Dude, strictly for novelty and interest; not for juvenile gross out value. If I was having an operation, I wouldn't mind signing waivers to have it web cast, it would be cool, AND interesting. How many of us ever get to see something like this?
Who pays for open source software? For the most part, noone: the effort is supported largely by those who WANT to make it work. Support of course is provided for a fee; some things will always require payment. I believe the same thing would be possible in most areas of science, if only it weren't illegal to make copyrighted data available at will.
Replicability of results is one issue. But this is really a technical issue; the real questions of interest are: how do your findings fair out of sample, and how does to exact same data look under different analytical methodologies?
Yes, you are correct, some journals and other institutions do endevour to make data publicly available, but only some (Review of Financial Studies is the only journal I can recall at the moment).
How do we pay for it? I don't think people would have a problem with providing funding for efforts to make data open source, but US$25,000 per year for one database is just way too much for some universities.
The open source model is something I readily dream about in finance research. The greatest proportion of papers published in journals come from universities with big reputations, not because of there being smarter people at Havard, for example, although to a certain extent that's true, but because of the amount of money they are free to spend on data, and the amount they already have, whereas smaller, lesser known universities, with never-the-less, a capacity for valuable output must scrape the bottom of the barrel just to get enough data for a decent conference paper.
I think there is a sort of un-uttered agreement that the journal review process exists, and that data is kept tightly protected because the establishment ensures proper quality of published output. In days gone by, perhaps that was a feasable approach, but with the advent of open source, and the thousands of developers forums throughout the web, I believe that finally there is a tangible example that argues for a complete overhaul in approach.
Despite the massive size and wide distribution of the community, there is still some, nay much, order to open source. There is some sort of consensus on the best distro, or the best app. for this or that. Reliable, secure and stable abound in open source; the bolts of excellent software are clear despite the storm. Open source has tipped arguments for the Cathederal, or a stuffy establishment that upholds integrity, on its head. The Bazaar works. People can be rewarded for the value of their output, and not their ability to horde.
In terms of a long term world view, I believe open information is the future in all areas of human endevour. How does a company create value by selling the same data, the same idea, a billion times. Let them be fairly recompensed for their effort in gathering the data, or putting the idea into a servicable form, and leave it at that.
Open source has demonstrated that the Bazaar is able to sort the elite from the mundane, and what's more, the volume and value of the output would not have been possible within a single closed establishment.
Oh, I wish, I wish, I wish that data and information was open. I wish that governments would legislate against data and idea hording; I believe such intervention passes the test of expediency. All hording does is hold back the ability of so many people to produce so much valuable output. This is the future for all important areas of human endevour if we ever want to make more of our precious progress before the earth is consumed by the sun.
I have much more to rant about, but you've probably stopped reading by now anyhow.
I have no formal IT/CS training. I'm not much of a programmer, at all. I've been using icewm on Debian with a Japanese environment for about a year now. The latest OpenOffice1.0.0 Japanese distro is here too, so I am serviceable to my M$ using contemporaries. I don't need any fancy desktop, I just use the krxvt terminal and man pages (with less as the pager). Once you've editted on.config file, you can do them all, in just a couple of secs.
But the best thing is that I have paid no money for any of this, and none of it is warezed. It really only took me 6 months to get fully comfortable, and I think slowly the people at work are starting to take me seriously when I say, I don't use Windows and I don't want to.
But I agree with most of the stuff this guy said; but I think that if you read the start of his article, it's clear that he never really had a strong feeling for the philosophy of GNU/Linux, which is something I believe in quite strongly. It is true, Linux still requires either prior know-how (ie, you have always been a computer person), or a strong belief that you really don't want to use M$ or any other proprietary money sucking, spying etc. corporate-ware.
This is not embarassing commercialism; it's a little give and take. There certainly would be no price fixing, and as long as conflicts of interests were avoided, why not? I would prefer to buy the same brand of network cables from "Keen's Cables" rather than from "Wombat's Wires" if I knew the former was supporting my OS. It's a way for the contributions to be made on a quasi-benefit ratio.
Is the philosophy of OSS an important issue here? Yes; everyone knows M$ is the standard for techno-plebs because of their marketing clout. We don't want to sell our souls here, so I think any activity like this would need to be monitored diligently by the community. I think the type of people that are involved in and use OSS at the moment are the type that would take the "no ads thanks" option. But aren't we trying to achieve a wider support base and acceptance of OSS.
We really need to consider how this sort of activity might affect wider plebian opinion; one possible concern is that it may make Linux as a whole a ripe target for M$ trolls if just one distro like this came out: "Oh look! Linux is soooo crap now they have to run ads on it, blah blah blah..." How can we avoid this?
So, I had a go on the Alice machine. For the most part "her" responses were disjointed, somewhat incoherent, and inconsistent. Really, one of the first bugs that needs ironing out is the inconsistency in the responses. Here is an extract (not verbatim, but this was the general gist):
ME: You're not able to learn about humans. ALICE: You don't know that, that's your opinion. What makes you think that? ME: Why aren't you able to learn about humans? ALICE: Because I was programmed that way.
To be fair, however, Alice did demonstrate some learning ability, demonstrating that "her" back end (ahem, you know what I mean) is not just a static database.
Anyway, following on from the discussion about the set of coherent English sentences; suppose all of these could be listed, not only would they contain "all" intelligence, as least as far as language can described, but all stupidity also. A coherent sentence does not preclude the absence of fact.
Moreover, and as was hinted to in the parent post, even if the number of sentenses is not infinite, the number of combinations, and subsets of said combinations, of sentences, split into paragraphs, chapters and so on, would most certainly approach infinity. I'm not going to speculate as to whether there is a finite number of atoms in the universe or not, but it is doubtful that such a list could be "stored" physically in any sense. Even if it was stored, can it really be argued that it is "known"? Further, this set would be continuously growing as new words for new phenomenon were invented. And this would still not encompass the total available knowlege. The true set of all knowledge is the universe itself. How can you "know" that? Intelligence lies in knowing a very small subset of total possible knowledge, extrapolating and inferring the rest in broad concepts, and acting based on that to achieve some end. What is intelligence without some sort of desired end?
What the hell am I getting at that wasn't already said? Yeah, that's right. How do humans engage in conversation? Surely they don't have predetermined responses to 40,000 questions. Perhaps that is the set of small talk available in current society. So what Wallace has invented is a sophisticated small talk machine.
But humans retain some set of knowledge, through experience and learning, and answer intelligently based upon this. The response to the same question, or argument changes over time, not just the number of questions answerable. So trying to record every possible question, and set of questions or arguments and their possible intelligent responses, and then randomly changing it over time, while adding to the database of knowledge in real time, while theoretically possible, and yes, physically, well it's just ridiculous to even consider how you might go about it physically, might produce something that appears intelligent. But these responses have to come from somewhere in the first place; so what you're really observing is the intelligence of the gimp typing the crap. And perhaps some learned responses from chat participants.
It seems to me that Alice would be one useful component of an artificially intelligent bot. The unconscious part that gets you through the routine parts of life, that you do over and over without thinking. Corresponding to the part of your brain that says, "I'll do it later," when your wife says "Do the dishes." It's the part that drives the car at 3 in the morning, so that when you get home you go, "What the hell, where did I just come from?"
Although this might form one part of an intelligent entity, the intelligence comes in knowing when to switch to autopilot, and when to develop new responses to stimuli.
Alice really has nothing to do with the part of intelligence that adventures through the theoretical set of knowledge, and retains efficiently mainly only what is needed, acting upon that to achieve some end. This approach to A.I. reminds me of using procedure based programming in BASIC with line numbers to build anything more complex than a hangman game. Yeck.
I forget what my sig is, so ignore what is below if it is not amusing.
Let's examine the logic of your cynical little post from a different perspective:
Women should not be informed of techniques to avoid rape. As long as information regarding ways of avoiding potential rapists - for example, not walking alone at night, carrying pepper spray etc. - is available to women, it is available to all. In which case, it is also available to rapists. Any new idea we come up with and disseminate to help women protect themselves from rapists will simply be rendered obsolete. Therefore, the best way to protect women from rape is to not talk about it.
Miss Word? What, are you fucking insane? Word is a piece of over chromed crap. I hate that fucking program. What the fuck is it good for? Ooh, it makes nice and fancy things. Excel, now there's M$s one good piece of software. A spreadsheet with an easy easy easy script language (even if VBA itself is a piece of shit). But fuck that, there are better tools around.
But seriously, what is it that Word is soooo good for. I fucking hate that piece of shit for writing.
Slashdot Mirror
Dude, strictly for novelty and interest; not for juvenile gross out value. If I was having an operation, I wouldn't mind signing waivers to have it web cast, it would be cool, AND interesting. How many of us ever get to see something like this?
What I'd like to see is a web cast of this sort of thing.
Who pays for open source software? For the most part, noone: the effort is supported largely by those who WANT to make it work. Support of course is provided for a fee; some things will always require payment. I believe the same thing would be possible in most areas of science, if only it weren't illegal to make copyrighted data available at will.
Replicability of results is one issue. But this is really a technical issue; the real questions of interest are: how do your findings fair out of sample, and how does to exact same data look under different analytical methodologies?
Yes, you are correct, some journals and other institutions do endevour to make data publicly available, but only some (Review of Financial Studies is the only journal I can recall at the moment).
How do we pay for it? I don't think people would have a problem with providing funding for efforts to make data open source, but US$25,000 per year for one database is just way too much for some universities.
The open source model is something I readily dream about in finance research. The greatest proportion of papers published in journals come from universities with big reputations, not because of there being smarter people at Havard, for example, although to a certain extent that's true, but because of the amount of money they are free to spend on data, and the amount they already have, whereas smaller, lesser known universities, with never-the-less, a capacity for valuable output must scrape the bottom of the barrel just to get enough data for a decent conference paper.
I think there is a sort of un-uttered agreement that the journal review process exists, and that data is kept tightly protected because the establishment ensures proper quality of published output. In days gone by, perhaps that was a feasable approach, but with the advent of open source, and the thousands of developers forums throughout the web, I believe that finally there is a tangible example that argues for a complete overhaul in approach.
Despite the massive size and wide distribution of the community, there is still some, nay much, order to open source. There is some sort of consensus on the best distro, or the best app. for this or that. Reliable, secure and stable abound in open source; the bolts of excellent software are clear despite the storm. Open source has tipped arguments for the Cathederal, or a stuffy establishment that upholds integrity, on its head. The Bazaar works. People can be rewarded for the value of their output, and not their ability to horde.
In terms of a long term world view, I believe open information is the future in all areas of human endevour. How does a company create value by selling the same data, the same idea, a billion times. Let them be fairly recompensed for their effort in gathering the data, or putting the idea into a servicable form, and leave it at that.
Open source has demonstrated that the Bazaar is able to sort the elite from the mundane, and what's more, the volume and value of the output would not have been possible within a single closed establishment.
Oh, I wish, I wish, I wish that data and information was open. I wish that governments would legislate against data and idea hording; I believe such intervention passes the test of expediency. All hording does is hold back the ability of so many people to produce so much valuable output. This is the future for all important areas of human endevour if we ever want to make more of our precious progress before the earth is consumed by the sun.
I have much more to rant about, but you've probably stopped reading by now anyhow.
I have no formal IT/CS training. I'm not much of a programmer, at all. I've been using icewm on Debian with a Japanese environment for about a year now. The latest OpenOffice1.0.0 Japanese distro is here too, so I am serviceable to my M$ using contemporaries. I don't need any fancy desktop, I just use the krxvt terminal and man pages (with less as the pager). Once you've editted on .config file, you can do them all, in just a couple of secs.
But the best thing is that I have paid no money for any of this, and none of it is warezed. It really only took me 6 months to get fully comfortable, and I think slowly the people at work are starting to take me seriously when I say, I don't use Windows and I don't want to.
But I agree with most of the stuff this guy said; but I think that if you read the start of his article, it's clear that he never really had a strong feeling for the philosophy of GNU/Linux, which is something I believe in quite strongly. It is true, Linux still requires either prior know-how (ie, you have always been a computer person), or a strong belief that you really don't want to use M$ or any other proprietary money sucking, spying etc. corporate-ware.
Yay! Why you only get 2 for your excellent post? Me no understand.
If noone ever reads this, did I ever really write it? (ahh, me in silly mood, hehee)
This is not embarassing commercialism; it's a little give and take. There certainly would be no price fixing, and as long as conflicts of interests were avoided, why not? I would prefer to buy the same brand of network cables from "Keen's Cables" rather than from "Wombat's Wires" if I knew the former was supporting my OS. It's a way for the contributions to be made on a quasi-benefit ratio.
Is the philosophy of OSS an important issue here? Yes; everyone knows M$ is the standard for techno-plebs because of their marketing clout. We don't want to sell our souls here, so I think any activity like this would need to be monitored diligently by the community. I think the type of people that are involved in and use OSS at the moment are the type that would take the "no ads thanks" option. But aren't we trying to achieve a wider support base and acceptance of OSS.
We really need to consider how this sort of activity might affect wider plebian opinion; one possible concern is that it may make Linux as a whole a ripe target for M$ trolls if just one distro like this came out: "Oh look! Linux is soooo crap now they have to run ads on it, blah blah blah..." How can we avoid this?
Popular press is the whore of the mind
So, I had a go on the Alice machine. For the most part "her" responses were disjointed, somewhat incoherent, and inconsistent. Really, one of the first bugs that needs ironing out is the inconsistency in the responses. Here is an extract (not verbatim, but this was the general gist):
ME: You're not able to learn about humans.
ALICE: You don't know that, that's your opinion. What makes you think that?
ME: Why aren't you able to learn about humans?
ALICE: Because I was programmed that way.
To be fair, however, Alice did demonstrate some learning ability, demonstrating that "her" back end (ahem, you know what I mean) is not just a static database.
Anyway, following on from the discussion about the set of coherent English sentences; suppose all of these could be listed, not only would they contain "all" intelligence, as least as far as language can described, but all stupidity also. A coherent sentence does not preclude the absence of fact.
Moreover, and as was hinted to in the parent post, even if the number of sentenses is not infinite, the number of combinations, and subsets of said combinations, of sentences, split into paragraphs, chapters and so on, would most certainly approach infinity. I'm not going to speculate as to whether there is a finite number of atoms in the universe or not, but it is doubtful that such a list could be "stored" physically in any sense. Even if it was stored, can it really be argued that it is "known"? Further, this set would be continuously growing as new words for new phenomenon were invented. And this would still not encompass the total available knowlege. The true set of all knowledge is the universe itself. How can you "know" that? Intelligence lies in knowing a very small subset of total possible knowledge, extrapolating and inferring the rest in broad concepts, and acting based on that to achieve some end. What is intelligence without some sort of desired end?
What the hell am I getting at that wasn't already said? Yeah, that's right. How do humans engage in conversation? Surely they don't have predetermined responses to 40,000 questions. Perhaps that is the set of small talk available in current society. So what Wallace has invented is a sophisticated small talk machine.
But humans retain some set of knowledge, through experience and learning, and answer intelligently based upon this. The response to the same question, or argument changes over time, not just the number of questions answerable. So trying to record every possible question, and set of questions or arguments and their possible intelligent responses, and then randomly changing it over time, while adding to the database of knowledge in real time, while theoretically possible, and yes, physically, well it's just ridiculous to even consider how you might go about it physically, might produce something that appears intelligent. But these responses have to come from somewhere in the first place; so what you're really observing is the intelligence of the gimp typing the crap. And perhaps some learned responses from chat participants.
It seems to me that Alice would be one useful component of an artificially intelligent bot. The unconscious part that gets you through the routine parts of life, that you do over and over without thinking. Corresponding to the part of your brain that says, "I'll do it later," when your wife says "Do the dishes." It's the part that drives the car at 3 in the morning, so that when you get home you go, "What the hell, where did I just come from?"
Although this might form one part of an intelligent entity, the intelligence comes in knowing when to switch to autopilot, and when to develop new responses to stimuli.
Alice really has nothing to do with the part of intelligence that adventures through the theoretical set of knowledge, and retains efficiently mainly only what is needed, acting upon that to achieve some end. This approach to A.I. reminds me of using procedure based programming in BASIC with line numbers to build anything more complex than a hangman game. Yeck.
I forget what my sig is, so ignore what is below if it is not amusing.
Remember when 120 meg hard drives started to come out, and you said, "I my f***ing God, I'll never fill it up or really need all that space."...
Let's examine the logic of your cynical little post from a different perspective:
Women should not be informed of techniques to avoid rape. As long as information regarding ways of avoiding potential rapists - for example, not walking alone at night, carrying pepper spray etc. - is available to women, it is available to all. In which case, it is also available to rapists. Any new idea we come up with and disseminate to help women protect themselves from rapists will simply be rendered obsolete. Therefore, the best way to protect women from rape is to not talk about it.
Miss Word? What, are you fucking insane? Word is a piece of over chromed crap. I hate that fucking program. What the fuck is it good for? Ooh, it makes nice and fancy things. Excel, now there's M$s one good piece of software. A spreadsheet with an easy easy easy script language (even if VBA itself is a piece of shit). But fuck that, there are better tools around. But seriously, what is it that Word is soooo good for. I fucking hate that piece of shit for writing.
Savvy has two "v"s. Isn't that interesting?