Slight clarification: if you invert the system using CPT invariance, the system may still look different. But all of the physical laws that describe its behavior will be unchanged. Obviously a person jumping into a pool and creating a splash looks very, very different from a bunch of water falling into a pool and pushing somebody out of it onto the side of the pool. But both situations follow the exact same laws of physics.
This brings up the whole question of the arrow of time: why do we experience a definite, fixed, unchangeable arrow of time if the fundamental laws don't have any such thing? That's really complicated and big subject, but if anybody reading this is curious, they may want to look up some of the stuff written by Sean Carroll, who has done a lot of work on the subject. He also has a number of YouTube videos.
The research itself is testing what is known as "CPT invariance" (you can see this stated clearly in the article link in the OP). A super short explanation is that CPT is charge-parity-time invariance. CPT invariance means if you flip the sign of the electric charge, swap parity (meaning exchange left with right), and reverse time, then the physical system will behave exactly as it did before. If CPT invariance were indeed violated, then that is one thing that could explain the asymmetry between matter and anti-matter.
But CPT invariance has never been the primary avenue of research for understanding the matter/anti-matter asymmetry. The general assumption that most theorists working in high-energy physics work with is that CPT invariance is absolute, and the matter/anti-matter asymmetry is explained by breaking CP invariance instaed. CP invariance means that if you flip the charge and reverse swap left with right, but don't reverse time, then the system will behave the same. Break that, and flipping charge and swapping left/right will change the behavior of the system.
So what this research really means is that the general assumption most theorists were working on was probably a solid foundation, and they can continue looking for the CP invariance they expected was there all along. But I guess that's not a very sensational headline.
It happens sometimes. My little brother did it when he was a young child (maybe 4). Had his thumbs on the prongs while he was plugging something in. Knocked him back, but he wasn't injured.
The real difference is that we only have 120V at the wall. At 240V, the probability of serious injury is much higher, which provides a good justification for plugs that are more difficult to touch.
Ugh, Schuko plugs are the worst. They fall out sooo easily. And there are about 3-4 different variants that are in use in various places in Europe. At least US plugs usually grip pretty well* and are strongly standardized.
* There are a couple of places I've been in in the US where the plugs often are rather loose. I have no idea why. But this doesn't happen in most places. In most places, they're snug pretty much all the time.
Yup. It takes quite a long time for new connector standards to become broadly adopted in the PC world. USB-C is still very new in most respects. As people pick up more USB-C devices, computer manufacturers will sell systems with more USB-C ports. As those ports reach broader use, they'll take more care to ensure that everything functions properly.
Desktop computers in general will continue a long, gradual decline over the next couple of decades as people find it easier and more convenient to use mobile devices instead.
Desktop computers will retain their usefulness mostly for work purposes. Even though many work functions will become quicker and easier on mobile, I can't see the keyboard or the desktop going away for any purpose that requires a noticeable amount of writing or programming. Especially programming.
But it may be pretty rare for people to bother with PCs in their homes in ten or twenty years.
The underlying science here says something quite different from the headline: it's a claim that quantum computing is fundamentally more efficient than classical computing. It has long been suspected that quantum computing wasn't necessarily more efficient: that the existence of an efficient quantum algorithm might actually imply an efficient classical algorithm must also exist. It's really, really difficult to prove that there is no efficient classical algorithm for a given problem.
What they tried to show is that a specific type of algorithm is fundamentally more efficient on a quantum computer, and can't be done efficiently on a classical computer. I don't know if they succeeded or not. From past experience, I know it's really really hard to prove this kind of thing definitively, so there may be many responses that shoot their conclusion down. But even if their conclusion is correct, it doesn't mean that our universe can't be simulated. If correct, it just means that the simulation must be performed by a quantum computer.
I wish him success. It'd be wonderful to have a new Tesla-like operation running (in the sense of a new purely-electric vehicle company). But the smart money is on this project utterly failing. There is a huge amount of technical and marketing expertise involved in designing something as complex as a car. If he's coming into this without involving a lot of people really experienced with all aspects of car development, the chances are really good that the project will be doomed to failure. Plus there's the whole manufacturing problem to tackle. Bringing a new car assembly line into production would be monumental, and even contracting with an existing manufacturer for this purpose would be extremely challenging (especially if the differences from existing car designs are substantial, as Dyson apparently wants to achieve).
And if the car is too different from existing designs, he's going to have a hell of a time convincing people to buy it.
I'd worry more about simple derailing than IEDs. Imagine, for example, a power interruption while the pod was undergoing some acceleration (e.g. going around a curve), resulting in the car leaving the rails and colliding with the exterior wall. Many of the hyperloop proposals involve the loop traveling above-ground near freeways, which opens up the possibility of vehicles crashing with the hyperloop wall (or supports, if it's sufficiently elevated), leading to a similar result.
It may be possible to create systems that limit the danger of such problems, but they will never go away entirely. And if such significant damage is caused to the track, it may cause the track to be offline for a significant stretch of time while it is repaired, leading to frequent service interruptions that make the entire prospect dubious at best.
I think they want Google to index their pages, to drive traffic from Google users to Yelp. But they don't want Google to be showing Yelp images to represent businesses on Google properties. I don't think robots.txt provides enough subtlety for this distinction.
Google and Yelp are direct rivals in the narrow field of searches/ratings for businesses, especially restaurants. Obviously Yelp is not a rival in many of the other areas that Google operates in.
Yup. The idea is a non-starter because it would require broad international agreement and massive funding. That's just not going to happen with such large uncertainties as to the potential downsides (and there will be downsides).
You're making the incorrect assumption that security patches only prevent you from doing certain things with your device. This is far from the case. By large, security patches are designed to prevent exploitation of your device by other actors. If, for example, you use your phone for banking or payments, you should be extremely motivated to ensure that you have minimized the possibility of anybody hijacking your device and gaining access to your money as a result.
I generally put Bitcoin in the same bucket as gold, whose value is also largely driven by speculation, and which also experiences very large price fluctuations.
Dollars are an extremely different beast, however. Dollars, and any other national fiat currency, are backed by their country's national bank, who manually adjusts the money supply to keep their value stable. This active management of the value of dollars prevents the sort of asset bubbles that gold and Bitcoin are subject to, because investors generally aren't foolish enough to bet against an actor with as much power to maintain the price of the currency as the central bank. The risks of a fiat currency are very, very different (and vastly lower) than the risks associated with Bitcoin or gold.
Primarily, the risks are that the forces driving inflation or deflation do not outpace the central bank's ability to correct. This can happen if the cost of servicing the national debt gets too high. Or if there is a deflationary shock that is just too large for the bank to correct for. There's also some risk that the central bank will change policy and choose a different inflation target (or a different target altogether). The first risk is tiny in most nations. The second risk is much higher, but is more a problem for the health of the economy than the value of the currency. The third risk is moderate, but the costs aren't very big.
Sort of. But most investments are tied to something tangible. Bitcoin's value is purely based upon speculation. Its price can and will drop precipitously, and then perhaps rise again once there's another press cycle. Because of this volatility, Bitcoin is utterly worthless as a currency. That means that its value will only be sustained so long as the illusion persists. Which could be a long time. Or it could crash so drastically in a few weeks that people abandon it forever. Nobody knows, which is why it is and will always be a gamble.
I agree. Also, semi trucks typically have much greater endurance requirements than consumer vehicles. Unless you have a truck that can drive for a full day (8-10 hours) on a single charge, I really doubt that many companies will want to bother with it at all.
Currently, I own a Chevy Volt, which I really like. It's got 40-60 miles of range on electric (depending upon weather), which is plenty for my work commute, and about 350 miles range on gas. If I'm on a long drive and running low on fuel, I can stop at a gas station, fill the tank in five minutes, and be on my way (the tank is about 9 gallons, so it takes very little time to fill). There is no electric-only vehicle around that offers that level of convenience: if it isn't long enough for the trip, I would have to wait, and in particular I'd have to plan my stops around where I could charge the vehicle.
With a semi truck, that problem becomes far worse, because now the amount of time spent is a matter of money for the shipping company. They really will not want a driver to drive for an hour and a half and then have to stop for 30 minutes to recharge the pack. And even 30 minutes may be generous because a semi truck is going to need a far bigger battery for the same mileage as a consumer vehicle.
I like the idea in the abstract of making fully-electric semi trucks, but it's really not going to be an easy sell.
This isn't a freedom of speech issue in that particular sense (at least not primarily). It's more about the fact that blocking a user makes it so that that user can no longer follow Trump's tweets. I don't think anybody contests that it's reasonable for Trump to be able to control how he is contacted. But he shouldn't be able to prevent people following him or linking his tweets if they want to. It would be perfectly reasonable (to me) if Twitter had a setting that allowed Trump to ignore certain users, but blocking also imposes on the user's access to public statements by Trump.
To me, it seems that if this fails it will be on the grounds that it's easy enough to work around a block by creating a sock puppet account and following him that way.
The check-in process taking too long is not something that happens everywhere. I've voted in California, New York, and Washington. In California and New York, the check-in process was just a few seconds. In Washington there was no check-in as I only had to drop the ballot in the ballot box and get back in my car.
These onerous check-ins are part of the problem. They don't secure the vote. They just provide a way to make it so that people of color have a more difficult time voting.
Yeah, no. It was a poorly-argued mishmash of unsubstantiated stereotypes and bullshit. It revealed an employee who is actively hostile to a large fraction of his coworkers, has a number of bigoted ideas, and a myopic, incorrect view of what it is a software engineer actually does.
I decided to see what people were saying about comparisons between the iPhone 7 and the Google Pixel phones in reviews online, which seems to me the most fair comparison. Every single review suggested both cameras are pretty similar, though most gave the edge to the Pixel, while recognizing that each had their strengths and weaknesses.
Essentially what you're saying is, "Well, we don't know. Therefore it's scary!"
What we do know is that we are exceptionally far away from an AI that can do anything like high-level planning. AI today is pattern matching. That's it. That's all it's designed for. It doesn't even resemble cognition in any meaningful sense. It can only operate in situations where there are extremely explicit goals. Human cognition, by contrast, operates in ambiguous situations with multiple poorly-defined and sometimes mutually-exclusive goals. AI just can't do that, and nothing is on the horizon that could allow AI to do that.
Why do you think the fact that there are self-driving cars has any bearing at all on this question?
A self-driving car uses AI for a relatively simple purpose: evaluating the configuration of the road immediately ahead, and any traffic or other obstacles. The AI in a self-driving car doesn't do any high-level planning at all.
So yeah, if a military organization built a fully-autonomous war machine, it'd probably go haywire at some point and shoot people because it's likely to make significant, unexpected errors. But what it won't do is engage in any high-level planning to stage anything like a takeover.
I don't think it's a lie at all, but headline unemployment can be misleading, especially in a period of protracted economic weakness (where many people are pushed out of the workforce and no longer counted in unemployment numbers).
The issue is that if you want to understand what's going on with the economy, you can't look at any one single number. The right thing to do is to look at a variety of numbers. With unemployment, for example, if it gets too low then the result should be rising wages, because employers are no longer to find willing workers. So if you see unemployment creeping down, but wages not rising, then that means that the unemployment number is hiding the real weakness in the job numbers.
I could imagine some situations where paid prioritization might be actively beneficial. Unfortunately, the possibilities for abuse are far too great to ignore.
Any positive effects from paid prioritization would stem not from some nebulous increase in speed, but from reduced interruptions due to congestion. I could easily imagine traffic being bifurcated into two categories, where if a router reaches capacity it ensures that traffic in one category takes priority so that it can continue with minimal disruption. It's possible to implement this in such a way that existing traffic isn't slowed, but only by adding new capacity. This kind of thing might be nice if, say, every customer had a certain percentage of their traffic designated as "fast lane", and they could choose which traffic to allocate. Low-latency, low-bandwidth operations like interactive gaming could benefit greatly if the bifurcation of traffic actually improved worst-case latency.
My worry is that what would happen if paid prioritization were to be allowed is that companies would feel free to use it as a means to force everybody onto the fast lane, and thus increase fees. Worse, it could dramatically worsen the Internet as small providers are priced out. I don't trust the current government to implement a rule that would prevent this outcome.
Slashdot Mirror
Slight clarification: if you invert the system using CPT invariance, the system may still look different. But all of the physical laws that describe its behavior will be unchanged. Obviously a person jumping into a pool and creating a splash looks very, very different from a bunch of water falling into a pool and pushing somebody out of it onto the side of the pool. But both situations follow the exact same laws of physics.
This brings up the whole question of the arrow of time: why do we experience a definite, fixed, unchangeable arrow of time if the fundamental laws don't have any such thing? That's really complicated and big subject, but if anybody reading this is curious, they may want to look up some of the stuff written by Sean Carroll, who has done a lot of work on the subject. He also has a number of YouTube videos.
The research itself is testing what is known as "CPT invariance" (you can see this stated clearly in the article link in the OP). A super short explanation is that CPT is charge-parity-time invariance. CPT invariance means if you flip the sign of the electric charge, swap parity (meaning exchange left with right), and reverse time, then the physical system will behave exactly as it did before. If CPT invariance were indeed violated, then that is one thing that could explain the asymmetry between matter and anti-matter.
But CPT invariance has never been the primary avenue of research for understanding the matter/anti-matter asymmetry. The general assumption that most theorists working in high-energy physics work with is that CPT invariance is absolute, and the matter/anti-matter asymmetry is explained by breaking CP invariance instaed. CP invariance means that if you flip the charge and reverse swap left with right, but don't reverse time, then the system will behave the same. Break that, and flipping charge and swapping left/right will change the behavior of the system.
So what this research really means is that the general assumption most theorists were working on was probably a solid foundation, and they can continue looking for the CP invariance they expected was there all along. But I guess that's not a very sensational headline.
It happens sometimes. My little brother did it when he was a young child (maybe 4). Had his thumbs on the prongs while he was plugging something in. Knocked him back, but he wasn't injured.
The real difference is that we only have 120V at the wall. At 240V, the probability of serious injury is much higher, which provides a good justification for plugs that are more difficult to touch.
Ugh, Schuko plugs are the worst. They fall out sooo easily. And there are about 3-4 different variants that are in use in various places in Europe. At least US plugs usually grip pretty well* and are strongly standardized.
* There are a couple of places I've been in in the US where the plugs often are rather loose. I have no idea why. But this doesn't happen in most places. In most places, they're snug pretty much all the time.
Yup. It takes quite a long time for new connector standards to become broadly adopted in the PC world. USB-C is still very new in most respects. As people pick up more USB-C devices, computer manufacturers will sell systems with more USB-C ports. As those ports reach broader use, they'll take more care to ensure that everything functions properly.
Desktop computers in general will continue a long, gradual decline over the next couple of decades as people find it easier and more convenient to use mobile devices instead.
Desktop computers will retain their usefulness mostly for work purposes. Even though many work functions will become quicker and easier on mobile, I can't see the keyboard or the desktop going away for any purpose that requires a noticeable amount of writing or programming. Especially programming.
But it may be pretty rare for people to bother with PCs in their homes in ten or twenty years.
The underlying science here says something quite different from the headline: it's a claim that quantum computing is fundamentally more efficient than classical computing. It has long been suspected that quantum computing wasn't necessarily more efficient: that the existence of an efficient quantum algorithm might actually imply an efficient classical algorithm must also exist. It's really, really difficult to prove that there is no efficient classical algorithm for a given problem.
What they tried to show is that a specific type of algorithm is fundamentally more efficient on a quantum computer, and can't be done efficiently on a classical computer. I don't know if they succeeded or not. From past experience, I know it's really really hard to prove this kind of thing definitively, so there may be many responses that shoot their conclusion down. But even if their conclusion is correct, it doesn't mean that our universe can't be simulated. If correct, it just means that the simulation must be performed by a quantum computer.
I wish him success. It'd be wonderful to have a new Tesla-like operation running (in the sense of a new purely-electric vehicle company). But the smart money is on this project utterly failing. There is a huge amount of technical and marketing expertise involved in designing something as complex as a car. If he's coming into this without involving a lot of people really experienced with all aspects of car development, the chances are really good that the project will be doomed to failure. Plus there's the whole manufacturing problem to tackle. Bringing a new car assembly line into production would be monumental, and even contracting with an existing manufacturer for this purpose would be extremely challenging (especially if the differences from existing car designs are substantial, as Dyson apparently wants to achieve).
And if the car is too different from existing designs, he's going to have a hell of a time convincing people to buy it.
I'd worry more about simple derailing than IEDs. Imagine, for example, a power interruption while the pod was undergoing some acceleration (e.g. going around a curve), resulting in the car leaving the rails and colliding with the exterior wall. Many of the hyperloop proposals involve the loop traveling above-ground near freeways, which opens up the possibility of vehicles crashing with the hyperloop wall (or supports, if it's sufficiently elevated), leading to a similar result.
It may be possible to create systems that limit the danger of such problems, but they will never go away entirely. And if such significant damage is caused to the track, it may cause the track to be offline for a significant stretch of time while it is repaired, leading to frequent service interruptions that make the entire prospect dubious at best.
I think they want Google to index their pages, to drive traffic from Google users to Yelp. But they don't want Google to be showing Yelp images to represent businesses on Google properties. I don't think robots.txt provides enough subtlety for this distinction.
Google and Yelp are direct rivals in the narrow field of searches/ratings for businesses, especially restaurants. Obviously Yelp is not a rival in many of the other areas that Google operates in.
Yup. The idea is a non-starter because it would require broad international agreement and massive funding. That's just not going to happen with such large uncertainties as to the potential downsides (and there will be downsides).
You're making the incorrect assumption that security patches only prevent you from doing certain things with your device. This is far from the case. By large, security patches are designed to prevent exploitation of your device by other actors. If, for example, you use your phone for banking or payments, you should be extremely motivated to ensure that you have minimized the possibility of anybody hijacking your device and gaining access to your money as a result.
I generally put Bitcoin in the same bucket as gold, whose value is also largely driven by speculation, and which also experiences very large price fluctuations.
Dollars are an extremely different beast, however. Dollars, and any other national fiat currency, are backed by their country's national bank, who manually adjusts the money supply to keep their value stable. This active management of the value of dollars prevents the sort of asset bubbles that gold and Bitcoin are subject to, because investors generally aren't foolish enough to bet against an actor with as much power to maintain the price of the currency as the central bank. The risks of a fiat currency are very, very different (and vastly lower) than the risks associated with Bitcoin or gold.
Primarily, the risks are that the forces driving inflation or deflation do not outpace the central bank's ability to correct. This can happen if the cost of servicing the national debt gets too high. Or if there is a deflationary shock that is just too large for the bank to correct for. There's also some risk that the central bank will change policy and choose a different inflation target (or a different target altogether). The first risk is tiny in most nations. The second risk is much higher, but is more a problem for the health of the economy than the value of the currency. The third risk is moderate, but the costs aren't very big.
Sort of. But most investments are tied to something tangible. Bitcoin's value is purely based upon speculation. Its price can and will drop precipitously, and then perhaps rise again once there's another press cycle. Because of this volatility, Bitcoin is utterly worthless as a currency. That means that its value will only be sustained so long as the illusion persists. Which could be a long time. Or it could crash so drastically in a few weeks that people abandon it forever. Nobody knows, which is why it is and will always be a gamble.
I agree. Also, semi trucks typically have much greater endurance requirements than consumer vehicles. Unless you have a truck that can drive for a full day (8-10 hours) on a single charge, I really doubt that many companies will want to bother with it at all.
Currently, I own a Chevy Volt, which I really like. It's got 40-60 miles of range on electric (depending upon weather), which is plenty for my work commute, and about 350 miles range on gas. If I'm on a long drive and running low on fuel, I can stop at a gas station, fill the tank in five minutes, and be on my way (the tank is about 9 gallons, so it takes very little time to fill). There is no electric-only vehicle around that offers that level of convenience: if it isn't long enough for the trip, I would have to wait, and in particular I'd have to plan my stops around where I could charge the vehicle.
With a semi truck, that problem becomes far worse, because now the amount of time spent is a matter of money for the shipping company. They really will not want a driver to drive for an hour and a half and then have to stop for 30 minutes to recharge the pack. And even 30 minutes may be generous because a semi truck is going to need a far bigger battery for the same mileage as a consumer vehicle.
I like the idea in the abstract of making fully-electric semi trucks, but it's really not going to be an easy sell.
This isn't a freedom of speech issue in that particular sense (at least not primarily). It's more about the fact that blocking a user makes it so that that user can no longer follow Trump's tweets. I don't think anybody contests that it's reasonable for Trump to be able to control how he is contacted. But he shouldn't be able to prevent people following him or linking his tweets if they want to. It would be perfectly reasonable (to me) if Twitter had a setting that allowed Trump to ignore certain users, but blocking also imposes on the user's access to public statements by Trump.
To me, it seems that if this fails it will be on the grounds that it's easy enough to work around a block by creating a sock puppet account and following him that way.
The check-in process taking too long is not something that happens everywhere. I've voted in California, New York, and Washington. In California and New York, the check-in process was just a few seconds. In Washington there was no check-in as I only had to drop the ballot in the ballot box and get back in my car.
These onerous check-ins are part of the problem. They don't secure the vote. They just provide a way to make it so that people of color have a more difficult time voting.
Yeah, no. It was a poorly-argued mishmash of unsubstantiated stereotypes and bullshit. It revealed an employee who is actively hostile to a large fraction of his coworkers, has a number of bigoted ideas, and a myopic, incorrect view of what it is a software engineer actually does.
I rather appreciated this ex-employee's response, which is spot-on
I decided to see what people were saying about comparisons between the iPhone 7 and the Google Pixel phones in reviews online, which seems to me the most fair comparison. Every single review suggested both cameras are pretty similar, though most gave the edge to the Pixel, while recognizing that each had their strengths and weaknesses.
Essentially what you're saying is, "Well, we don't know. Therefore it's scary!"
What we do know is that we are exceptionally far away from an AI that can do anything like high-level planning. AI today is pattern matching. That's it. That's all it's designed for. It doesn't even resemble cognition in any meaningful sense. It can only operate in situations where there are extremely explicit goals. Human cognition, by contrast, operates in ambiguous situations with multiple poorly-defined and sometimes mutually-exclusive goals. AI just can't do that, and nothing is on the horizon that could allow AI to do that.
Why do you think the fact that there are self-driving cars has any bearing at all on this question?
A self-driving car uses AI for a relatively simple purpose: evaluating the configuration of the road immediately ahead, and any traffic or other obstacles. The AI in a self-driving car doesn't do any high-level planning at all.
So yeah, if a military organization built a fully-autonomous war machine, it'd probably go haywire at some point and shoot people because it's likely to make significant, unexpected errors. But what it won't do is engage in any high-level planning to stage anything like a takeover.
No, Musk is talking about some imaginary future that has about as much connection to reality as Lord of the Rings.
I don't think it's a lie at all, but headline unemployment can be misleading, especially in a period of protracted economic weakness (where many people are pushed out of the workforce and no longer counted in unemployment numbers).
The issue is that if you want to understand what's going on with the economy, you can't look at any one single number. The right thing to do is to look at a variety of numbers. With unemployment, for example, if it gets too low then the result should be rising wages, because employers are no longer to find willing workers. So if you see unemployment creeping down, but wages not rising, then that means that the unemployment number is hiding the real weakness in the job numbers.
I could imagine some situations where paid prioritization might be actively beneficial. Unfortunately, the possibilities for abuse are far too great to ignore.
Any positive effects from paid prioritization would stem not from some nebulous increase in speed, but from reduced interruptions due to congestion. I could easily imagine traffic being bifurcated into two categories, where if a router reaches capacity it ensures that traffic in one category takes priority so that it can continue with minimal disruption. It's possible to implement this in such a way that existing traffic isn't slowed, but only by adding new capacity. This kind of thing might be nice if, say, every customer had a certain percentage of their traffic designated as "fast lane", and they could choose which traffic to allocate. Low-latency, low-bandwidth operations like interactive gaming could benefit greatly if the bifurcation of traffic actually improved worst-case latency.
My worry is that what would happen if paid prioritization were to be allowed is that companies would feel free to use it as a means to force everybody onto the fast lane, and thus increase fees. Worse, it could dramatically worsen the Internet as small providers are priced out. I don't trust the current government to implement a rule that would prevent this outcome.