"If you go to a hotel and ask the concierge for reservations to Morton's and he says, "ah, but here is a better steakhouse that my buddy runs" -- can you imagine that being illegal?"
Google isn't doing that, Google is saying "hey go to Frank's" because Frank paid them 50 cents to tell you that. It may not be illegal, but it is arguably unethical.
My guess: RAM prices come down in steps. RAM gets cheaper because new fab technologies come out, but upgrading to these new factories costs a lot of money. This loss is recouped by selling the current technology at a higher price, while discounting the old technology to have half a chance in hell of selling enough to maintain a profit. This is the way CPU prices work (see current article). While prices come down slowly (I purchased a gig of RAM for 100 dollars last year, which is almost 50% more than current prices), it is not noticeable until one of these new technologies come out. I usually time my computer upgrades to coincide with these events.
I could be totally wrong. Sometimes catastrophic events can actually increase RAM prices, such as the earthquake in taiwan a few years back that drastically cut production of RAM.
"the real problem will be making cars that are manufacturable... AND sellable."...and street legal in the USA. The USA is one of the biggest potential markets for something like this, and has some pretty stringent street laws.
You are making an argument based on laws that IMO haven't strongly been tested in courts (we won't really know the outcome until this goes to court.) I purposefully tried to avoid this line of discussion when I made my original post. I am talking about it from an even more abstract POV: the moral point of view. I think turnitin improves the world, and is therefore morally OK. I don't think it hurts anyone, so it is OK. If you reread my original post, you will see that was my *only* point. If it makes you feel any better, I will admit there is a good chance what turnitin is doing is illegal. That doesn't bother me in the least, and I will grudgingly let the idiot lawyers and courts battle that out at a huge cost to the taxpayer. The IP and copyright laws are too broken in the USA to have any sort of moral meaning.
BTW, google records what you search and visit even if you don't log in. They still have your IP, so they can weakly attach it to you. There is no way to opt-out, except to not use Google. I am not attacking Google, I also find them to be morally A-OK, I am just trying to use them as a grounded moral benchmark to compare against.
But those same people don't complain when their searches come out great on Google. From a machine learning perspective, you can't do much useful without data. Real world data. It's a dangerous game using real world data, but IMO it is essential to truly take advantage of what computers can do.
And ignoring legality, I think Google recording what you search, attaching it to your name, and recording what websites you visit, is far worse of an infraction on our rights and privacy than the anti-plagiarism website. I am willing to sacrifice that privacy because I know Google's 150 billion dollar market cap depends on it (in other words I trust them) and because I know they end up making a better product because of it.
So far, every comment I have read has been resting on the assumption that something wrong is happening. I think what Turnitin does it great - they catch people who cheat, which (relatively) helps the students who don't cheat. They have created a service that applies some pretty cool research in machine learning, and are paying honest programmers and researchers in the process (perhaps this is my bias as a machine learning researcher). They help teachers do their job, without interfering with the creative or subjective process of the teachers grading.
The only thing that could be construed as "wrong" in all this is that the students' papers are being added to a database. I don't find this to be a very valid complaint. The company is not stealing their papers, it is making sure no one copies them. Companies employ this kind of pattern all the time in our daily lives - every time you search in google, they record what you searched for along with plenty of other contextual clues (including what sites you visited). These records aren't being "stolen" or published, they are used to improve the service. Likewise, these students papers aren't being published, they are being added to a machine learning database to improve the service in the future.
I understand the intuition that something is wrong, but I think deeper thought shows that in the long run this is pretty win-win.
It all depends on what you want to do with your portable computer. A lot of people have managed to turn their cell phones into functional computers with games, messaging, music/video, web browsing and (of course) voice. All without a keyboard. My optimal portable computer would not have a keyboard at all, but would instead be able to understand me through other means: touch screen for drawing equations/thoughts, voice recognition for writing paragraphs or commands, and mind reading for deciding what music I want to listen to. But seriously, I don't want my portable computer to be constrained by the size of keyboards - if I want a keyboard I will use a modern laptop or my desktop. But what's the fun in that?
"Currently, people just 'think' they don't know how to use MS Word."
I have never met anyone who didn't know how to use MS Word (or admitted it, at the very least). One thing I have to give MSFT credit for is they made Word pretty damn easy to use. I haven't used the new Office, but I have heard good things. You are comparing Office to a theoretical competitor that never was, which is not fair.
Who cares if Word was hard to use 10 years ago? MSFT Office became a defacto standard because they convinced people to use it, not because they used any existing monopolies to force others out of the business (AFAIK). TODAY, Microsoft office is probably the best office suite. I hope OO will match is someday, but I am still waiting.
"If you enjoy, or are into the work, it is very difficult to become a segmentor"
I enjoy what I do, and I think I am good at it, but I was raised to be a segmentor. I am still relatively young, and I think you may be right that this attitude will push me away from being a coder when my career picks up, but meanwhile I don't see it hurting my coding abilities in the *hours I am actually working.* If I have a thought, I tuck it deep into my head and implement it first thing in the morning or write it on my to-do list.
Being a segmentor has its (subjective) pros and cons.
Cons: You care less about the company, and therefore your job The job sometimes seems like a chore You get less done because you aren't always thinking about the problems at hand
Pros: You have more free time for family and friends You don't feel as obligated to do things you shouldn't have to do (e.g. stay late) You have a life
Read that last pro. Very few integrators, IMO, can really claim that last one. Perhaps it is my bias, but I never thought working 50-80 hour weeks and then thinking about those hours at home was conducive to having a life.
It's more complicated than that. From the economist:
"The legal situation is ambiguous. At issue is America's Digital Millennium Copyright Act, which became law in 1998, when Google, founded that year, was unknown and YouTube did not exist. It includes a "safe-harbour" provision for anybody who removes copyrighted content as soon as the owner requests it.
YouTube has been doing that--most notably a month ago, when Viacom demanded that 100,000 clips be taken down. But the safe-harbour clause applies only as long as a site does not gain financially from infringement."
Does youtube make any sort of profit off its videos?
I believe he is working on hardware solutions to that. One of the things he emphasizes in his book "On Intelligence" is that machines need to have more memory to do what AI wants to do.
I did not rtfa, who has time for that anymore;) But regardless, in the Slashdot fashion, here is my opinion:
I played around with some of his publicly available code a few months ago. It was pretty impressive on a toy problem (recognizing a small set of characters) but was very, very slow at training (on the order of hours or days to learn the simple problem).
But on the other hand, I can't think of any sort of technology that could do better than it (I am into machine learning and AI.) Also, it is not a big deal if it trains slowly if it can compute fast - the human brain took millions of years to evolve.
From what I could tell, his technology is (was?) a glorified bayes net with time forced into the model. To train the net to recognize images he just moved an image around a bunch and had the net brute-force learn all the possible patterns. Training this could get tedious, to say the least. In theory it sounds like he's on to something, but it came off as a pretty simple modification to an old algorithm. I found his book entertaining, and insightful at times, but not revolutionary. He reasons out some really cool ideas, but it comes off more as philosophy more than science sometimes.
"Just because you're used to something in Windows doesn't make it the best or even the natural method."
Sorry, but I have used linux enough to know that the current system is no better than Windows for a lay user. Yeah, I am pretty advanced in Windows, but it is really damn easy to use for just about everybody.
"In Windows you need to find out what program you want, then you have to download it, then you have to install it, then you have to make sure that you keep it updated"
I fail to see how anything could possibly be different. Does Linux figure out what programs you want for you?;)
According to my game developer friend, this isn't really possible. He says that the output from the controller is way too noisy to get good positional data. That is why, in most games, everything is controlled by simple movements of the controller instead of precise gestures.
If this is true, I think it is purposefully being hidden by Nintendo. Everyone is waiting for killer games to come out for the Wii that truly utilize the hardware, but the deep down secret may be that this will never occur.
I don't own the system, and have only played it for a few hours, so I may be wrong. Please correct me if I am.
It seems to me that this way is the best method for 90% of people (I'm a Linux n00b): 4) Installing from distribution-independent binaries (most proprietary software is delivered this way),
I apologize if the answer to this is obvious, but why isn't everything packaged that way?
"Although Bill Gates rejected Mikhail Gorbachev's personal appeal for mercy on behalf of the teacher..."
I haven't been following the story since the last Slashdot article, but did Bill Gates reject the personal appeal, or did Microsoft? Huge difference. I can't find any source indicating what exactly happened from their side.
BT is efficient for bandwidth distribution, but not storage space. The typical bit torrent file is stored on 10s - 1000s of different computers. For a huge collection of videos, like youtube, dozens of copies are wasteful.
"I have seen WW-II era footage of soldiers ing tiny generators by hand or by legs to power their radio sets."
It sounds like they are going for something that would require no additional effort from the astronauts. They have better things to do than rotate a generator in a huge spaceship that already has ample power (compared to a hand-cranked generator at least).
If I owned stock in ebay, I would sell it. I seriously don't trust them with any sort of common-sense, monetary decisions. Why does every great company start out so cool, and then end up succumbing to business school morons who drive the company into the ground with their lack of intelligence and overconfidence?
"If the FTC or whomever must review the ENTIRE content of a video game, does that mean every possible combination of levels/characters/interactions?"
Funny, that was my first thought too. Then I remembered they aren't testing the game for bugs, they are just looking for anything offensive. Generally, a cursory run through a game will give a pretty good indication of the rating. In the very rare cases where a developer is stupid enough to put something *hidden* into the game that will ruin its rating, a simple but direct question to the developers making them disclose it should cover the rest. I am not claiming this is airtight, but it's still "better" than the current system.
I think you missed the point. While I agree this is not revolutionary, it is different in a few ways: -It's a neuroscience project more than a machine learning project (simulating the brain, not a function to be learned) -It's trying to mimic the *hardware* of the brain; it's not software written for a general purpose CPU -It's probably more powerful
I frankly think this project is stupid, because it's the connections in the brain that make intelligence, not the neurons. We don't understand the connections and how they work. But I guess we'll see if it works.
Slashdot Mirror
I haven't seen a doubleclick ad in years thanks to adblock
Oh, and just for this, I am going to set up adblock to block google ads too
"If you go to a hotel and ask the concierge for reservations to Morton's and he says, "ah, but here is a better steakhouse that my buddy runs" -- can you imagine that being illegal?"
Google isn't doing that, Google is saying "hey go to Frank's" because Frank paid them 50 cents to tell you that. It may not be illegal, but it is arguably unethical.
"But everyone wins, because HP this summer will debut a special new calculator model."
Subtle...
My guess:
RAM prices come down in steps. RAM gets cheaper because new fab technologies come out, but upgrading to these new factories costs a lot of money. This loss is recouped by selling the current technology at a higher price, while discounting the old technology to have half a chance in hell of selling enough to maintain a profit. This is the way CPU prices work (see current article). While prices come down slowly (I purchased a gig of RAM for 100 dollars last year, which is almost 50% more than current prices), it is not noticeable until one of these new technologies come out. I usually time my computer upgrades to coincide with these events.
I could be totally wrong. Sometimes catastrophic events can actually increase RAM prices, such as the earthquake in taiwan a few years back that drastically cut production of RAM.
"delivery infrastructure"
What about the delivery infrastructure of coal shipments? I think the last thing Google wants to do is get into the power plant business.
"the real problem will be making cars that are manufacturable... AND sellable." ...and street legal in the USA. The USA is one of the biggest potential markets for something like this, and has some pretty stringent street laws.
You are making an argument based on laws that IMO haven't strongly been tested in courts (we won't really know the outcome until this goes to court.) I purposefully tried to avoid this line of discussion when I made my original post. I am talking about it from an even more abstract POV: the moral point of view. I think turnitin improves the world, and is therefore morally OK. I don't think it hurts anyone, so it is OK. If you reread my original post, you will see that was my *only* point. If it makes you feel any better, I will admit there is a good chance what turnitin is doing is illegal. That doesn't bother me in the least, and I will grudgingly let the idiot lawyers and courts battle that out at a huge cost to the taxpayer. The IP and copyright laws are too broken in the USA to have any sort of moral meaning.
_ moral_development
Our discussion is happening at different levels on this scale:
http://en.wikipedia.org/wiki/Kohlberg's_stages_of
BTW, google records what you search and visit even if you don't log in. They still have your IP, so they can weakly attach it to you. There is no way to opt-out, except to not use Google. I am not attacking Google, I also find them to be morally A-OK, I am just trying to use them as a grounded moral benchmark to compare against.
But those same people don't complain when their searches come out great on Google. From a machine learning perspective, you can't do much useful without data. Real world data. It's a dangerous game using real world data, but IMO it is essential to truly take advantage of what computers can do.
And ignoring legality, I think Google recording what you search, attaching it to your name, and recording what websites you visit, is far worse of an infraction on our rights and privacy than the anti-plagiarism website. I am willing to sacrifice that privacy because I know Google's 150 billion dollar market cap depends on it (in other words I trust them) and because I know they end up making a better product because of it.
So far, every comment I have read has been resting on the assumption that something wrong is happening. I think what Turnitin does it great - they catch people who cheat, which (relatively) helps the students who don't cheat. They have created a service that applies some pretty cool research in machine learning, and are paying honest programmers and researchers in the process (perhaps this is my bias as a machine learning researcher). They help teachers do their job, without interfering with the creative or subjective process of the teachers grading.
The only thing that could be construed as "wrong" in all this is that the students' papers are being added to a database. I don't find this to be a very valid complaint. The company is not stealing their papers, it is making sure no one copies them. Companies employ this kind of pattern all the time in our daily lives - every time you search in google, they record what you searched for along with plenty of other contextual clues (including what sites you visited). These records aren't being "stolen" or published, they are used to improve the service. Likewise, these students papers aren't being published, they are being added to a machine learning database to improve the service in the future.
I understand the intuition that something is wrong, but I think deeper thought shows that in the long run this is pretty win-win.
It all depends on what you want to do with your portable computer. A lot of people have managed to turn their cell phones into functional computers with games, messaging, music/video, web browsing and (of course) voice. All without a keyboard. My optimal portable computer would not have a keyboard at all, but would instead be able to understand me through other means: touch screen for drawing equations/thoughts, voice recognition for writing paragraphs or commands, and mind reading for deciding what music I want to listen to. But seriously, I don't want my portable computer to be constrained by the size of keyboards - if I want a keyboard I will use a modern laptop or my desktop. But what's the fun in that?
"Currently, people just 'think' they don't know how to use MS Word."
I have never met anyone who didn't know how to use MS Word (or admitted it, at the very least). One thing I have to give MSFT credit for is they made Word pretty damn easy to use. I haven't used the new Office, but I have heard good things. You are comparing Office to a theoretical competitor that never was, which is not fair.
Who cares if Word was hard to use 10 years ago? MSFT Office became a defacto standard because they convinced people to use it, not because they used any existing monopolies to force others out of the business (AFAIK). TODAY, Microsoft office is probably the best office suite. I hope OO will match is someday, but I am still waiting.
"If you enjoy, or are into the work, it is very difficult to become a segmentor"
I enjoy what I do, and I think I am good at it, but I was raised to be a segmentor. I am still relatively young, and I think you may be right that this attitude will push me away from being a coder when my career picks up, but meanwhile I don't see it hurting my coding abilities in the *hours I am actually working.* If I have a thought, I tuck it deep into my head and implement it first thing in the morning or write it on my to-do list.
Being a segmentor has its (subjective) pros and cons.
Cons:
You care less about the company, and therefore your job
The job sometimes seems like a chore
You get less done because you aren't always thinking about the problems at hand
Pros:
You have more free time for family and friends
You don't feel as obligated to do things you shouldn't have to do (e.g. stay late)
You have a life
Read that last pro. Very few integrators, IMO, can really claim that last one. Perhaps it is my bias, but I never thought working 50-80 hour weeks and then thinking about those hours at home was conducive to having a life.
It's more complicated than that. From the economist:
"The legal situation is ambiguous. At issue is America's Digital Millennium Copyright Act, which became law in 1998, when Google, founded that year, was unknown and YouTube did not exist. It includes a "safe-harbour" provision for anybody who removes copyrighted content as soon as the owner requests it.
YouTube has been doing that--most notably a month ago, when Viacom demanded that 100,000 clips be taken down. But the safe-harbour clause applies only as long as a site does not gain financially from infringement."
Does youtube make any sort of profit off its videos?
I believe he is working on hardware solutions to that. One of the things he emphasizes in his book "On Intelligence" is that machines need to have more memory to do what AI wants to do.
I did not rtfa, who has time for that anymore ;) But regardless, in the Slashdot fashion, here is my opinion:
I played around with some of his publicly available code a few months ago. It was pretty impressive on a toy problem (recognizing a small set of characters) but was very, very slow at training (on the order of hours or days to learn the simple problem).
But on the other hand, I can't think of any sort of technology that could do better than it (I am into machine learning and AI.) Also, it is not a big deal if it trains slowly if it can compute fast - the human brain took millions of years to evolve.
From what I could tell, his technology is (was?) a glorified bayes net with time forced into the model. To train the net to recognize images he just moved an image around a bunch and had the net brute-force learn all the possible patterns. Training this could get tedious, to say the least. In theory it sounds like he's on to something, but it came off as a pretty simple modification to an old algorithm. I found his book entertaining, and insightful at times, but not revolutionary. He reasons out some really cool ideas, but it comes off more as philosophy more than science sometimes.
"Just because you're used to something in Windows doesn't make it the best or even the natural method."
;)
Sorry, but I have used linux enough to know that the current system is no better than Windows for a lay user. Yeah, I am pretty advanced in Windows, but it is really damn easy to use for just about everybody.
"In Windows you need to find out what program you want, then you have to download it, then you have to install it, then you have to make sure that you keep it updated"
I fail to see how anything could possibly be different. Does Linux figure out what programs you want for you?
"b) The Wiisaber. That alone will double sales."
According to my game developer friend, this isn't really possible. He says that the output from the controller is way too noisy to get good positional data. That is why, in most games, everything is controlled by simple movements of the controller instead of precise gestures.
If this is true, I think it is purposefully being hidden by Nintendo. Everyone is waiting for killer games to come out for the Wii that truly utilize the hardware, but the deep down secret may be that this will never occur.
I don't own the system, and have only played it for a few hours, so I may be wrong. Please correct me if I am.
It seems to me that this way is the best method for 90% of people (I'm a Linux n00b):
4) Installing from distribution-independent binaries (most proprietary software is delivered this way),
I apologize if the answer to this is obvious, but why isn't everything packaged that way?
"Although Bill Gates rejected Mikhail Gorbachev's personal appeal for mercy on behalf of the teacher..."
I haven't been following the story since the last Slashdot article, but did Bill Gates reject the personal appeal, or did Microsoft? Huge difference. I can't find any source indicating what exactly happened from their side.
BT is efficient for bandwidth distribution, but not storage space. The typical bit torrent file is stored on 10s - 1000s of different computers. For a huge collection of videos, like youtube, dozens of copies are wasteful.
"I have seen WW-II era footage of soldiers ing tiny generators by hand or by legs to power their radio sets."
It sounds like they are going for something that would require no additional effort from the astronauts. They have better things to do than rotate a generator in a huge spaceship that already has ample power (compared to a hand-cranked generator at least).
"too bad ebay execs are a bunch of anal fucks."
If I owned stock in ebay, I would sell it. I seriously don't trust them with any sort of common-sense, monetary decisions. Why does every great company start out so cool, and then end up succumbing to business school morons who drive the company into the ground with their lack of intelligence and overconfidence?
"If the FTC or whomever must review the ENTIRE content of a video game, does that mean every possible combination of levels/characters/interactions?"
Funny, that was my first thought too. Then I remembered they aren't testing the game for bugs, they are just looking for anything offensive. Generally, a cursory run through a game will give a pretty good indication of the rating. In the very rare cases where a developer is stupid enough to put something *hidden* into the game that will ruin its rating, a simple but direct question to the developers making them disclose it should cover the rest. I am not claiming this is airtight, but it's still "better" than the current system.
I think you missed the point. While I agree this is not revolutionary, it is different in a few ways:
-It's a neuroscience project more than a machine learning project (simulating the brain, not a function to be learned)
-It's trying to mimic the *hardware* of the brain; it's not software written for a general purpose CPU
-It's probably more powerful
I frankly think this project is stupid, because it's the connections in the brain that make intelligence, not the neurons. We don't understand the connections and how they work. But I guess we'll see if it works.
According to your post 98% of the Republicans would have been kicked out of office by now.