I find it interesting that all the subjects were already HIV positive. It looks like this study only shows that it is ok to inject into a human, not that it does anything useful.
That is all Phase I testing is: identify a safe dosage range and screening for side effects... Phase II, they will be attempt to determine if it does anything useful.
There's likely no conspiracy, just a few crappy uncalibrated voting machines out there. Nothing like a hotly contested election to get the spotlight out on something that probably nets out to nothing in the end.
The democratic party isn't doing so great in SC. According to the wikipedia...
The South Carolina Democratic Party controls none of the statewide offices and holds the minority in both the South Carolina Senate and the South Carolina House of Representatives. Democrats hold one of the state's six U.S. House seats.
...so who cares if a FPGA core is a bit slower than dedicated hardware or ASICs...
FPGAs? Let's see. One of the top of the line FPGAs available today is the Virtex 7 from Xilinx. Virtex 7 is 20M gates, 21Kbits of ram, DDR 1866. A run-of-the-mill Core i5 is 550M gates, 4MB cache, DDR 2666. I'm not aware of an Intel compatible core available to be put in an FPGA, but for reference, the built-in 32-bit PowerPC core that Xilinx put into a few of their FPGAs ran about 500MHz. Core i5 runs ~3GHz.
The FPGA is probably gonna be a little bit more than a "bit slower"... Probably enough slower for most folks that it would indeed be ground-breaking...
Samsung is the biggest investment competitor to Intel in the chip market, right? [ http://tinyurl.com/samsungintel ] What does Apple need Intel for, give the guys at Samsung a call. What could go wrong?
Apple and Samsung are getting along so well these days... I'm sure the thought to buy chips from Samsung crossed a few minds at Apple.... NOT!
Yes to velcro cable ties. The supposedly reusable plastic cable ties are mostly just releaseable, not reusable. You can also staple/nail right through the vecro anywhere you want and don't have to deal with the nail-in plastic cable holders.
I don't think that phrase invokes the same idea as most of the folks on/. The "neuromorphic" algorithms they allude to are the kind that run on highly specialized hardware (e.g., this beast). This type of hardware really just works similarly to synapses (integrate & fire architecture). Of course you could simulate the algorithm on a more conventional processor, but it would probably lose much of it's low-power attribute.
FWIW, the algorithm they propose is attempt to identify objects that project up from the ground. To do this, they attempt to label parts of the image as obstacle (or not) taking a raw initial guess and filtering it with a pre-trained neural net (using some sort of adjacent region belief propagation technique).
I think they may have "cheated" a bit in that in some papers, they describe decomposing the image with oriented Gabor filters (edge orientation detectors), but they admit that this decompsition doesn't currently work well on their ultra-low-power computing platform.
I've had issues with resistive screens, but multi-touch capacitive screens have all been fine.
Just like any technology, once you learn how MT-capactive screens work you may find you trust them less. Basically, they are scanning across x/y positions and measuring blobs of capacitive bleed. Then they go through many stages of proprietary noise-removal, segmentation, and centroid computation algorithms to come up with the number of touches and their x/y positions. Until recently, all of this was done is some embedded microcontroller in the "touch" controller (where any tampering of the microcode for this touch controller would be nearly impossible to detect). In some more modern devices, to save money, some of this processing is done on the main CPU (even more prone to tampering).
Of course all this embedded software is perfect. Just yesterday, my 2yo son splashed a couple very small drops of water on the corner of the display and found that when he was playing angrybirds he couldn't pull the slingshot straight back and told me it was "broken". That's robust software for you.
Also nobody in their right mind would ever think of tampering with this stuff. Nobody.;^)
Of course no tech is perfect or tamper proof, but perhaps this is just a road to understanding that more work needs to be done on the best way to measure voter intent rather than concentrating on some tech vs other tech to count votes (after the measument has corrupted the voter intent).
Most humans have had a mom who, without any commercial "efficiency need", devoted years and years into raising a sensible human who is capable of responding to anything unexpected...
There are humans that didn't have moms, or humans that had moms that didn't raise sensible beings, and many human aren't particularly capable of responding to unexpected driving events (in the various mental states that they choose to drive), yet the state still issues these humans driver's licences, and people willingly get into cars with those humans as drivers.
The point of fear that many folks have about turning their lives over to machines is that folks don't feel that they can predict what they will do, or if they will be ultimatly responsible for the failures that occur as a result of the machines action. Some people had the same fears about "horseless carriages". It's not that these fears are unfounded, but sadly the risk is just the price of progress. Some folks jumped on those "horseless carriages" right away, others never did and treat them as the encarnation of the devil, but most folks waited until the tech was shown to be useful and affordable until they jumped in and as a result in the USA we have 8 cars for every 10 people...
Apple should be paying close attention to this. One day, people will demand "right to repair" for thier iDevice and Apple will be sitting where the car manufacturers are today...
That's why developers are typically hired in "at-will" jurisdictions. If California wasn't at-will, Silicon Valley would be elsewhere.
Although California being at-will, might be a minor contribution to silicon valley, it is probably not in the top 5 main reasons.
The number 1 reason for the location of Silicon Valley is probably California's comprehensive ban on no-compete clauses. This created a situation where people could change jobs easily and frequently and startup new companies easily. The history behind this California anomoly (passed way back in 1872) was part of it's gold-mining heritage. To avoid the civil disruptions of attempts to enforce no-compete clauses between gold miners and the firms holding competing gold claims that attempted to hire-off scare labor, California basically banned any form of no-compete clauses that didn't have to do with business transfer (e.g. sale of goodwill, mergers & acquisitions, and partnerships). This ancient piece of law enabled the Silicon Valley as we know it. Only prohibitions on disemination of trade-secrets are upheld, not employment.
The other reasons are likely the historical availability of cheap land in Santa Clara County, the proximity of top tech schools Cal and Stanford, the proximity of support schools (SJ State, + community collegies) to help staff companies, the casual attitudes of the West-Coast (socialization outside companies, less formality of business contacts, risk taking by starting new companies), and even the weather (relative to the east coast). Of course after Silicon Valley was started, you got other things as well (easier availablity of local start-up capital, serial entrepenuers, etc)
I think you are mistaken about Samsung and Sony. Sony has never really been a very big industrial company. Although it hit it big in a few big consumer electronics markets over time (transistor radios, TVs, cd-player, Playstation) and diversified into media holding, compared to the mega industrial companies in Japan, it never really was a big industrial player, say like Panasonic (Matsushita), JVC, or Mitsubishi. On the other side, Samsung has always been one of the big Korean Industrial giants (chaebols or family owned industrial conglomerates).
Samsung's goals have been simply to just dominate every market they enter (basically the war-cry of it's ex-Founder/Chairman Lee). I don't think they ever specifically looked at Sony as a company to emulate or even take-out, but just yet another company to trounce along the way. I don't know if you remember but back in early 2000 there was a wave of Samsung Trinitron TVs. As I understand it, this was because Sony needed a supplier of LCD for the transition to up-and-comming LCD-TV market and traded a licence to the (dying) Trinitron name and tube-technology to sweeeten things up for Samsung. If they wanted to, Samsung could have killed Sony right there.
Sure some Sony-Samsung rivalry makes a good fiction for a cable-tv miniseries, but doesn't really jive with reality.
FWIW, the NURP... at the end of the message simply identifies the carrier pigeons (NURP stands for National Union of Racing Pigeons). This probably has nothing to do with the message.
The 40 and 37 indicate the year of registration and TW194 and DK76 are the "serial numbers" of the pigeon.
Actually, now that I think about it some mroe, given the digraph letter frequencies, it's probably more likely a double playfair cipher (which was in common use at the time). The ADFGX-like cipher wouldn't have this type of letter frequency distribution...
The basic idea they have is somewhat novel, but requires some understanding on how modern output stage power amplifiers are designed.
As mentioned by many, a "C" class amplifier can be designed so it's pretty linear over a certain range, but with variable-envelope modulation schemes used in modern communications, they waste quite a bit of power.
On the other hand, classic pwm (pulse-width modulation) switch mode power amplifers (aka D-class amplifiers), allow trading pusle-width modulation for amplitude modulation at low frequencies (e.g., audio), but don't work at communications frequencies.
So, in order to get the best performance at low power, switch voltage power amplifiers have been the most commonly used technique. The simplistic polar technique splits the signal into phase and amplitude and sends the amplitude to the voltage switch and the phase to a variable phase delay element before the amplifier. As you might imagine this isn't that efficient.
The outphasing technique relies on math to split the signal into two constant amplitude phase modulated signals which when amplified separatly and non-linearly can be recombined into a linear amplification the two phase combine in the output (which is why it is called outphasing). The idea is that you have a few amplifiers (maybe only 2) at different voltages and you split the signal between these amplifiers and sum them up (usually with a passive wilkinson combiner). The trick is how you split the signal up in such a way that you can use the full range of each amplifier (not just the super-linear part in the middle), and still have them sum up in the antenna with the minimal amount of distortion (e.g., get linear behavior).
There are many ways to split the signals up for outphasing. Having the same amplitude for each of the two signals is the most common implementation of this and is known by the acronym LINC (linear amplfication with non-linear components). A generalization of this where the two signals aren't the same ampltitude (ML-LINC or multi-level LINC) is an obvious power saver (one amplfier wastes less power than the other) because to get a low level output with two signals of the same amplitude requires the two output to cancel each other out in phase, but can't be applied efficiently to all moduation techinques.
The technique proposed by these guys is yet another advance on ML-LINC (called AMO). The math is simply fancier as it tries to minimize the amount that the signals are just out-of-phase and cancelling each other (wasting power).
Unfortunatly, although the Navy continues to fund research into Polywell style fusion reactors, there are several big hurdles to overcome. The biggest ones (to me) are that the concept has unknown scaling constants (e.g, does a "big" version lose too much efficinecy), and they most expensive component (the magnets) are inside the reactor and get bombarded with radiation which creates and equally big material science headache as some of the alternate approaches.
Sudden advances in IT don't expose latent social problems. All the social problems I can think of were exposed long before any advances in IT. There have always been nosey neighbors, bullied kids, crank phone calls, voyers, stalkers, and perversions of every kind imaginable before there was anywhere near the current IT technology. Telephone operators listened to calls, voyers with telescopes looked through windows, nosey neighbors snooped through mail.
The problem with advances in IT are that now people who work for corporations have the ability to tempt the masses and apply technology to mixit up with the masses to serve their customers. By allowing IT into our lives (through social networking) we've created a new social problem for ourselves by deliberatly sharing information with someone that we don't know, and doesn't have a duty to serve us. Since we don't pay them, and we aren't customers (advertisers are the customers), they don't naturally have any kind of fidicual duty to us.
It's as if we handed a briefcase full of $100 bills to a stranger on the street to give to our neighbor and trusted the laws to make sure that they did it. Sure that stranger might do it today, but they might also tell a friend that we handed them a briefcase and it had a lot of money in it. I don't see how any theory of law protects us in this situation. Nobody is peeking through a window, nobody is listening to a phone call, nobody is opening up mail. Basically there is a briefcase. No milliseconds are necessary.
Is privacy worth anything to the typical person? I'm afraid the answer is no. Nearly everyone seems willing to turn over a phone number or address to a total stranger for a 1 in a billion chance for a stupid prize. Maybe that's not the society that you want to live in, but sadly, it is the society that you live in. The only way to change it is to change the masses mind about privacy (which is something they don't really care much about right now). Once the masses mind is changed, government will often follow, but aiming at the government as the way of attacking this is really just one person thinking they know better than the rest (which they may know better, but that's not a very populist approach and sometimes frowned upon).
Actually, no. Stock price is simply an instantaneous measure of the relative attractiveness of a stock vs it's peers.
Historically, potential attractiveness is measured by some folks as proportional to projected future earnings (PE ratio), but that is a created reality, not anything fundamental. Simply because others think that is an appropriate measure, it becomes a somewhat usable measure (in that it approximates the "attractiveness" utility function of people competing to buy the stock). In the '90's bubble, when net earnings per share for some hot companies weren't high enough to justify high prices via the PE metric, the stock papperazzi dug up an old out-of-favor metric called Price Earnings to Growth (PEG). By dividing by growth, the that made these low PE, high-growth companies suddenly appear more attractive. That didn't work for firms with negative net earnings, so they made up new ways to measure potential earnings (e.g., normalized revenue with sustainable margins). This just goes to show that the measures of attractiveness of a stock can and will change over time.
Another big factor in a stock price is the total amount of money being invested in the stock market. As more money pours in, stocks get boosted somewhat in proportion to their relative attractiveness, so even in projected future earnings are the same, the stock price will go up. As 401k and pension funds have more money to invest or if say bond interest rates hover at all time historic lows, more money will pour into the stock market. In this environment, stocks will go up regardless of changes in measures of earnings (same supply of stock, more demand results in higher prices). To help satisfy this demand more companies will issue stock (e.g., IPOs, secondary offerings, etc) to attempt to sastify demand.
Of course a stock price or and index has little to do with unemployment, but because of money flow pressures, a stock price does bear some relation to an index (which is a rough measure of money flow into a basket of stocks). If you believe that a stock is priced "efficiently", projected corporate profit is "built-in" to the stock price and thus it's relative attractiveness. Only if stocks "crush" or "miss" their projections, is there a forcing function to change attractiveness.
Although previous generation AMD used to be more flops/watt, the new generation of Kepler GPUs from Nvidia are quite a bit better than AMD's current generation (GCN / southern islands).
Well, if we are going with personal antecdotes, I've debated a few folks about evolution vs ID, and most "evolutists" are simply working off script themselves (often fanciful diversions about dinosaurs or great apes and neaderthals and fossil evidence), which they don't seem to have the foggiest idea about. I can summarize most of their arguments in the phrase "creationists are wrong, so I must be right". Unfortunatly, that's one of the biggest logical fallacies (certainly, both could be wrong).
I'd wager that many of the folks that support evolution are equally as ignorant about what the actual evidence for evolution is, or the logical chain of reasoning that supports it. Certainly not all are ignorant, but the vast majority.
BTW, I sort-of believe in evolution, but sometimes I just enjoy debating folks on the subject because many evolutionists are so passionate about their belief, yet so ignorant about how tenuous some of the evidence is. For instance, originally, Homo Habilis was thought to be the oldest direct ancestors to H. Sapiens, but more recent evidence make it appears that HH might be a "dead" branch of H. Erectus or H. Rudolfensis. Also all the evidence is pretty much based on cranial size on very a couple of skeletons and carbon dating of some surrounding rocks. There's no dna, or any direct evidence. As you might say in a court, all the evidence is circumstantial.
Also, now that we know more about how environmental stresses can change the expression of genes, all that bunk about the only thing that can explain adaptation is collections of "fit" mutations passed along generation is also under question. As we learn more about dna and gene expression in progeny, we may find that the simplified explanation of evolution that we teach must change in much the same way alchemy changed into chemistry as we discovered atoms, electron orbits, isomers, and covalent bonding...
As another aside, the evidence pretty much shows that the human ancestors pretty much all died out (~5000 or so about 100,000 years ago), but somehow miraculously survived to populate the whole world. Nobody can say for sure why this is true or if in fact we are actually decendents of 5000 space aliens (e.g, some superbeing buying more goldfish for the tank after they all died out, although I think that is highly unlikely).
Just because creation isn't likely correct by most stretches of the imagination, we are still woefully ignorant about our own orgins, or how what we currently call evolution might work to say that we really know what we are talking about. We are the cocky teenagers who think they know it all, but apparently do not.
Perhaps this will force mensa to change their requirements (under the threat of false advertising).
Assuming that you have to be in the top 2% to be a member of mensa and younger people getting higher IQ score, for someone with an IQ enough to clear the 98% level, after a while, it probably won't be good enough (their raw test score won't be in the top 2% anymore). So you'd have to score high enough so that during your life expectancy, you wouldn't fall below the 98% level. If the current trends hold, you might expect to need a pretty high score on an intelligence tests to rank in the top 2% for the duration of your life.
Strangely, although new members of Mensa would be getting smarter over time, the average intelligence of a Mensa member relative to the rest of the world would likely be going down over time (unless they kicked out older members that scored borderline in the past order to maintain their stated top 2% charter, or were growing the group fast enough to overcome this trend).
Food for thought? (not that it matters any iota, but just a curiosity)...
Slashdot Mirror
I find it interesting that all the subjects were already HIV positive. It looks like this study only shows that it is ok to inject into a human, not that it does anything useful.
That is all Phase I testing is: identify a safe dosage range and screening for side effects...
Phase II, they will be attempt to determine if it does anything useful.
http://en.wikipedia.org/wiki/Clinical_trial#Phases
FWIW, it apparently happens both ways...
http://www.krdo.com/news/Pueblo-GOP-Machines-switched-Romney-votes-to-Obama/-/417220/17252566/-/qpcqxr/-/index.html
There's likely no conspiracy, just a few crappy uncalibrated voting machines out there. Nothing like a hotly contested election to get the spotlight out on something that probably nets out to nothing in the end.
FYI: In House District 2 in South Carolina, apparantly no democrat registered to oppose incumbant Joe Wilson (yes he was the same person that shouted out "you lie").
The democratic party isn't doing so great in SC. According to the wikipedia...
The South Carolina Democratic Party controls none of the statewide offices and holds the minority in both the South Carolina Senate and the South Carolina House of Representatives. Democrats hold one of the state's six U.S. House seats.
...so who cares if a FPGA core is a bit slower than dedicated hardware or ASICs...
FPGAs? Let's see. One of the top of the line FPGAs available today is the Virtex 7 from Xilinx. Virtex 7 is 20M gates, 21Kbits of ram, DDR 1866. A run-of-the-mill Core i5 is 550M gates, 4MB cache, DDR 2666. I'm not aware of an Intel compatible core available to be put in an FPGA, but for reference, the built-in 32-bit PowerPC core that Xilinx put into a few of their FPGAs ran about 500MHz. Core i5 runs ~3GHz.
The FPGA is probably gonna be a little bit more than a "bit slower"... Probably enough slower for most folks that it would indeed be ground-breaking...
Samsung is the biggest investment competitor to Intel in the chip market, right? [ http://tinyurl.com/samsungintel ] What does Apple need Intel for, give the guys at Samsung a call. What could go wrong?
Apple and Samsung are getting along so well these days... I'm sure the thought to buy chips from Samsung crossed a few minds at Apple.... NOT!
http://www.reuters.com/article/2012/09/07/net-us-apple-samsung-supply-idUSBRE88603A20120907
http://www.zdnet.com/samsung-to-stop-providing-lcds-to-apple-7000006182/
Yes to velcro cable ties. The supposedly reusable plastic cable ties are mostly just releaseable, not reusable. You can also staple/nail right through the vecro anywhere you want and don't have to deal with the nail-in plastic cable holders.
I don't think that phrase invokes the same idea as most of the folks on /. The "neuromorphic" algorithms they allude to are the kind that run on highly specialized hardware (e.g., this beast). This type of hardware really just works similarly to synapses (integrate & fire architecture). Of course you could simulate the algorithm on a more conventional processor, but it would probably lose much of it's low-power attribute.
FWIW, the algorithm they propose is attempt to identify objects that project up from the ground. To do this, they attempt to label parts of the image as obstacle (or not) taking a raw initial guess and filtering it with a pre-trained neural net (using some sort of adjacent region belief propagation technique).
I think they may have "cheated" a bit in that in some papers, they describe decomposing the image with oriented Gabor filters (edge orientation detectors), but they admit that this decompsition doesn't currently work well on their ultra-low-power computing platform.
FYI: MAV=micro aerial vehicle
I've had issues with resistive screens, but multi-touch capacitive screens have all been fine.
Just like any technology, once you learn how MT-capactive screens work you may find you trust them less. Basically, they are scanning across x/y positions and measuring blobs of capacitive bleed. Then they go through many stages of proprietary noise-removal, segmentation, and centroid computation algorithms to come up with the number of touches and their x/y positions. Until recently, all of this was done is some embedded microcontroller in the "touch" controller (where any tampering of the microcode for this touch controller would be nearly impossible to detect). In some more modern devices, to save money, some of this processing is done on the main CPU (even more prone to tampering).
Of course all this embedded software is perfect. Just yesterday, my 2yo son splashed a couple very small drops of water on the corner of the display and found that when he was playing angrybirds he couldn't pull the slingshot straight back and told me it was "broken". That's robust software for you.
Also nobody in their right mind would ever think of tampering with this stuff. Nobody. ;^)
Of course no tech is perfect or tamper proof, but perhaps this is just a road to understanding that more work needs to be done on the best way to measure voter intent rather than concentrating on some tech vs other tech to count votes (after the measument has corrupted the voter intent).
Most humans have had a mom who, without any commercial "efficiency need", devoted years and years into raising a sensible human who is capable of responding to anything unexpected...
There are humans that didn't have moms, or humans that had moms that didn't raise sensible beings, and many human aren't particularly capable of responding to unexpected driving events (in the various mental states that they choose to drive), yet the state still issues these humans driver's licences, and people willingly get into cars with those humans as drivers.
The point of fear that many folks have about turning their lives over to machines is that folks don't feel that they can predict what they will do, or if they will be ultimatly responsible for the failures that occur as a result of the machines action. Some people had the same fears about "horseless carriages". It's not that these fears are unfounded, but sadly the risk is just the price of progress. Some folks jumped on those "horseless carriages" right away, others never did and treat them as the encarnation of the devil, but most folks waited until the tech was shown to be useful and affordable until they jumped in and as a result in the USA we have 8 cars for every 10 people...
Apple: All your ideas are belong to us.
What more to understand?
Apple should be paying close attention to this. One day, people will demand "right to repair" for thier iDevice and Apple will be sitting where the car manufacturers are today...
That's why developers are typically hired in "at-will" jurisdictions. If California wasn't at-will, Silicon Valley would be elsewhere.
Although California being at-will, might be a minor contribution to silicon valley, it is probably not in the top 5 main reasons.
The number 1 reason for the location of Silicon Valley is probably California's comprehensive ban on no-compete clauses. This created a situation where people could change jobs easily and frequently and startup new companies easily. The history behind this California anomoly (passed way back in 1872) was part of it's gold-mining heritage. To avoid the civil disruptions of attempts to enforce no-compete clauses between gold miners and the firms holding competing gold claims that attempted to hire-off scare labor, California basically banned any form of no-compete clauses that didn't have to do with business transfer (e.g. sale of goodwill, mergers & acquisitions, and partnerships). This ancient piece of law enabled the Silicon Valley as we know it. Only prohibitions on disemination of trade-secrets are upheld, not employment.
The other reasons are likely the historical availability of cheap land in Santa Clara County, the proximity of top tech schools Cal and Stanford, the proximity of support schools (SJ State, + community collegies) to help staff companies, the casual attitudes of the West-Coast (socialization outside companies, less formality of business contacts, risk taking by starting new companies), and even the weather (relative to the east coast). Of course after Silicon Valley was started, you got other things as well (easier availablity of local start-up capital, serial entrepenuers, etc)
I think you are mistaken about Samsung and Sony. Sony has never really been a very big industrial company. Although it hit it big in a few big consumer electronics markets over time (transistor radios, TVs, cd-player, Playstation) and diversified into media holding, compared to the mega industrial companies in Japan, it never really was a big industrial player, say like Panasonic (Matsushita), JVC, or Mitsubishi. On the other side, Samsung has always been one of the big Korean Industrial giants (chaebols or family owned industrial conglomerates).
Samsung's goals have been simply to just dominate every market they enter (basically the war-cry of it's ex-Founder/Chairman Lee). I don't think they ever specifically looked at Sony as a company to emulate or even take-out, but just yet another company to trounce along the way. I don't know if you remember but back in early 2000 there was a wave of Samsung Trinitron TVs. As I understand it, this was because Sony needed a supplier of LCD for the transition to up-and-comming LCD-TV market and traded a licence to the (dying) Trinitron name and tube-technology to sweeeten things up for Samsung. If they wanted to, Samsung could have killed Sony right there.
Sure some Sony-Samsung rivalry makes a good fiction for a cable-tv miniseries, but doesn't really jive with reality.
FWIW, the NURP... at the end of the message simply identifies the carrier pigeons (NURP stands for National Union of Racing Pigeons).
This probably has nothing to do with the message.
The 40 and 37 indicate the year of registration and TW194 and DK76 are the "serial numbers" of the pigeon.
Actually, now that I think about it some mroe, given the digraph letter frequencies, it's probably more likely a double playfair cipher (which was in common use at the time). The ADFGX-like cipher wouldn't have this type of letter frequency distribution...
This is most likely a ADFGX-style cipher given the 5-character length units.
The basic idea they have is somewhat novel, but requires some understanding on how modern output stage power amplifiers are designed.
As mentioned by many, a "C" class amplifier can be designed so it's pretty linear over a certain range, but with variable-envelope modulation schemes used in modern communications, they waste quite a bit of power.
On the other hand, classic pwm (pulse-width modulation) switch mode power amplifers (aka D-class amplifiers), allow trading pusle-width modulation for amplitude modulation at low frequencies (e.g., audio), but don't work at communications frequencies.
So, in order to get the best performance at low power, switch voltage power amplifiers have been the most commonly used technique. The simplistic polar technique splits the signal into phase and amplitude and sends the amplitude to the voltage switch and the phase to a variable phase delay element before the amplifier. As you might imagine this isn't that efficient.
The outphasing technique relies on math to split the signal into two constant amplitude phase modulated signals which when amplified separatly and non-linearly can be recombined into a linear amplification the two phase combine in the output (which is why it is called outphasing). The idea is that you have a few amplifiers (maybe only 2) at different voltages and you split the signal between these amplifiers and sum them up (usually with a passive wilkinson combiner). The trick is how you split the signal up in such a way that you can use the full range of each amplifier (not just the super-linear part in the middle), and still have them sum up in the antenna with the minimal amount of distortion (e.g., get linear behavior).
There are many ways to split the signals up for outphasing. Having the same amplitude for each of the two signals is the most common implementation of this and is known by the acronym LINC (linear amplfication with non-linear components). A generalization of this where the two signals aren't the same ampltitude (ML-LINC or multi-level LINC) is an obvious power saver (one amplfier wastes less power than the other) because to get a low level output with two signals of the same amplitude requires the two output to cancel each other out in phase, but can't be applied efficiently to all moduation techinques.
The technique proposed by these guys is yet another advance on ML-LINC (called AMO). The math is simply fancier as it tries to minimize the amount that the signals are just out-of-phase and cancelling each other (wasting power).
You must be talking about those EMC2 folks...
Unfortunatly, although the Navy continues to fund research into Polywell style fusion reactors, there are several big hurdles to overcome. The biggest ones (to me) are that the concept has unknown scaling constants (e.g, does a "big" version lose too much efficinecy), and they most expensive component (the magnets) are inside the reactor and get bombarded with radiation which creates and equally big material science headache as some of the alternate approaches.
Read more about it here...
I can think of no technology which has comparable levels of continued failure.
I can, it's call the NIF...
Sudden advances in IT don't expose latent social problems. All the social problems I can think of were exposed long before any advances in IT. There have always been nosey neighbors, bullied kids, crank phone calls, voyers, stalkers, and perversions of every kind imaginable before there was anywhere near the current IT technology. Telephone operators listened to calls, voyers with telescopes looked through windows, nosey neighbors snooped through mail.
The problem with advances in IT are that now people who work for corporations have the ability to tempt the masses and apply technology to mixit up with the masses to serve their customers. By allowing IT into our lives (through social networking) we've created a new social problem for ourselves by deliberatly sharing information with someone that we don't know, and doesn't have a duty to serve us. Since we don't pay them, and we aren't customers (advertisers are the customers), they don't naturally have any kind of fidicual duty to us.
It's as if we handed a briefcase full of $100 bills to a stranger on the street to give to our neighbor and trusted the laws to make sure that they did it. Sure that stranger might do it today, but they might also tell a friend that we handed them a briefcase and it had a lot of money in it. I don't see how any theory of law protects us in this situation. Nobody is peeking through a window, nobody is listening to a phone call, nobody is opening up mail. Basically there is a briefcase. No milliseconds are necessary.
Is privacy worth anything to the typical person? I'm afraid the answer is no. Nearly everyone seems willing to turn over a phone number or address to a total stranger for a 1 in a billion chance for a stupid prize. Maybe that's not the society that you want to live in, but sadly, it is the society that you live in. The only way to change it is to change the masses mind about privacy (which is something they don't really care much about right now). Once the masses mind is changed, government will often follow, but aiming at the government as the way of attacking this is really just one person thinking they know better than the rest (which they may know better, but that's not a very populist approach and sometimes frowned upon).
Anyhow...
So now we know what Jim Keller is back at AMD to do...
Actually, no. Stock price is simply an instantaneous measure of the relative attractiveness of a stock vs it's peers.
Historically, potential attractiveness is measured by some folks as proportional to projected future earnings (PE ratio), but that is a created reality, not anything fundamental. Simply because others think that is an appropriate measure, it becomes a somewhat usable measure (in that it approximates the "attractiveness" utility function of people competing to buy the stock). In the '90's bubble, when net earnings per share for some hot companies weren't high enough to justify high prices via the PE metric, the stock papperazzi dug up an old out-of-favor metric called Price Earnings to Growth (PEG). By dividing by growth, the that made these low PE, high-growth companies suddenly appear more attractive. That didn't work for firms with negative net earnings, so they made up new ways to measure potential earnings (e.g., normalized revenue with sustainable margins). This just goes to show that the measures of attractiveness of a stock can and will change over time.
Another big factor in a stock price is the total amount of money being invested in the stock market. As more money pours in, stocks get boosted somewhat in proportion to their relative attractiveness, so even in projected future earnings are the same, the stock price will go up. As 401k and pension funds have more money to invest or if say bond interest rates hover at all time historic lows, more money will pour into the stock market. In this environment, stocks will go up regardless of changes in measures of earnings (same supply of stock, more demand results in higher prices). To help satisfy this demand more companies will issue stock (e.g., IPOs, secondary offerings, etc) to attempt to sastify demand.
Of course a stock price or and index has little to do with unemployment, but because of money flow pressures, a stock price does bear some relation to an index (which is a rough measure of money flow into a basket of stocks). If you believe that a stock is priced "efficiently", projected corporate profit is "built-in" to the stock price and thus it's relative attractiveness. Only if stocks "crush" or "miss" their projections, is there a forcing function to change attractiveness.
Although previous generation AMD used to be more flops/watt, the new generation of Kepler GPUs from Nvidia are quite a bit better than AMD's current generation (GCN / southern islands).
http://hothardware.com/Reviews/NVIDIA-GeForce-GTX-680-Review-Kepler-Debuts/?page=15
FYI: The K20 used in the Titan system are Kepler based.
Well, if we are going with personal antecdotes, I've debated a few folks about evolution vs ID, and most "evolutists" are simply working off script themselves (often fanciful diversions about dinosaurs or great apes and neaderthals and fossil evidence), which they don't seem to have the foggiest idea about. I can summarize most of their arguments in the phrase "creationists are wrong, so I must be right". Unfortunatly, that's one of the biggest logical fallacies (certainly, both could be wrong).
I'd wager that many of the folks that support evolution are equally as ignorant about what the actual evidence for evolution is, or the logical chain of reasoning that supports it. Certainly not all are ignorant, but the vast majority.
BTW, I sort-of believe in evolution, but sometimes I just enjoy debating folks on the subject because many evolutionists are so passionate about their belief, yet so ignorant about how tenuous some of the evidence is. For instance, originally, Homo Habilis was thought to be the oldest direct ancestors to H. Sapiens, but more recent evidence make it appears that HH might be a "dead" branch of H. Erectus or H. Rudolfensis. Also all the evidence is pretty much based on cranial size on very a couple of skeletons and carbon dating of some surrounding rocks. There's no dna, or any direct evidence. As you might say in a court, all the evidence is circumstantial.
Also, now that we know more about how environmental stresses can change the expression of genes, all that bunk about the only thing that can explain adaptation is collections of "fit" mutations passed along generation is also under question. As we learn more about dna and gene expression in progeny, we may find that the simplified explanation of evolution that we teach must change in much the same way alchemy changed into chemistry as we discovered atoms, electron orbits, isomers, and covalent bonding...
As another aside, the evidence pretty much shows that the human ancestors pretty much all died out (~5000 or so about 100,000 years ago), but somehow miraculously survived to populate the whole world. Nobody can say for sure why this is true or if in fact we are actually decendents of 5000 space aliens (e.g, some superbeing buying more goldfish for the tank after they all died out, although I think that is highly unlikely).
Just because creation isn't likely correct by most stretches of the imagination, we are still woefully ignorant about our own orgins, or how what we currently call evolution might work to say that we really know what we are talking about. We are the cocky teenagers who think they know it all, but apparently do not.
Perhaps this will force mensa to change their requirements (under the threat of false advertising).
Assuming that you have to be in the top 2% to be a member of mensa and younger people getting higher IQ score, for someone with an IQ enough to clear the 98% level, after a while, it probably won't be good enough (their raw test score won't be in the top 2% anymore). So you'd have to score high enough so that during your life expectancy, you wouldn't fall below the 98% level. If the current trends hold, you might expect to need a pretty high score on an intelligence tests to rank in the top 2% for the duration of your life.
Strangely, although new members of Mensa would be getting smarter over time, the average intelligence of a Mensa member relative to the rest of the world would likely be going down over time (unless they kicked out older members that scored borderline in the past order to maintain their stated top 2% charter, or were growing the group fast enough to overcome this trend).
Food for thought? (not that it matters any iota, but just a curiosity)...