go bad in days
Not a problem. Eat it in days.
I can buy the most expensive deli meats and organic buns and still come out ahead of McDonalds and Subway. Never mind the more mundane stuff and "real food".
It's a RESTAURANT. It has a 5x margin by it's very nature. It has to. Otherwise it could not possibly stay in business.
A 5x margin is easy to beat.
For every dollar a restaurant charges, they roughly average 33% food costs (and roughly 25-30% labor costs, the rest go to capital/equipment/rent/profit, etc). Fast food the labor costs are generally on the lower side (less staff per customer and lower wage scales). The profit margin per food category varies tremendously, with beverages being the highest by far and stuff like featured entrees being extremely low.
Since we are talking about fast food, for example, the dollar menu at McD's is nearly $0.90/item fully loaded cost, where a drink costs about $0.20 and they sell it for a dollar. The McDouble broke the bank (started about ~$0.96** and eventually was money losing) and they had to take it off the menu.
Of course if you are looking strictly at food, you might conclude you can do significantly better, but don't forget they have a bulk wholesale cost-of-food advantage too. Several estimates of the home cost (not including labor, or energy) of reproducing a McDouble range from about $1.25 to $2.00. But of course the home burger would be more tasty and probably healthier for you... Of course $2 is less that the so-called $6-burger at a quick-service restaurant.
**McD's originally wanted the McDouble to have 2 slices of cheese, but franchisees protested that would produce a loss so it was launched with 1 slice.
First off, there are a number of bios manufacturers...
Maybe in number, but not marketshare where there are basically 2: AMI and Phoenix/Award. The market share of all others is a rounding error.
Second, there are numerous bugs still existent.
True, but see point #1
This sounds like the NSA found someone in China using an exploit in a BIOS to hack computers. Alerted the manufacturer who was probably already aware of the fact after numerous Linux users had reported it years ago.
Probably likely, but not a consequence of your first two points.
You missed a few other open issues related to decommisioning (e.g., mostly what to do with the salt).
In any case, the only efforts I know of are: FUJI which I think died in the fund-raising stage back in 2011. TTS an attempt to resurrect this. Thor Energy
MSRE showed that the physics worked, however, as with many things, the engineering problems remain. AFAIK, most people are attempting to figure out the salt problem. The metal problem is currently unsolved (and a much more important problem since you need the reactor to have a reasonable operating life to make the whole thing economical in the first place).
Acutally you can sometimes get a cell signal from a phone on a plane (some lightly used rural towers can pretty high power/range). Unfortunatly the high-altitude and velocity of a plane used to confuse cell networks (esp., the tower-handoff protocols between say a 3G and a CDMA tower on a multi-mode phone) and consume too much of the tower's antenna power-bandwidth envelope leaving less for other calls. Because so many folks leave their cell-phones onboard planes, nowdays many networks are configured detect if you are hitting too many different types of towers and simply refuse to track/connect a phone exhibiting such a pattern.
Of course at 5miles up in cruise phase, you'd have to have a very good phone to get a signal because the signal you would get isn't from a tower straight down from you (the cell towers don't focus their energy up), but on a far away tower where your signal looks more parallel to the ground. But during a slow decent phase over a relatively unpopulated area, it's quite common to be able to get a decent signal somewhere between 3-4 miles in altitude.
But they aren't going to do that, they are going to put a femto-cell on the plane and probably back-haul it through a satellite network (likely the same one they use to provide wifi internet service). I imagine the call will look like a high-cost roaming fee on your cell-phone bill and they will block VoIP (if they don't do that already).
The quiet section is already monetized. It's called "first class", that wonderful place where the screaming of infants and children is reduced to a dull howl.
Was flying in first class once and two seats were occupied with 2 parents and 2 lap-children. Even though it was a short haul flight (2 hours on a 737), I felt sorry for the people in the seats next to them (because of the availability of oxygen masks, lap-children were restricted to one per row so they couldn't put them next to each other in the same row).
The cost moderated the usage so much that they're mostly no longer enabled even if they're present. The remaining service will be shutting down completely at the end of this year.
Here is a thought. An aeroplane is a public place, and you should not necessarily expect quiet on a plane. What next, do you ban people talking to each other on planes because is disturbs your peace and quiet?
Why not buy some earplugs if you want to drown out the noise around you? Why not let people use their time as productively as they wish.
If it is a red eye flight, then I fully agree that people should be quiet at certain times. but to mandate it for all flights, whatever the time is ridiculous.
True story. I was on a flight having an animated conversation with someone (probably a bit louder than we thought) and the flight attendant came over and gave us both a free in-flight movies and a drink in an effort to shut us up (well she was quite disarming polite about it). Don't know if someone complained or if it was a pre-emptive move, but it was quite effective.
On the other hand, I've seen a passenger yelling at a poor parent that was trying to comfort a baby crying. The flight attendant offered to change that person's seat, but she refused and instead had to find 2 adjacent seat for the parent and his 2 kids (fwiw, I offered mine, but my seat-neighbors weren't willing, maybe they didn't want to sit next to that passenger after all of her yelling).
Personally, noise doesn't bother me too much, but yelling is much more distracting than a baby crying. I really hate the personal-space battles, though. It's the arm rest tug-o-war that really gets on my nerve. Also the person that want you to hold their drink or put it on your tray table whilst they type a quick note on their monster laptop and that quick note lasts 45-minutes as they go over a presenation and write a few emails (okay that only happened twice, but seriously?). Often, I just pull out a laptop and put it on the tray-table and do nothing just to avoid that situation in the future. As long as their productivity doesn't intrude on my physical space, maybe it's okay...
Slashdot Mirror
Theoretical answer:
Recode your algorithm in SystemC (a c++ library that can be used to implement a register transfer language representation of your algorithm) and synthesize it with one of the available tools (e.g., Accelera, Synopsys, Calypto, etc) targeting a typical library (e.g, 28nm TSMC), at a particular clock frequency.
Practical answer:
Ask someone with hw design experience to estimate it for you...
FWIW, nobody wants an "exact" size in logic gates, all they want an idea in complexity. The big ticket items people care about are the size in bits of RAMs (and how many simultaneous read/write ports it might need) and complicated math that is likely to take more than 1 clock cycle to complete (e.g., like a floating point math operation) and the data-width of the main data path at the throughput that you want to have. Simply multiplying the data path width by the estimated number of pipeline cycles is generally proportional to the eventual area minus the RAMS and special math ops (which is why you need to identify those parts separately).
Generally, I've found that naïve "software" algorithms have not been very amenable to HW implementation without some amount of rework and the fact that you do not have an answer to the posed question would likely lead me (and probably your potential customer), that your algorithm is half-baked from a HW implementation point of view... Just food for thought...
I would disagree with this. Although most of the CPU instruction set designs and programmer paradigms are directly descended from microcontroller architectures in the 60's and 70's, but most of the lineage of CPUs today is from the early super-scalar designs from the '80s...
Some might credit the CDC6600 (a early generation super-computer, not a microcontroller) as the first super scalar design that inspired this, but it was more akin to a unified instruction stream with several non-pipelined functional units (kind of like how early CPU's dispatched instructions to an FPU or other co-processor device).
Generic malware that mimic alerts to fool computer users to click to download an exploit might be largely ineffective and often ignored. Researchers, however, have proposed a change to the status quo, believing instead that malware should be re-architected to include the same concrete, specific warnings that will be used in the future to maintain the status quo.
The more things change, the more they remain the same...
-40F == -40C. Clearly a pedant had to bother putting in the name of the scale used.
Now what's that in Kelvin (K)? That's what this pedant uses.
Cheers
Jon
Of course -40K is really, really HOT...
Binary thinking is for the young. The dichotomy of the idealist and the pragmatist is really a failing of idealism that pragmatism is thought to be unprincipled. In most cases, there is no true understanding of an ideal, it is really just an unembodied idea with unforeseen consequences. The truism is that world is changed by people that can change the world, not people that only wish it. Of course one would hope to change the world for the better rather than the worse, but it's the height of arrogance that a single person always know which direction is better.
It's not easy to raise a pragmatist that realizes every hand is a winner and a loser, and not to count their money when it's sitting at the table... It's much easier to raise an idealist that ignores these facts of life and laments the world for not seeing things the way they do. That's what my 3-yo son does now. He doesn't know the rules to all the games that we play together and the surprising (to him) strategies that seem to work yet, so he makes some assumptions which sometimes don't turn out to be true and that disappoints him (which is reflected in different severity of negative behaviors depending on his energy status / mood).
I'm hoping to prepare him to experience his own chosen path in life by pointing him in what I think is a good direction and giving him some tools to continue to make his way as he discovers the true rules of various "games" he will experience. I'm sure I too will regret not giving my son enough tools someday, but he owns his own life path...
On principle (one of the few that I have), my son doesn't have to change the world to be better for me (or the rest of society). Actually, I think that deliberately sending my son on such a mission would be sociopathic behavior on *my* part (might be akin to sending him on a suicide mission if I can be momentarily extra insensitive to the above mentioned situation). However, if he choses to walk that path, I wish him all the luck.
Has it actually been proven that it is mathematically impossible for a quantum algorithm to exist capable of defeating this system? I'm sure you could prove that any particular known algorithm wouldn't work, but the only system resistant to unknown algorithms that I'm aware of is the one-time pad.
If this has been proven I'm genuinely interested. I will confess I'm not a cryptographer.
I don't know about ring-learning-with-errors, but if it indeed reduces to an integer lattice problem, I suspect it would eventually prove to be vulnerable to some sort of attack that could be executed by a quantum computer.
As a silly example, here's a proposed attack on lattices that employs a quantum computer implementing a partial Grover's algorithm to speed up looking for solutions...
http://www.cdc.informatik.tu-darmstadt.de/reports/TR/TI-03-03.QSamplingPaper.pdf
As with many things, I doubt there is a negative proof. There's much about quantum computability that we do not understand yet (of course there is much about regular computability that we don't understand either, starting with P ?= NP). When people usually say it's resistant to quantum computers, they actually are implying that it's resistant to a quantum computer employing Shor's algorithm (and similar quantum fourier transform techiques) to factor large numbers and compute discrete logarithms (the basis behind the RSA and DH public key cryptosystems). There are other algorithms that quantum computer can run, most of which people have not even discovered yet.
They are not an engineering organization and they will not be at the forefront of qubit manufacturing.
How do you know this? The NSA purchased an old abandoned Sony chip fab in San Antonio and started to re-commission it back in 2006, who knows what they are doing with it for the last 6 years? One of the promising target architectures for a large scale qubit is a cryogenically-cooled silicon double quantum-dot scheme. They might have more going on in this area than you might guess...
Although under maritime conventions, the cost of rescuing people entails no obligation of reimbursement, freeing boat from ice might be considered a salvage operation. If so, the costs for freeing the boat from the ice may mean that the owners of the boat might be liable for the cost of any salvage operation (if successful) maybe even up to 50% of the value of the boat.
Coca-Cola is typically produced not by the main company, but a local distributor/bottler who has a license from the main company.
FWIW: I think fountains (including the new "freestyle" machine) are leased directly from the Coca-Cola company (as opposed to the stuff in cans and bottles in vending machines which are produced and sold by distributors).
Perhaps someone thought that since Jupiter's diameter is about 10x earth, it would have 10x the gravity?
Given g = G*m*(M/r^2)
If somehow the density of Jupiter was similar to earth...
Me = De*4*pi*re^3
Mj = De*4*pi*rj^3
The force of gravity experienced by a mass at the surface would be...
ge = G*m*(Me/re^2) = G*m*(De*4*pi*re/3)
gj = G*m*(Mj/rj^2) = G*m*(De*4*pi*rj/3) (proportional to the radius or diameter)
But of course, Jupiter is much less dense than the earth, so that analysis is totally bogus...
As I remember hearing in a journalism conference once upon a time: Noteworthy-Entertainment-With-Substance
This has now apparently morphed into: Narcissistic-Enterprise-With-Sarcasm
Sad, as journalism used to be a fun endeavor. You had to juggle 3 things, your sponsors, your subscribers, and your conscience to produce the best product you can. Now there are no subscribers and the advertising revenue are supplied by 3rd parties in bulk (e.g., adwords, and similar platforms) so you are left with only the conscience of the publisher which is the most susceptible to human failing...
Note that in the news there was never objective truth (or so journalism professors tell us). The point of news was always entertainment. The goal of a journalist is to produce information of substance. Of course if the facts are wrong, that isn't substance, so fact checking is always important, but it's important to remember that news being limited in time and duration is always a lie-by-omission.
When people talk about News they often confuse it with Features. Features are the heavily researched reporting that is done in a longer time cycle. News has always been on a short cycle.
Nearly all the stuff they are talking about requires your machine to be compromised somehow (virus/malware allowing malicious code execution). Once that happens, it's game over. It's just that the NSA has considerable resources to both attack a OS network access (looking for zero-day equivalent exploits to do that), and open source probably just makes it easier to look for likely problems in the source code (instead of trying to steal the source code from some company or disassembling the code).
The way to protect yourself is to make your own distribution that doesn't show up on their radar (they don't have infinite resources to attack everything, just the most common stuff).
The thing that they are actually talking about in this article is once the OS is compromised, the NSA has developed attacks on the firmware of many commonly available peripherals (e.g., routers, drives, etc) to insert compromises that live underneath the OS (and thus undetectable by the OS). The other articles talk about intercepting packages and actually physically compromising the firmware of these peripherals in transit.
The expressed protein is a Bt toxin, which is approved for and used on organic farms as a natural pesticide.
Although the expressed protein is a Bt toxin, and application of Bt is considered as a organic natural pesticide spray, they aren't quite the same.
In the GM variant, it is actually produced in the corn itself, where in the organic farming case, an inactivated Bt bacteria solution is sprayed on corn and you are only consuming the pesticide residue. The difference is in the quantity that you might ingest.
Personally, I don't think GM modifications like MIR162 as a major consumption hazard (I'm sure I've eaten some already as I live in the USA), but I'm a believer in full disclosure and letting people decide. On the other hand, I think of cultivation of stuff like MIR162 as a latent environmental hazard not unlike anti-bacterial soap or people overusing anti-biotics. Low-level doses that permeate the environment seem like they will inevitably breed resistance in the pests. Indeed this is happening with Bt toxin as some pests have already developed resistance to GM-corn with Bt genes.
In case anyone was wondering, the way they find out if there is GM contamination, is that the GM variant has a special marker gene inserted in it. The tests are for the presence of the marker, not for the toxin itself. http://www.syngentabiotech.com/biotech_licensing/pmi_technology.aspx
FYI, MIR162 was approved for use by the EU.
Of course, this might be because it was developed by a Swiss company (Syngenta)...
The modification made to MIR162 (insertion of a gene from Bacillus thuringiensis aka Bt which creates the Vip3Aa protein) is in some ways the complement to what was known as BtCorn (which is the generic moniker for many varieties of corn which inserted one of the other genes from Bt and created a different protein Cry1Ac and was developed by the non-European company Monsanto). Apparently, many pest have developed a resistance to the Cry1Ac expression BtCorn.
There is another GM-corn which was called Starlink (owned by Bayer-Aventis, a French company) which used yet another protein from Bt (Cry9C), and there were problems in testing (the protein is apparently not broken down as easily and testing got prohibitively expensive as it showed some possibility for human allergic response) and it was eventually withdrawn from the market.
The deviations they are talking about aren't things like mountains or bumps, but a systematic non-spherical bias.
For example, the earth isn't spherical either, it's basically a bit fatter around the equator pretty close to an oblate spheroid (e.g., an M&M is a more exaggerated oblate spheroid). Like a baseball, if the electron isn't totally spherical, you can detect a systematic bias as it's being thrown around (you can think of the LHC as throwing an electron spit-ball or a knuckle-ball).
Although even in the standard model, the electron at some energy level probably has a detectable dipole moment (e.g., the charge wouldn't be uniformly spherically distributed in the electron), it is my understanding that it is predicted to be too small to be validated by current experiments. However, some versions of super-symmetry apparently would predict that the electron at some energy levels would have a larger detectable dipole moment . I guess these super-symmetry predictions didn't pan out.
Actually, no.
This attack was made possible because of a specific optimization made in the GnuPG implementation. Because modular exponentiation of big numbers is time consuming, a general purpose optimization in question isn't always the fastest (specifically karatsuba multiplication technique) and branch is taken when the library heuristically thinks a naïve algorithm with some special case partial operand code would be faster. The branch is attacked by using a carefully chosen sequence of chosen ciphertexts to attack the branch that tickles the special-case code.
In contrast, most block ciphers do not have interior branches, but are repetitive data operations of substitution and permutation that would not exhibit this type of vulnerability (even more so with AES as there are specific CPU instructions that implement the inner loops of the algorithm making them highly resistant to side channel attacks like this).
And of course counter mode is not a csprng, but this has nothing to do with this. DRBGs, however can be based on block ciphers and they appear to be state of the art (unlike the well known backdoor-ed Dual-EC-DRBG).
Yours is the most useful comment on this fascinating story. Since you RTFA (sucka), what is the source of the CPU noise? What are its characteristics (frequency range, dynamic range) such that it can cut through other noises to register onto an at-distance audio mic?
BTW, I would have thought it bunk but for the name Adi Shamir in the author list, you know, the S of RSA.
The authors speculate that this is electro/magnetic-mechanical noise coming from the voltage regulators near the CPU (e.g., inductor "hum"). They note that it is hard to localize because the whole motherboard assembly is of course mechanically coupled together. The audio spectrum was characterized in a small range (35-39kHz) which is something that you can pick up with a standard microphone pickup. The dynamic range varied on different platforms, but in one case 0.5Hz separated the '0' and the '1' case.
So should academics use the same research and release standards with security bugs/exploits as government "hacks"?
The general consensus is that researchers that find vulnerabilities in commercial software at least notify the vendor and give them a chance to mitigate the problem before releasing a potential attack blue-print to the universe. Should researchers notify government entities before releasing their research on their "hacks", or are they to treat the government as a black-hat?
It's an interesting choice either way... One way you are colluding the other way you apparently might be considered a traitor (or hero depending on your politics).
> Spend the same $4 on fresh foods and they will go bad in days
Not a problem. Eat it in days.
I can buy the most expensive deli meats and organic buns and still come out ahead of McDonalds and Subway. Never mind the more mundane stuff and "real food".
It's a RESTAURANT. It has a 5x margin by it's very nature. It has to. Otherwise it could not possibly stay in business.
A 5x margin is easy to beat.
For every dollar a restaurant charges, they roughly average 33% food costs (and roughly 25-30% labor costs, the rest go to capital/equipment/rent/profit, etc). Fast food the labor costs are generally on the lower side (less staff per customer and lower wage scales). The profit margin per food category varies tremendously, with beverages being the highest by far and stuff like featured entrees being extremely low.
Since we are talking about fast food, for example, the dollar menu at McD's is nearly $0.90/item fully loaded cost, where a drink costs about $0.20 and they sell it for a dollar. The McDouble broke the bank (started about ~$0.96** and eventually was money losing) and they had to take it off the menu.
Of course if you are looking strictly at food, you might conclude you can do significantly better, but don't forget they have a bulk wholesale cost-of-food advantage too. Several estimates of the home cost (not including labor, or energy) of reproducing a McDouble range from about $1.25 to $2.00. But of course the home burger would be more tasty and probably healthier for you... Of course $2 is less that the so-called $6-burger at a quick-service restaurant.
**McD's originally wanted the McDouble to have 2 slices of cheese, but franchisees protested that would produce a loss so it was launched with 1 slice.
First off, there are a number of bios manufacturers...
Maybe in number, but not marketshare where there are basically 2: AMI and Phoenix/Award. The market share of all others is a rounding error.
Second, there are numerous bugs still existent.
True, but see point #1
This sounds like the NSA found someone in China using an exploit in a BIOS to hack computers. Alerted the manufacturer who was probably already aware of the fact after numerous Linux users had reported it years ago.
Probably likely, but not a consequence of your first two points.
You missed a few other open issues related to decommisioning (e.g., mostly what to do with the salt).
In any case, the only efforts I know of are:
FUJI which I think died in the fund-raising stage back in 2011.
TTS an attempt to resurrect this.
Thor Energy
MSRE showed that the physics worked, however, as with many things, the engineering problems remain. AFAIK, most people are attempting to figure out the salt problem. The metal problem is currently unsolved (and a much more important problem since you need the reactor to have a reasonable operating life to make the whole thing economical in the first place).
Acutally you can sometimes get a cell signal from a phone on a plane (some lightly used rural towers can pretty high power/range). Unfortunatly the high-altitude and velocity of a plane used to confuse cell networks (esp., the tower-handoff protocols between say a 3G and a CDMA tower on a multi-mode phone) and consume too much of the tower's antenna power-bandwidth envelope leaving less for other calls. Because so many folks leave their cell-phones onboard planes, nowdays many networks are configured detect if you are hitting too many different types of towers and simply refuse to track/connect a phone exhibiting such a pattern.
Of course at 5miles up in cruise phase, you'd have to have a very good phone to get a signal because the signal you would get isn't from a tower straight down from you (the cell towers don't focus their energy up), but on a far away tower where your signal looks more parallel to the ground. But during a slow decent phase over a relatively unpopulated area, it's quite common to be able to get a decent signal somewhere between 3-4 miles in altitude.
But they aren't going to do that, they are going to put a femto-cell on the plane and probably back-haul it through a satellite network (likely the same one they use to provide wifi internet service). I imagine the call will look like a high-cost roaming fee on your cell-phone bill and they will block VoIP (if they don't do that already).
The quiet section is already monetized. It's called "first class", that wonderful place where the screaming of infants and children is reduced to a dull howl.
Was flying in first class once and two seats were occupied with 2 parents and 2 lap-children. Even though it was a short haul flight (2 hours on a 737), I felt sorry for the people in the seats next to them (because of the availability of oxygen masks, lap-children were restricted to one per row so they couldn't put them next to each other in the same row).
The cost moderated the usage so much that they're mostly no longer enabled even if they're present. The remaining service will be shutting down completely at the end of this year.
And good riddance ;^)
Here is a thought. An aeroplane is a public place, and you should not necessarily expect quiet on a plane. What next, do you ban people talking to each other on planes because is disturbs your peace and quiet?
Why not buy some earplugs if you want to drown out the noise around you? Why not let people use their time as productively as they wish.
If it is a red eye flight, then I fully agree that people should be quiet at certain times. but to mandate it for all flights, whatever the time is ridiculous.
True story. I was on a flight having an animated conversation with someone (probably a bit louder than we thought) and the flight attendant came over and gave us both a free in-flight movies and a drink in an effort to shut us up (well she was quite disarming polite about it). Don't know if someone complained or if it was a pre-emptive move, but it was quite effective.
On the other hand, I've seen a passenger yelling at a poor parent that was trying to comfort a baby crying. The flight attendant offered to change that person's seat, but she refused and instead had to find 2 adjacent seat for the parent and his 2 kids (fwiw, I offered mine, but my seat-neighbors weren't willing, maybe they didn't want to sit next to that passenger after all of her yelling).
Personally, noise doesn't bother me too much, but yelling is much more distracting than a baby crying. I really hate the personal-space battles, though. It's the arm rest tug-o-war that really gets on my nerve. Also the person that want you to hold their drink or put it on your tray table whilst they type a quick note on their monster laptop and that quick note lasts 45-minutes as they go over a presenation and write a few emails (okay that only happened twice, but seriously?). Often, I just pull out a laptop and put it on the tray-table and do nothing just to avoid that situation in the future. As long as their productivity doesn't intrude on my physical space, maybe it's okay...