Huh? Maybe my explanation was lacking. The bomb is either there or not there, and not in a quantum way. There is a bomb, or no bomb, with the same status in BOTH universes. The only difference between the universes is the path the photon takes. And in fact the device ALWAYS gives the right answer. The only uncertainty comes into play when there IS a bomb: then it is uncertain whether it will explode in this universe, or another one. Regardless, we will detect the event. And we can make the probability it explodes in this universe arbitrarily small by biasing the photon to take the bomb path with arbitrarily small probability.
This is not really an abuse of terminology. The fact is that in OUR possible world, the quantum computer was not EVER TOUCHED by a single input photon. If it were a photon-activated bomb, it would not have been set off. Sure, the computer was run somewhere in our MULTIVERSE (in another possible world)... but saying it was not run in this world is fairly correct (your definition of universe may vary). See my comment above.
You don't seem to understand the issue fully... See the "bomb testing" link provided above for more info. I will briefly summarize:
Say you know with 50% probability that there may be a bomb in some (very dark) position. The problem is that if even a single photon comes in contact with the bomb, it will set it off, taking out you and your lab assistants. You would like to detect whether the bomb is there, without setting it off. This is where interaction-free measurement (e.g. counterfactual computation) comes into play.
Now send a photon through a beam splitter, such that half goes through the location where the bomb may be and half goes thorugh ordinary space. Then if a bomb exists, you will set if off with 50% probability (constituting a "measurement", although you won't live to see it). The kicker is that in all other cases, you can still determine whether a bomb was there. By effectively doing a "double slit" experiment recombining the two paths, the presence of an interference pattern means that no bomb was there, and the lack of it means that the bomb WAS there (but you DID NOT actually physically interact with it).
The easiest way to interpret this result is using the many-worlds interpretation. Basically, you can detect the bomb (with 50% probability) when it in fact explodes in another possible world. You can try to "merge" that world back with this one, which will work iff no bomb was set off there (please forgive my oversimplicfications), a detectable fact. By tweaking the experiment, you can make this work over an arbitrary number of possible worlds, with the bomb going off in exactly one. Thus, you can detect the bomb while making the probability of physically interacting with it in ANY WAY arbitrarily small.
Re:IPods are the only reason why Apple still exist
on
Woz On Apple's Success
·
· Score: 1
It's not for teenage girls who want a "cute" computer (although that was my little sis's reason for going Apple). Here in the computer science department at Berkeley (and in other scientific disciplines) I would say about half the students have switched to Macs (largely in the last year or two). At the coffee shop down the street from the CS building, I see more P/iBooks than all other laptops combined. I figured there must be a reason, so I myself switched a few weeks ago (after a lifetime of professing my hatred for Macs)... the Unix-based OS X, I think, is one of the main selling points.
We can't find out. From WP: Like all forms of U.S. Public Diplomacy (propaganda), its broadcast is forbidden in the U.S. itself under the 1948 Smith-Mundt Act.
What goes up must come down. There's only a finite amount of oil, and so production must start decreasing at sp,e point. Obligatory Wikipedia link for lots more info on peak oil theory.
I have, in fact, read several neuroscience textbooks. Of course, in an approach like the one you suggest you would have to include the computations involved in integrating all the inputs of a single neuron in deciding to produce a single pulse, which could be substantial... this is the figure I was referring to. A simpler thought experiment: when computers hit the atomic scale, they will almost certianly be more efficient in terms of computations per unit volume X time than the brain is (the brain needs a lot of "support" hardware, due to biological limitations). Once we get there, a brain-sized hunk of the stuff would certainly be more than powerful enough to simulate a brain (considerations of programming the thing aside).
That makes the brain by far the most complicated structure in the known universe.
More complex than a group of monkeys? Or human society? Or how about the sun? Don't get me wrong, the brain is pretty damn complex. But OTOH, direct numerical simulations of the brain should be reaching feasibility in a couple decades (based on current estimates of the brain's pure processing power).
Somewhat offtopic, but dammit children's access to internet porn is NOT the same as child porn . Please stop perpetuating this misinformation. Thank you.
...But when the effort of doing so exceeds the reasonable effort of walking into the library and scanning the entirety of said literary work, I would contend that Google has met it's burden...
The analogy breaks down when someone goes to copy a second book. Presumably, a second trip to the copy machine at the library will take just as long as the first. OTOH, once a hole is discovered in Google's book protection, it could be scripted/otherwise automated such that downloading their entire catalog takes only a single click. When I want to crack the CSS on a DVD, I don't have to reinvent the wheel... I just do what any script kiddie does best. Now some kind of argument could be made about Google's due diligence to quickly find and close such holes once they are opened... I'm just saying that the issue isn't as simple as your comparison might make it seem.
BTW, I'm completely in support of Google Print, just playing devil's advocate here.
I tried this out today, and I'm sold. After seeing stats on how Opera is significantly faster than Firefox in almost every category, I finally decided to check it out. While I miss one or two extensions (Bugmenot and Forecast Firefox), I can do without these. Other than that, the built-in mouse gestures, keyboard + location bar shortcuts, ad block, torrents, better download manager, fast forward (hit the button or press ctrl-x and automatically go to the next page of google search results, next part of any article,...), and so on means that out of the box it is a firefox killer, and much faster to boot.
You've never programmed extensively in Javascript/DTML, I take it. I haven't done any serious web design in 5 or 6 years, but at least back then browser compatbility was a huge headache (each browser & version had a slightly different object model that had to be taken into account explicitly to make things work and degrade cleanly).
I think the use of the word "fuel" is just downright confusing, since although the stuff being referred to is expended during acceleration, it is not itself providing any actual energy. It is just arbitrary junk being thrown off the back of the spaceship to speed it up (ala Newton's 3rd law), with the potential difference presumably coming from some other energy-producing substance that should more accurately be referred to as the fuel.
The Cell won't be terribly well suited for AI either
I don't see why not. Each vector processor should be particularly good at, e.g., evaluating or training a single neural net. And while it is true that most AI algorithms today are not suited to parallelism, don't forget that most games have a bunch of stuff going on at once (maybe 10 enemies on screen at a time that need AI, plus physics, graphics, etc). The game itself is inherently parallelizable, so the AI algorithms don't have to be.
What, and Microsoft's model is much better? They release software that SHOULD be in beta as a final release, and then deal with the million remaining vulnerabilities, etc. later. IMHO at least Google is being honest about the status of their products.
Exactly.... I'm sure improvements to the interface will be coming as they get more of this kind of feedback. As for the paucity of content, remember that anyone can sell a video on video.google.com (with google taking 30%? of the fee). Thus, it is in their best interest to launch the store as early as possible, to entice more copyright owners to sell their videos through their service. This is in stark contrast (I assume) to the model taken by, i.e., iTunes, where content is solicited from a few large corporations.
Since all of the other answers to your ? seem a bit off, I'll take a shot at it. Qubits can be used as ordinary bits with no problem, by simply avoiding superpositions of states. The issue is a practical, not theoretical one; why would you go to all the trouble of lasers and ion traps yaddy yadda to build a 10 bit (i.e. only 10 computing elements) computer that runs at perhaps 1000 Hz, when you could easily build a chip with 10 billion transistors clocked at 100 GHz instead (ballpark figures for today's chips). It will be a long time, if ever, before a quantum computer would also be a practial, efficient classical computer. The gain from QC comes in from using special-purpose quantum algorithms such as Shor's algorithm (exponential speedup over classical computers specific to the factoring problem) or Grover's algorithm (quadratic speedup on all NP complete problems), plus more algorithms yet to be discovered (the field is fairly new). On an ordinary classical algorithm, a decent QC will perform at the level of a classical computer from before I was born.
I think people around here forget that DRM is not inherently evil (although perhaps it is inherently flawed). The problem we have with DRM is that content producers use it to impose unnecessary requirements (your e-book expires in 3 months, you can play this song on your computer but not your ipod, etc.). Keep in mind that the content producers get to decide what requirements they want, and would refuse to release most current works without some form of DRM. Google is not setting the rules here; they're just trying to stay in the game.
By this point, I think availability is growing faster than the body of useful knowledge. Even if the total amount of available information has doubled in the last 20 years, new search technologies make it 1000 times faster to find what you want (approximately, of course). While TFA talks about emerging technologies like del.ico.us and personalized search, I think the real boom is still to come, in the form of real AI.
When computers are fast enough and new algorithms are developed to really harness this power (I give it 10 years, give or take, for this to begin), computers will finally be able to at least have the semblance of understanding the body of knowledge rather than just syntactically sifting through it. This will give us another order of magnitude change comparable to the introduction of search technologies in the first place. Imagine being able to ask google "in one paragraph, summarize the most influential inventions of 2015". Not the most interesting or illuminating example, but you get the idea.
Ummm... I'll guess you just did, along with a million other Slashdotters. Given that this started making the rounds a couple months ago IIRC, and has been linked from many mainstream news stories, I'm sure he's seen a lot of traffic. Who cares where the visitors came from once they're there... there's not even any content to distract from the ads! I'm sure a year from now nobody will visit the stupid page, but I think a few hundred bucks for a few million page views is not a bad deal in the advertising business.
Slashdot Mirror
Huh? Maybe my explanation was lacking. The bomb is either there or not there, and not in a quantum way. There is a bomb, or no bomb, with the same status in BOTH universes. The only difference between the universes is the path the photon takes. And in fact the device ALWAYS gives the right answer. The only uncertainty comes into play when there IS a bomb: then it is uncertain whether it will explode in this universe, or another one. Regardless, we will detect the event. And we can make the probability it explodes in this universe arbitrarily small by biasing the photon to take the bomb path with arbitrarily small probability.
This is not really an abuse of terminology. The fact is that in OUR possible world, the quantum computer was not EVER TOUCHED by a single input photon. If it were a photon-activated bomb, it would not have been set off. Sure, the computer was run somewhere in our MULTIVERSE (in another possible world) ... but saying it was not run in this world is fairly correct (your definition of universe may vary). See my comment above.
You don't seem to understand the issue fully ... See the "bomb testing" link provided above for more info. I will briefly summarize:
Say you know with 50% probability that there may be a bomb in some (very dark) position. The problem is that if even a single photon comes in contact with the bomb, it will set it off, taking out you and your lab assistants. You would like to detect whether the bomb is there, without setting it off. This is where interaction-free measurement (e.g. counterfactual computation) comes into play.
Now send a photon through a beam splitter, such that half goes through the location where the bomb may be and half goes thorugh ordinary space. Then if a bomb exists, you will set if off with 50% probability (constituting a "measurement", although you won't live to see it). The kicker is that in all other cases, you can still determine whether a bomb was there. By effectively doing a "double slit" experiment recombining the two paths, the presence of an interference pattern means that no bomb was there, and the lack of it means that the bomb WAS there (but you DID NOT actually physically interact with it).
The easiest way to interpret this result is using the many-worlds interpretation. Basically, you can detect the bomb (with 50% probability) when it in fact explodes in another possible world. You can try to "merge" that world back with this one, which will work iff no bomb was set off there (please forgive my oversimplicfications), a detectable fact. By tweaking the experiment, you can make this work over an arbitrary number of possible worlds, with the bomb going off in exactly one. Thus, you can detect the bomb while making the probability of physically interacting with it in ANY WAY arbitrarily small.
It's not for teenage girls who want a "cute" computer (although that was my little sis's reason for going Apple). Here in the computer science department at Berkeley (and in other scientific disciplines) I would say about half the students have switched to Macs (largely in the last year or two). At the coffee shop down the street from the CS building, I see more P/iBooks than all other laptops combined. I figured there must be a reason, so I myself switched a few weeks ago (after a lifetime of professing my hatred for Macs) ... the Unix-based OS X, I think, is one of the main selling points.
or better yet, +5 troll. come on, we just need two more underrated's and 1 troll.
that is all
We can't find out. From WP: Like all forms of U.S. Public Diplomacy (propaganda), its broadcast is forbidden in the U.S. itself under the 1948 Smith-Mundt Act.
Does this strike anyone else as, well, fucked up?
I want one !!!!!!!!
What goes up must come down. There's only a finite amount of oil, and so production must start decreasing at sp,e point. Obligatory Wikipedia link for lots more info on peak oil theory.
I have, in fact, read several neuroscience textbooks. Of course, in an approach like the one you suggest you would have to include the computations involved in integrating all the inputs of a single neuron in deciding to produce a single pulse, which could be substantial ... this is the figure I was referring to. A simpler thought experiment: when computers hit the atomic scale, they will almost certianly be more efficient in terms of computations per unit volume X time than the brain is (the brain needs a lot of "support" hardware, due to biological limitations). Once we get there, a brain-sized hunk of the stuff would certainly be more than powerful enough to simulate a brain (considerations of programming the thing aside).
That makes the brain by far the most complicated structure in the known universe.
More complex than a group of monkeys? Or human society? Or how about the sun? Don't get me wrong, the brain is pretty damn complex. But OTOH, direct numerical simulations of the brain should be reaching feasibility in a couple decades (based on current estimates of the brain's pure processing power).
Somewhat offtopic, but dammit children's access to internet porn is NOT the same as child porn . Please stop perpetuating this misinformation. Thank you.
...But when the effort of doing so exceeds the reasonable effort of walking into the library and scanning the entirety of said literary work, I would contend that Google has met it's burden...
... I just do what any script kiddie does best. Now some kind of argument could be made about Google's due diligence to quickly find and close such holes once they are opened... I'm just saying that the issue isn't as simple as your comparison might make it seem.
The analogy breaks down when someone goes to copy a second book. Presumably, a second trip to the copy machine at the library will take just as long as the first. OTOH, once a hole is discovered in Google's book protection, it could be scripted/otherwise automated such that downloading their entire catalog takes only a single click. When I want to crack the CSS on a DVD, I don't have to reinvent the wheel
BTW, I'm completely in support of Google Print, just playing devil's advocate here.
I tried this out today, and I'm sold. After seeing stats on how Opera is significantly faster than Firefox in almost every category, I finally decided to check it out. While I miss one or two extensions (Bugmenot and Forecast Firefox), I can do without these. Other than that, the built-in mouse gestures, keyboard + location bar shortcuts, ad block, torrents, better download manager, fast forward (hit the button or press ctrl-x and automatically go to the next page of google search results, next part of any article, ...), and so on means that out of the box it is a firefox killer, and much faster to boot.
Last.FM is similiarly trivial to rip.
To help out those of us who are not so savvy with such things, please post your script! I thought about trying to do this but was stumpted as to how.
and these people are running our country...
You've never programmed extensively in Javascript/DTML, I take it. I haven't done any serious web design in 5 or 6 years, but at least back then browser compatbility was a huge headache (each browser & version had a slightly different object model that had to be taken into account explicitly to make things work and degrade cleanly).
I think the use of the word "fuel" is just downright confusing, since although the stuff being referred to is expended during acceleration, it is not itself providing any actual energy. It is just arbitrary junk being thrown off the back of the spaceship to speed it up (ala Newton's 3rd law), with the potential difference presumably coming from some other energy-producing substance that should more accurately be referred to as the fuel.
The Cell won't be terribly well suited for AI either
I don't see why not. Each vector processor should be particularly good at, e.g., evaluating or training a single neural net. And while it is true that most AI algorithms today are not suited to parallelism, don't forget that most games have a bunch of stuff going on at once (maybe 10 enemies on screen at a time that need AI, plus physics, graphics, etc). The game itself is inherently parallelizable, so the AI algorithms don't have to be.
What, and Microsoft's model is much better? They release software that SHOULD be in beta as a final release, and then deal with the million remaining vulnerabilities, etc. later. IMHO at least Google is being honest about the status of their products.
Exactly.... I'm sure improvements to the interface will be coming as they get more of this kind of feedback. As for the paucity of content, remember that anyone can sell a video on video.google.com (with google taking 30%? of the fee). Thus, it is in their best interest to launch the store as early as possible, to entice more copyright owners to sell their videos through their service. This is in stark contrast (I assume) to the model taken by, i.e., iTunes, where content is solicited from a few large corporations.
Since all of the other answers to your ? seem a bit off, I'll take a shot at it. Qubits can be used as ordinary bits with no problem, by simply avoiding superpositions of states. The issue is a practical, not theoretical one; why would you go to all the trouble of lasers and ion traps yaddy yadda to build a 10 bit (i.e. only 10 computing elements) computer that runs at perhaps 1000 Hz, when you could easily build a chip with 10 billion transistors clocked at 100 GHz instead (ballpark figures for today's chips). It will be a long time, if ever, before a quantum computer would also be a practial, efficient classical computer. The gain from QC comes in from using special-purpose quantum algorithms such as Shor's algorithm (exponential speedup over classical computers specific to the factoring problem) or Grover's algorithm (quadratic speedup on all NP complete problems), plus more algorithms yet to be discovered (the field is fairly new). On an ordinary classical algorithm, a decent QC will perform at the level of a classical computer from before I was born.
I think people around here forget that DRM is not inherently evil (although perhaps it is inherently flawed). The problem we have with DRM is that content producers use it to impose unnecessary requirements (your e-book expires in 3 months, you can play this song on your computer but not your ipod, etc.). Keep in mind that the content producers get to decide what requirements they want, and would refuse to release most current works without some form of DRM. Google is not setting the rules here; they're just trying to stay in the game.
By this point, I think availability is growing faster than the body of useful knowledge. Even if the total amount of available information has doubled in the last 20 years, new search technologies make it 1000 times faster to find what you want (approximately, of course). While TFA talks about emerging technologies like del.ico.us and personalized search, I think the real boom is still to come, in the form of real AI.
When computers are fast enough and new algorithms are developed to really harness this power (I give it 10 years, give or take, for this to begin), computers will finally be able to at least have the semblance of understanding the body of knowledge rather than just syntactically sifting through it. This will give us another order of magnitude change comparable to the introduction of search technologies in the first place. Imagine being able to ask google "in one paragraph, summarize the most influential inventions of 2015". Not the most interesting or illuminating example, but you get the idea.
Ummm... I'll guess you just did, along with a million other Slashdotters. Given that this started making the rounds a couple months ago IIRC, and has been linked from many mainstream news stories, I'm sure he's seen a lot of traffic. Who cares where the visitors came from once they're there ... there's not even any content to distract from the ads! I'm sure a year from now nobody will visit the stupid page, but I think a few hundred bucks for a few million page views is not a bad deal in the advertising business.