You are, I think, talking about the versions that already exist.
OTOH, I doubt that we will intentionally create an AI that desires to eliminate humans. The tricky thing will be creating AIs that are useful for all these other goals, like winning the next war, and not creating one (or possibly a pair, created by different combatants) that end up wiping out humanity.
IIRC it was sometime last year, or possibly the year before that, that I read about a place in Japan that was deploying totally automated security robots that could (under unspecified circumstances) kill people. Now that's not a very smart AI, but it's still an AI. And it can kill people (who invade it's turf?).
So this has already been reported as happening. (Was the report accurate? Was it just a proposal someone made? Who knows?)
Then there's the work that's being done for the army on target identification. So far none of those have officially been given weapons. And there's the "drone" operations which are telefactors rather than robots, but which would be easy enough to automate once the target identification groups get "good enough".
It seems to me that various groups have already made the decision to go ahead.
IIUC, they're talking about two groups of people who are disconnected.
Group 1 is people who save browsing history or links into a pocket. That gets shared.
Group 2 is people who are typing links into the browser, that gets suggested to from the database created by group 1.
OTOH, this seems to mean that you need to store locally all the most prominent links used by anyone. But it would fit my understanding of what they are proposing.
Perhaps that was designed a few years too early. The designs I saw looked a bit overly ambitious. More so than CERN. But perhaps CERNs super collider couldn't have been build without the preliminary work on superconducting magnets done for the Texas project.
Sometimes things really do need to be attempted several times, because of what you learn on the earlier attempts.
That, of course, is the problem. There's no real sign that the standard model needs correction, except that it appears incompatible with Relativity. Whoops. And doesn't handle gravity well. Whoops!
So we know it needs to be changed, but just how is a real question. Perhaps sterile neutrinos will show us a path forwards, but not unless someone discovers another way to detect them. Otherwise it's just another anomaly that doesn't lead anywhere.
There are lots of "things with names that we don't have handles for", but without a handle, the name is just name magic, it doesn't point to anything real. Anomalies show us places where we can tell that our theories are incomplete, but don't, in and of themselves, tell us how to make that useful.
Perhaps particle theory is all wrong, and everything is fields that sometimes look like particles. There was a time when I kept hearing about Quantum Field Theory, but I haven't heard much about it recently. Probably quarks made it too difficult to handle. But this doesn't mean it isn't the right approach. Of course, it also doesn't mean that it is.
What you're ignoring is that if they do create a black hole it will be moving above escape velocity, and will have such a small capture cross-section that it could sit harmlessly in the center of the Earth.
The question is, "Could it survive long enough without emitting too much Hawking radiation to be noticed?". I have heard some people speculate that there is a minimum size for black holes, below which they can neither emit Hawking radiation nor capture incoming material even if embedded within neutronium, but I sort of doubt that conjecture. OTOH, if that could exist, perhaps it could explain "dark matter". They'd need to have been created during the big bang to avoid disrupting the Lithium balance, but why not. They should have been a time back then when things were dense enough to seed them, and if they're that size, the seeds should have been sterile (i.e., neither able to evaporate nor to grow). But this depends on such a minimum size existing.
If it's from Adobe, I don't trust it. Back when fonts were just bit maps, I wouldn't have needed to trust them, and this wouldn't be important. These days though... Postscript was a Turing-complete computer language. I don't know about the languages used to program fonts today, but I'm going to assume that if it's from Adobe it can't be trusted. If I'm wrong, little is lost.
That probably true. The Fedora tests were more directly comparable. But if what you want is what Clear Linux offers, this tells you it's a bit faster.
OTOH, even for the Fedora tests, the desktop chosen is quite important. I would be surprised if xfce turned in the same results as Gnome for many of the tests. If nothing else, the memory use would be different. And I often use KDE which is a lot heavier, and would probably be even slower...on many tasks.
But if Clear Linux suits your needs, then this test shows you what speed advantage you would have (on this particular hardware, WHOOPS!).
It depends on what kind of performance interests you. It's not relevant to speed, but it's relevant to many use cases for a portable. It's also an area where Linux has frequently been reported to have problems. (I don't know about any of the other systems...and I suspect it may depend heavily on the chosen desktop, but that's just a suspicion. [An extremely reasonable suspicion, but I've never seen it measured.])
Microsoft's history is a lot more malign that that. They have a history of actively working to break the competition. For one example, look into the standardization of word processor file formats.
Actually, it's not that clear. Certain quantum processes have been applied successfully on reasonably large scale to Quantum Computing, but they aren't (or don't seem) sufficient to build a general quantum computer, but only a specialized variety that can handle some problems well, but can't touch others.
Other techniques have been shown to work in the lab, but getting the error rates under control has been quite a challenge, and nobody has proven that they can do this in a stable fashion.
Even then, for many problems it has not been shown that quantum computers provide any advantage. Perhaps they do, but this will require that new algorithms be developed, and until the computers are actually available, little effort is going to be expended in this direction. And maybe they don't exist. It's also possible that one *can't* get the error rates under control.
So it's possible that quantum computers have only a niche use, and can't really do things like Shor's algorithm in any practical case.
Nobody knows...but it general quantum computers are reasonably realizable, then they'll make current encryption essentially worthless...so there will be an important niche use for them even if no new algorithms get developed...and there's no reason to think they won't be.
This depends on your precise definition of Quantum Computer. D-Wave sells something that they reasonably call a Quantum Computer, but it's not a general quantum computer. And all their customers have been cagey about how effective it is.
That said, simulated annealing through quantum mechanics is reasonably called quantum computing. (I think I remember that that's what D-Wave claims to do.)
No, the market was not free during the 1960's. Not for legal goods, anyway.
I cannot think of a time or place where it can be shown that an actual free market existed in concert with some government. There was a free market in parts of the 1860-70's US, but that was because governmental control was absent. Prior to that the indigenous population exerted control over the economies, after that the US government exerted control. There was a period in between, however, when there was essentially no government economic control in large parts of that area. The death toll was high. Fraud was rampant. It was hardly anything ideal.
That said, heavy control over economies can be just as bad. All benefit is in the intermediate situations. The current problem is uneven control favoring those who are rich and powerful. The average degree of control seems pretty reasonable, but average is not what is experienced in any one transaction.
OK, here's someone who actually understands what the article was about. It's interesting and informative, but the short is, not enough info to decide about bio-synthesis: http://blogs.sciencemag.org/pi...
Well, one article linked above seemed to indicate that about all the determined was the melting, or possibly sublimation, temperature and possibly the mass. Some of the stuff they tested for "does it contain sulfur?", It also sounded like they tested, somehow, for carbon rings. This is wildly out of my area of expertise, and someone whose organic chem was a few decades closer than mine could probably have gotten a lot more out of it.
I wasn't questioning that UDP was faster, I was questioning the assumption that you could depend on it for reliable packet delivery. Your response indicates that you accept that you can't depend on it for reliable packet delivery, but the post I was responding to (was that you?) said that they depended on UDP reliably delivering packets...which is NOT it's designed use case.
These equations would predict the same experimental observation. That's what's meant by being compatible.
As a simple analogy, 4 + 2 = 6, but so does 3 + 3 = 6. The equations are different, but the underlying reality is the same. The thing is, in physics the underlying reality is subtler than we can directly detect, so even with choosing the simplest equations that fit the data we end up with multiple consistent interpretations. What this means nobody really knows, though you'll certainly get people who will tell you that this or that interpretation is correct or unreasonable.
Some people prefer to believe the simpler equation, so you get the heliocentric theories. Then it turns out that you need to recast those into the galactic central theories, but that's not really defensible either. Any coordinate system that matches the observations can have at least one set of equations fitted to it. But some are more difficult to work with than others. The ones with a stable fixed point on earth are easy for local values, but become increasingly difficult as you need to consider things more distant. However, if you do them correctly, they remain correct. It's just that epicycles were a gross oversimplification.
Somehow people tend to prefer the forms that are easier to calculate on. And there's nothing wrong with that. There are isomorphic transforms between them.
Now models that deform the earth into a form not topologically equivalent to a sphere are another case. I have a strong suspicion that those cannot be consistent models.
WHAT!!! Your system requires "ISP to deliver UDP packets reliably"??? UDP?? User Datagram Protocol??? Reliably?
UDP was designed for systems that don't need the packets delivered reliably! If you're requiring it to reliably delivered you are misusing the protocol. TCP is for reliably delivered packets. UDP doesn't even tell you if the packet wasn't delivered.
Nevertheless, he's correct about ocean current being pivotal in the climate and weather. What he's missing is that they don't change without cause. Yes, they've changed, and that has had a great impact on the weather. But their change was not some random happening, but has a multitude of causal antecedents, many of which are summed up in the phrase "global warming".
Slashdot Mirror
The theory about Hoover is wrong, because marijuana became illegal under federal law a lot earlier than the Black Panthers even existed.
OTOH, if you were to say that's why they started pushing enforcement of an existing law, you might have a point.
Not entirely. If you think so, you're living in a fool's paradise.
You are, I think, talking about the versions that already exist.
OTOH, I doubt that we will intentionally create an AI that desires to eliminate humans. The tricky thing will be creating AIs that are useful for all these other goals, like winning the next war, and not creating one (or possibly a pair, created by different combatants) that end up wiping out humanity.
IIRC it was sometime last year, or possibly the year before that, that I read about a place in Japan that was deploying totally automated security robots that could (under unspecified circumstances) kill people. Now that's not a very smart AI, but it's still an AI. And it can kill people (who invade it's turf?).
So this has already been reported as happening. (Was the report accurate? Was it just a proposal someone made? Who knows?)
Then there's the work that's being done for the army on target identification. So far none of those have officially been given weapons. And there's the "drone" operations which are telefactors rather than robots, but which would be easy enough to automate once the target identification groups get "good enough".
It seems to me that various groups have already made the decision to go ahead.
IIUC, they're talking about two groups of people who are disconnected.
Group 1 is people who save browsing history or links into a pocket. That gets shared.
Group 2 is people who are typing links into the browser, that gets suggested to from the database created by group 1.
OTOH, this seems to mean that you need to store locally all the most prominent links used by anyone. But it would fit my understanding of what they are proposing.
Perhaps that was designed a few years too early. The designs I saw looked a bit overly ambitious. More so than CERN. But perhaps CERNs super collider couldn't have been build without the preliminary work on superconducting magnets done for the Texas project.
Sometimes things really do need to be attempted several times, because of what you learn on the earlier attempts.
That, of course, is the problem. There's no real sign that the standard model needs correction, except that it appears incompatible with Relativity. Whoops. And doesn't handle gravity well. Whoops!
So we know it needs to be changed, but just how is a real question. Perhaps sterile neutrinos will show us a path forwards, but not unless someone discovers another way to detect them. Otherwise it's just another anomaly that doesn't lead anywhere.
There are lots of "things with names that we don't have handles for", but without a handle, the name is just name magic, it doesn't point to anything real. Anomalies show us places where we can tell that our theories are incomplete, but don't, in and of themselves, tell us how to make that useful.
Perhaps particle theory is all wrong, and everything is fields that sometimes look like particles. There was a time when I kept hearing about Quantum Field Theory, but I haven't heard much about it recently. Probably quarks made it too difficult to handle. But this doesn't mean it isn't the right approach. Of course, it also doesn't mean that it is.
What you're ignoring is that if they do create a black hole it will be moving above escape velocity, and will have such a small capture cross-section that it could sit harmlessly in the center of the Earth.
The question is, "Could it survive long enough without emitting too much Hawking radiation to be noticed?". I have heard some people speculate that there is a minimum size for black holes, below which they can neither emit Hawking radiation nor capture incoming material even if embedded within neutronium, but I sort of doubt that conjecture. OTOH, if that could exist, perhaps it could explain "dark matter". They'd need to have been created during the big bang to avoid disrupting the Lithium balance, but why not. They should have been a time back then when things were dense enough to seed them, and if they're that size, the seeds should have been sterile (i.e., neither able to evaporate nor to grow). But this depends on such a minimum size existing.
But what about all the other bobsmiths?
If it's from Adobe, I don't trust it.
Back when fonts were just bit maps, I wouldn't have needed to trust them, and this wouldn't be important. These days though... Postscript was a Turing-complete computer language. I don't know about the languages used to program fonts today, but I'm going to assume that if it's from Adobe it can't be trusted. If I'm wrong, little is lost.
That probably true. The Fedora tests were more directly comparable. But if what you want is what Clear Linux offers, this tells you it's a bit faster.
OTOH, even for the Fedora tests, the desktop chosen is quite important. I would be surprised if xfce turned in the same results as Gnome for many of the tests. If nothing else, the memory use would be different. And I often use KDE which is a lot heavier, and would probably be even slower...on many tasks.
But if Clear Linux suits your needs, then this test shows you what speed advantage you would have (on this particular hardware, WHOOPS!).
Benchmarks for general use cannot be perfect.
It depends on what kind of performance interests you. It's not relevant to speed, but it's relevant to many use cases for a portable. It's also an area where Linux has frequently been reported to have problems. (I don't know about any of the other systems...and I suspect it may depend heavily on the chosen desktop, but that's just a suspicion. [An extremely reasonable suspicion, but I've never seen it measured.])
Microsoft's history is a lot more malign that that. They have a history of actively working to break the competition. For one example, look into the standardization of word processor file formats.
The "Linux Foundation" has a history of promoting things that are bad for Linux users, though not necessarily for Linux companies.
Actually, it's not that clear. Certain quantum processes have been applied successfully on reasonably large scale to Quantum Computing, but they aren't (or don't seem) sufficient to build a general quantum computer, but only a specialized variety that can handle some problems well, but can't touch others.
Other techniques have been shown to work in the lab, but getting the error rates under control has been quite a challenge, and nobody has proven that they can do this in a stable fashion.
Even then, for many problems it has not been shown that quantum computers provide any advantage. Perhaps they do, but this will require that new algorithms be developed, and until the computers are actually available, little effort is going to be expended in this direction. And maybe they don't exist. It's also possible that one *can't* get the error rates under control.
So it's possible that quantum computers have only a niche use, and can't really do things like Shor's algorithm in any practical case.
Nobody knows...but it general quantum computers are reasonably realizable, then they'll make current encryption essentially worthless...so there will be an important niche use for them even if no new algorithms get developed...and there's no reason to think they won't be.
This depends on your precise definition of Quantum Computer. D-Wave sells something that they reasonably call a Quantum Computer, but it's not a general quantum computer. And all their customers have been cagey about how effective it is.
That said, simulated annealing through quantum mechanics is reasonably called quantum computing. (I think I remember that that's what D-Wave claims to do.)
No, the market was not free during the 1960's. Not for legal goods, anyway.
I cannot think of a time or place where it can be shown that an actual free market existed in concert with some government. There was a free market in parts of the 1860-70's US, but that was because governmental control was absent. Prior to that the indigenous population exerted control over the economies, after that the US government exerted control. There was a period in between, however, when there was essentially no government economic control in large parts of that area. The death toll was high. Fraud was rampant. It was hardly anything ideal.
That said, heavy control over economies can be just as bad. All benefit is in the intermediate situations. The current problem is uneven control favoring those who are rich and powerful. The average degree of control seems pretty reasonable, but average is not what is experienced in any one transaction.
It probably works. The question is at what price and for how long.
OK, here's someone who actually understands what the article was about. It's interesting and informative, but the short is, not enough info to decide about bio-synthesis:
http://blogs.sciencemag.org/pi...
Well, one article linked above seemed to indicate that about all the determined was the melting, or possibly sublimation, temperature and possibly the mass. Some of the stuff they tested for "does it contain sulfur?", It also sounded like they tested, somehow, for carbon rings. This is wildly out of my area of expertise, and someone whose organic chem was a few decades closer than mine could probably have gotten a lot more out of it.
The question is "What does 'complex organic matter' mean?" It's not just carbon, and you wouldn't find methane in a lake bed. It's a gas at Mars STP.
I wasn't questioning that UDP was faster, I was questioning the assumption that you could depend on it for reliable packet delivery. Your response indicates that you accept that you can't depend on it for reliable packet delivery, but the post I was responding to (was that you?) said that they depended on UDP reliably delivering packets...which is NOT it's designed use case.
These equations would predict the same experimental observation. That's what's meant by being compatible.
As a simple analogy, 4 + 2 = 6, but so does 3 + 3 = 6. The equations are different, but the underlying reality is the same. The thing is, in physics the underlying reality is subtler than we can directly detect, so even with choosing the simplest equations that fit the data we end up with multiple consistent interpretations. What this means nobody really knows, though you'll certainly get people who will tell you that this or that interpretation is correct or unreasonable.
Some people prefer to believe the simpler equation, so you get the heliocentric theories. Then it turns out that you need to recast those into the galactic central theories, but that's not really defensible either. Any coordinate system that matches the observations can have at least one set of equations fitted to it. But some are more difficult to work with than others. The ones with a stable fixed point on earth are easy for local values, but become increasingly difficult as you need to consider things more distant. However, if you do them correctly, they remain correct. It's just that epicycles were a gross oversimplification.
Somehow people tend to prefer the forms that are easier to calculate on. And there's nothing wrong with that. There are isomorphic transforms between them.
Now models that deform the earth into a form not topologically equivalent to a sphere are another case. I have a strong suspicion that those cannot be consistent models.
WHAT!!! Your system requires "ISP to deliver UDP packets reliably"??? UDP?? User Datagram Protocol??? Reliably?
UDP was designed for systems that don't need the packets delivered reliably! If you're requiring it to reliably delivered you are misusing the protocol. TCP is for reliably delivered packets. UDP doesn't even tell you if the packet wasn't delivered.
Nevertheless, he's correct about ocean current being pivotal in the climate and weather. What he's missing is that they don't change without cause. Yes, they've changed, and that has had a great impact on the weather. But their change was not some random happening, but has a multitude of causal antecedents, many of which are summed up in the phrase "global warming".