Define faster. The single core performance may only increase by a few percent but the number of cores keeps increasing. So if your algorithm can use multiple cores it will be faster if not then it won't.
Not really. We have come to an obstacle that requires a change in tactic. Processing power is still increasing at Moore's law but only for multi-core applications. Once we adapt to the new paradigm where to double your performance you need to double the number of cores, things will pick up again. However, this is a really hard change to make and it is going to take some time to adapt.
Just as in a particle phenomenology class, you should learn Standard Model as if it is correct
No you should not. You should learn it as the best, most accurate model of particle physics that we have, how we have tested it and what the limitations of it are. We are talking about a university-level education, not force feeding information at high school. It is extremely important that, near the edges of our knowledge, we do not teach things as "correct" but as our best current understanding and the data and observations that lead us to that conclusion.
It really comes down to this: you can't replace a paradigm—no matter how flawed—with nothing.
I don't - I let the students themselves replace it with their own developed intuition about QM. The Copenhagen interpretation is exactly that - an interpretation. It is not a model or a theory. It makes no predictions and there are no observables. Its only purpose is to provide a framework to try and map our physical intuition of the human-scale physical world onto the quantum world. As such it is a useful crutch to help students get an initial grip on QM by providing a bridge but ultimately, as they become more familiar with QM, they have to discard this crutch and replace it with the intuition they have developed themselves. There is no replacement for it because, ultimately, it is trying to do something that cannot work: you cannot understand QM by analogy with the large-scale world because there are quantum phenomena which have no consistent classical analogy.
Overall you seem to have a school-level approach to teaching physics, not a university-level one. University physics students are generally intelligent, motivated people. They are not scared by learning that there are somethings that we do not yet fully understand nor are they incapable of developing their own models and understanding of phenomena. Physics professors are not some modern-day Moses coming down from the mountain with the Standard Model lagrangian carved into a stone tablet telling everyone that this is how the world works. We are people who piece together models and then spend ages trying to test them to destruction so we can learn how to make a better model and imporove our understanding of the world. It is important that we expose students to the limits of our understanding and how we got to those limits because some of those we teach are going to be the ones who may eventually expand our understanding further.
The problem is that in the past the growth used to apply to both the company and the consumer: the company sold new gadgets and so grew and the gadgets had new functionality which meant that their usefulness to the consumer grew. Now though it seems that there is very little growth in capability particularly from Apple. New iPhones have faster processors, nicer screens and better cameras but no really new functionality: they are just epsilon better than the previous model with fancier fonts and animations.
Well, the original Copenhagen thing referred to an "observer" and everybody jumped to the absurd conclusion that it must be a conscious human observer.
No, nobody jumped to that conclusion they were just left without any guidance as to what an observer was, or was not, and started trying to make guesses that had some vague semi-consistency. You can indeed extend observers to include cats and geiger counters. But why did you stop there? Both are made of atoms so can atoms themselves count as observers? If so everything is observed all the time and, if the Copenhagen interpretation was right, QM would cease to work.
So perhaps not atoms but then where between a geiger counter and an atom do you draw the line? We can make silicon devices at the tens of nanometre scale that can detect charged particles like a geiger counter so do these count? This is why the notion of an observer is a complete non-starter: there is no consistent definition of one.
If you believe that, you haven't taken a single course in quantum mechanics.
Not only have I taken some courses in QM I now give them! The Copenhagen interpretation is taught in the same manner than the Bohr atom is taught: it is a useful toy model to get students used to some of the ideas of QM but ultimately it is clearly wrong as, indeed, Schrodinger's Cat successfully showed.
Making more things here makes *more* sense with the tariffs in place than without
No it does not because you have forgotten that the tariffs are bi-directional. When the US imposes tariffs on goods the countries affected impose tariffs on US goods. If the rest of the world outside the US continues to remove trade barriers then because the non-US combined economy is far larger than the US one you will end up being far better off making your things outside the US than inside because you can then sell them without tariffs to far more people.
The whole point of Schrodinger's cat experiment [phys.org] was to show that trying to apply certain quantum physics theories to reality resulted in absurd results.
No, it is more subtle than that. It was designed to show that one interpretation of the results of QM was wrong by showing that it leads to an absurd explanation for every-day scale objects like cats. Nobody ever believed that the cat was in some weird superposition: that was indeed the entire point. The interpretation of QM, called the Copenhagen interpretation, was clearly wrong which is why nobody believes it today. However, everyone believes in quantum mechanics itself and that it works when describing reality (it's the second most precisely tested scientific theory that has ever existed). The problem is trying to get brains that are used to a world that works in the large-scale limit of QM to really grasp the rather different underlying reality.
Rubbish. In my field we use Comic Sans for our most important discoveries...but that is because we are more interested in the information than the font it is written in.
If a prosecutor or judge uses an algorithm to set sentencing or determine parole, the individual prosecutor or judge should still be held accountable if he was in error.
If the algorithm that the law requires them to use gives an inappropriate result then it is not the judge or prosecutor who is to blame but those who passed the law requiring that they use the algorithm. This is the same for all algorithms: someone, somewhere has made a decision to use the algorithm and that decision makes that person accountable.
Even for more complex system which may design or adapt their own algorithms someone, somewhere has decided to put them in charge of something and that's where the accountability lies.
So you cannot legally restrict advertising based on these and other criteria.
Yes, but suppose you advertize to those with an interest in car maintenance (which purely for the sake of this argument let's assume is a male-dominated group)? The effect is that you would be targetting the ad to a predominantly male group but is this illegal gender discrimination?
If you are doing this for a job working as a car mechanic this seems like a very reasonable thing to do. However, if your job is for an investment banker your motivation is likely to be illegal gender discrimination. This makes the situation complex and, for Facebook, rather challenging since I doubt there is any algorithm that can possibly decide whether targetting a job ad to a specific group is valid or invalid discrimination because it depends heavily on context.
I can't believe some people are really going to defend something like a job posting site offering the ability to employers to say "I only want men to know about this job."
If that is what they are doing then there is absolutely no argument at all. However, I suspect that they are doing something more nuanced and, rather than just selecting "men-only" they are selecting to display the ads to people with certain interests and then selecting interests that are biased towards men.
This makes the situation a lot less clear. If the interest group targetted has an interest related to the job being advertized then this is reasonable discrimination regardless of the gender balance of the group. However, if employers are selecting groups only because of their gender balance then this is clearly wrong since it is outright gender discrimination.
This makes the problem a complex one: whether a particular targeting approach is legal depends heavily on the context. However, in today's world of black-and-white politics you can forget any sort of reasoned, sensible response to this. One side will want to ban it outright and the other will want to allow everything.
It's an economic disadvantage for the UK if it doesn't stop using DST.
It's a massive economic disadvantage for the UK if it doesn't stop Brexit. There is no sign of that happening and I doubt the effect of DST will be noticeable on top of that.
It's worse than that because the UK will never be able to do this now without appearing to be following the EU lead which the brexiteers will deem completely unacceptable. The UK is probably now stuck with GMT and BST for the foreseeable future.
You should never go on a cruise to get financial planning advice: you go on the cruise after the financial planning advice you got from somewhere reputable pays off.
In fairness, all that stuff is only half on him...
You are missing a part though that is entirely on him. If he wants to really help low-income people a good place to start would be by paying his employees a living wage instead of a minimum wage. His attitude smacks of the old Victorian factory owners in the UK who would make tons of money off the back of their workers while paying them a pittance and would then turn around with their profits and fund "charitable" initiatives e.g. decent housing which came with additional requirements such as no drinking, regular church attendance etc. to get them to behave as the factory owner thought they should behave.
The issue is the fact that these new jobs require less general education and more focused education.
That's largely not true. The attitude of society has changed considerably but the actual knowledge required has not (although technology means that the details have altered e.g. we no longer show people how to look up stuff in a library - all science papers are now online). Companies used to value university graduates with a deep, broad understanding of a particular subject because these people can draw on their education to handle new situations and come up with innovative and different solutions to problems and opportunities as they arise.
However, such people have some initial, up-front training costs for the precise job that you want them to do plus, with a broader range of skills, they can more easily jump ship to someone else to do a different job if you treat them like crap. With modern companies increasingly focused on short-term profits to fuel the CEO's bonus, investing in long-term human resource development so that you have a high quality, well-educated workforce that will keep the company operating and growing for the next several decades is not a priority. As long as the company lasts as long as the CEO's contract it's fine. Interestingly the companies that eshew this short-termism seem to be the ones that are now doing really well...
I think the point is if you imagine a photon to travel from A to B and then to C, and thing P happens at A and Q at B, and detection of state at C, then you expect the events to happen PQ.
You only expect this if you don't have a rudimentary understanding of QM. Wave-particle duality means that a photon can be at both A and B at the same time just like a wave can hit the beach of a range of times and places. Given that which event happens first is no longer clear at all.
Really it means that the concept of motion is different at the quantum level.
Not at all. The photon is not behaving as a point-like particle here where it passes through A and then B. Think of it as a wave which can be at both A and B at the same time because of its finite size.
In the classical limit the probabilities do go to extremes: either a lot, lot closer to 100% than 99.99% or incredibly close to zero. That is why classical physics works. However, in the quantum limit the probabilities are nowhere near so cut and dried and so you cannot approximate them to 0% or 100% and the system becomes highly non-deterministic.
The point is though that, at a fundamental level (as we understand it currently) the universe is non-deterministic. Which is a very good thing if you believe in free-will since, while a non-deterministic universe allows for the possibility of free-will (but does not guarantee it) a deterministic universe completely rules it out.
This seems to be a similar, but more complex, approach to boosted decision trees where the training samples that the algorithm initially mis-classifies are fed back through with a higher weight to make the algorithm pay more attention.
You got drunk because you were going to break the table.
This and the summary fails to understand QM properly. In QM the time of an event is not well determined. The smaller the energy change associated with an event the less well determined the time of the event.
To think of a classically equivalent question it would be like asking exactly when does a wave hit a beach? The wave has a finite size and hits the beach over a range of time as it comes in at a slight angle to the shore so it is impossible to pick on exact instant and say that is exactly when the wave hit the beach.
Hence the summary is wrong to say that A then B were performed on the particle. The time of A and B is not well defined and so it is literally impossible to say whether A or B occurred first: the time of both events overlapped.
Why do we assume we only move forward in time? We move forward and backward in all the other dimensions, why assume we aren't oscillating back and forth in time? What difference would that make if we were? How would we know?
Physics is not invariant under time reversal and we can unambiguously determine if time were suddenly reversed. Note that this is not related to entropy e.g. a shattered plate would not leap off the floor and reassemble itself if time were reversed this is far more fundamental than that.
Oscillations of kaon and B mesons have been shown to violate time-reversal symmetry so, if the "direction" of time were suddenly reversed we would know this because these oscillations would have the opposite bias.
There is a relativistic quantum behaviour similar to what you describe - called zitterbewegung - but only in space. Essentially electrons can be shown to hypothetically be travelling back and forth at the speed of light but direction biased so that the net movement is in the classical electron's direction at the classical speed.
Slashdot Mirror
No, we can't. That is the point.
Define faster. The single core performance may only increase by a few percent but the number of cores keeps increasing. So if your algorithm can use multiple cores it will be faster if not then it won't.
We have hit that wall.
Not really. We have come to an obstacle that requires a change in tactic. Processing power is still increasing at Moore's law but only for multi-core applications. Once we adapt to the new paradigm where to double your performance you need to double the number of cores, things will pick up again. However, this is a really hard change to make and it is going to take some time to adapt.
Just as in a particle phenomenology class, you should learn Standard Model as if it is correct
No you should not. You should learn it as the best, most accurate model of particle physics that we have, how we have tested it and what the limitations of it are. We are talking about a university-level education, not force feeding information at high school. It is extremely important that, near the edges of our knowledge, we do not teach things as "correct" but as our best current understanding and the data and observations that lead us to that conclusion.
It really comes down to this: you can't replace a paradigm—no matter how flawed—with nothing.
I don't - I let the students themselves replace it with their own developed intuition about QM. The Copenhagen interpretation is exactly that - an interpretation. It is not a model or a theory. It makes no predictions and there are no observables. Its only purpose is to provide a framework to try and map our physical intuition of the human-scale physical world onto the quantum world. As such it is a useful crutch to help students get an initial grip on QM by providing a bridge but ultimately, as they become more familiar with QM, they have to discard this crutch and replace it with the intuition they have developed themselves. There is no replacement for it because, ultimately, it is trying to do something that cannot work: you cannot understand QM by analogy with the large-scale world because there are quantum phenomena which have no consistent classical analogy.
Overall you seem to have a school-level approach to teaching physics, not a university-level one. University physics students are generally intelligent, motivated people. They are not scared by learning that there are somethings that we do not yet fully understand nor are they incapable of developing their own models and understanding of phenomena. Physics professors are not some modern-day Moses coming down from the mountain with the Standard Model lagrangian carved into a stone tablet telling everyone that this is how the world works. We are people who piece together models and then spend ages trying to test them to destruction so we can learn how to make a better model and imporove our understanding of the world. It is important that we expose students to the limits of our understanding and how we got to those limits because some of those we teach are going to be the ones who may eventually expand our understanding further.
The problem is that in the past the growth used to apply to both the company and the consumer: the company sold new gadgets and so grew and the gadgets had new functionality which meant that their usefulness to the consumer grew. Now though it seems that there is very little growth in capability particularly from Apple. New iPhones have faster processors, nicer screens and better cameras but no really new functionality: they are just epsilon better than the previous model with fancier fonts and animations.
Well, the original Copenhagen thing referred to an "observer" and everybody jumped to the absurd conclusion that it must be a conscious human observer.
No, nobody jumped to that conclusion they were just left without any guidance as to what an observer was, or was not, and started trying to make guesses that had some vague semi-consistency. You can indeed extend observers to include cats and geiger counters. But why did you stop there? Both are made of atoms so can atoms themselves count as observers? If so everything is observed all the time and, if the Copenhagen interpretation was right, QM would cease to work.
So perhaps not atoms but then where between a geiger counter and an atom do you draw the line? We can make silicon devices at the tens of nanometre scale that can detect charged particles like a geiger counter so do these count? This is why the notion of an observer is a complete non-starter: there is no consistent definition of one.
If you believe that, you haven't taken a single course in quantum mechanics.
Not only have I taken some courses in QM I now give them! The Copenhagen interpretation is taught in the same manner than the Bohr atom is taught: it is a useful toy model to get students used to some of the ideas of QM but ultimately it is clearly wrong as, indeed, Schrodinger's Cat successfully showed.
Making more things here makes *more* sense with the tariffs in place than without
No it does not because you have forgotten that the tariffs are bi-directional. When the US imposes tariffs on goods the countries affected impose tariffs on US goods. If the rest of the world outside the US continues to remove trade barriers then because the non-US combined economy is far larger than the US one you will end up being far better off making your things outside the US than inside because you can then sell them without tariffs to far more people.
The whole point of Schrodinger's cat experiment [phys.org] was to show that trying to apply certain quantum physics theories to reality resulted in absurd results.
No, it is more subtle than that. It was designed to show that one interpretation of the results of QM was wrong by showing that it leads to an absurd explanation for every-day scale objects like cats. Nobody ever believed that the cat was in some weird superposition: that was indeed the entire point. The interpretation of QM, called the Copenhagen interpretation, was clearly wrong which is why nobody believes it today. However, everyone believes in quantum mechanics itself and that it works when describing reality (it's the second most precisely tested scientific theory that has ever existed). The problem is trying to get brains that are used to a world that works in the large-scale limit of QM to really grasp the rather different underlying reality.
The academic fonts are Computer Modern.
Rubbish. In my field we use Comic Sans for our most important discoveries...but that is because we are more interested in the information than the font it is written in.
When I see these car free articles my first though is, "what about maintenance and repair?"
In a pedestrianized city? You are clearly talking a load of old cobblers.
the only time in human history when people in power are held to account is when they were lynched by a mob.
If a prosecutor or judge uses an algorithm to set sentencing or determine parole, the individual prosecutor or judge should still be held accountable if he was in error.
If the algorithm that the law requires them to use gives an inappropriate result then it is not the judge or prosecutor who is to blame but those who passed the law requiring that they use the algorithm. This is the same for all algorithms: someone, somewhere has made a decision to use the algorithm and that decision makes that person accountable.
Even for more complex system which may design or adapt their own algorithms someone, somewhere has decided to put them in charge of something and that's where the accountability lies.
So you cannot legally restrict advertising based on these and other criteria.
Yes, but suppose you advertize to those with an interest in car maintenance (which purely for the sake of this argument let's assume is a male-dominated group)? The effect is that you would be targetting the ad to a predominantly male group but is this illegal gender discrimination?
If you are doing this for a job working as a car mechanic this seems like a very reasonable thing to do. However, if your job is for an investment banker your motivation is likely to be illegal gender discrimination. This makes the situation complex and, for Facebook, rather challenging since I doubt there is any algorithm that can possibly decide whether targetting a job ad to a specific group is valid or invalid discrimination because it depends heavily on context.
I can't believe some people are really going to defend something like a job posting site offering the ability to employers to say "I only want men to know about this job."
If that is what they are doing then there is absolutely no argument at all. However, I suspect that they are doing something more nuanced and, rather than just selecting "men-only" they are selecting to display the ads to people with certain interests and then selecting interests that are biased towards men.
This makes the situation a lot less clear. If the interest group targetted has an interest related to the job being advertized then this is reasonable discrimination regardless of the gender balance of the group. However, if employers are selecting groups only because of their gender balance then this is clearly wrong since it is outright gender discrimination.
This makes the problem a complex one: whether a particular targeting approach is legal depends heavily on the context. However, in today's world of black-and-white politics you can forget any sort of reasoned, sensible response to this. One side will want to ban it outright and the other will want to allow everything.
It's an economic disadvantage for the UK if it doesn't stop using DST.
It's a massive economic disadvantage for the UK if it doesn't stop Brexit. There is no sign of that happening and I doubt the effect of DST will be noticeable on top of that.
It's worse than that because the UK will never be able to do this now without appearing to be following the EU lead which the brexiteers will deem completely unacceptable. The UK is probably now stuck with GMT and BST for the foreseeable future.
You should never go on a cruise to get financial planning advice: you go on the cruise after the financial planning advice you got from somewhere reputable pays off.
In fairness, all that stuff is only half on him...
You are missing a part though that is entirely on him. If he wants to really help low-income people a good place to start would be by paying his employees a living wage instead of a minimum wage. His attitude smacks of the old Victorian factory owners in the UK who would make tons of money off the back of their workers while paying them a pittance and would then turn around with their profits and fund "charitable" initiatives e.g. decent housing which came with additional requirements such as no drinking, regular church attendance etc. to get them to behave as the factory owner thought they should behave.
The issue is the fact that these new jobs require less general education and more focused education.
That's largely not true. The attitude of society has changed considerably but the actual knowledge required has not (although technology means that the details have altered e.g. we no longer show people how to look up stuff in a library - all science papers are now online). Companies used to value university graduates with a deep, broad understanding of a particular subject because these people can draw on their education to handle new situations and come up with innovative and different solutions to problems and opportunities as they arise.
However, such people have some initial, up-front training costs for the precise job that you want them to do plus, with a broader range of skills, they can more easily jump ship to someone else to do a different job if you treat them like crap. With modern companies increasingly focused on short-term profits to fuel the CEO's bonus, investing in long-term human resource development so that you have a high quality, well-educated workforce that will keep the company operating and growing for the next several decades is not a priority. As long as the company lasts as long as the CEO's contract it's fine. Interestingly the companies that eshew this short-termism seem to be the ones that are now doing really well...
I think the point is if you imagine a photon to travel from A to B and then to C, and thing P happens at A and Q at B, and detection of state at C, then you expect the events to happen PQ.
You only expect this if you don't have a rudimentary understanding of QM. Wave-particle duality means that a photon can be at both A and B at the same time just like a wave can hit the beach of a range of times and places. Given that which event happens first is no longer clear at all.
Really it means that the concept of motion is different at the quantum level.
Not at all. The photon is not behaving as a point-like particle here where it passes through A and then B. Think of it as a wave which can be at both A and B at the same time because of its finite size.
In the classical limit the probabilities do go to extremes: either a lot, lot closer to 100% than 99.99% or incredibly close to zero. That is why classical physics works. However, in the quantum limit the probabilities are nowhere near so cut and dried and so you cannot approximate them to 0% or 100% and the system becomes highly non-deterministic.
The point is though that, at a fundamental level (as we understand it currently) the universe is non-deterministic. Which is a very good thing if you believe in free-will since, while a non-deterministic universe allows for the possibility of free-will (but does not guarantee it) a deterministic universe completely rules it out.
This seems to be a similar, but more complex, approach to boosted decision trees where the training samples that the algorithm initially mis-classifies are fed back through with a higher weight to make the algorithm pay more attention.
You got drunk because you were going to break the table.
This and the summary fails to understand QM properly. In QM the time of an event is not well determined. The smaller the energy change associated with an event the less well determined the time of the event.
To think of a classically equivalent question it would be like asking exactly when does a wave hit a beach? The wave has a finite size and hits the beach over a range of time as it comes in at a slight angle to the shore so it is impossible to pick on exact instant and say that is exactly when the wave hit the beach.
Hence the summary is wrong to say that A then B were performed on the particle. The time of A and B is not well defined and so it is literally impossible to say whether A or B occurred first: the time of both events overlapped.
The implications to Determinism should be interesting
Not really because there is no determinism in quantum mechanics, only probabilities of outcomes.
Why do we assume we only move forward in time? We move forward and backward in all the other dimensions, why assume we aren't oscillating back and forth in time? What difference would that make if we were? How would we know?
Physics is not invariant under time reversal and we can unambiguously determine if time were suddenly reversed. Note that this is not related to entropy e.g. a shattered plate would not leap off the floor and reassemble itself if time were reversed this is far more fundamental than that.
Oscillations of kaon and B mesons have been shown to violate time-reversal symmetry so, if the "direction" of time were suddenly reversed we would know this because these oscillations would have the opposite bias.
There is a relativistic quantum behaviour similar to what you describe - called zitterbewegung - but only in space. Essentially electrons can be shown to hypothetically be travelling back and forth at the speed of light but direction biased so that the net movement is in the classical electron's direction at the classical speed.