There are two main problems with the Chinese room experiment.
1. Thought experiments don't really prove things, only real experiments do.
2. John Searle ignored the person who created the room, who did most of the intellectual work, making the rest derivative and predictable.
It is the person who created the room who is the real intellect there, not the room itself or the person manipulating symbols inside the room. It is the room's creator who solved the general class of problem of language translation in a manner superior to the present entire output of humankind, and who therefore deserves the label "intelligent" more than the room itself.
It's as though I were having a discussing with Searle and I say, "You should read this book, it's intelligent." and he replies, "What's so intelligent about it? Ink? Glue? Dead tree matter? The book-binder or person who sold you the book, perhaps?" "No," I would reply, "when I say a book is intelligent, I mean it's author is intelligent, not the book itself!" Likewise, the creator of the Chinese room, were such a person to exist, would be more intelligent than the room itself, although the room itself would admittedly be impressive in its own right.
The room would not need to display "understanding", because that need would have been eliminated by its creator, who somehow reduced the problem of general understanding to one of rote mechanization.
So... now every time I hear one human being describe another as "tasty" or "quite the dish" I'm going to remember your unfortunate remarks. Thanks a lot for that mental image!
Possible reasons for their failure:
1. Fizzbuzz contains a hidden trap. The cases are given to you in a different order than in what order you are supposed to apply them. If you apply the cases as given, the 15 test comes last and fails, because it is silently passed over due to being redundant to the other cases.
2. The applicants probably thought the problem looked simple and didn't test their code. They assumed that it worked. Or the tester took the first invocation of their program as their final answer, depriving them of the opportunity for trial and error that professionals do in practice.
3. Most of the applicants were probably college grads, not professional programmers. Classes probably accept their code if it is close enough and don't ask them to fix it. Anyway, college is not a coding factory, to drill them in basic coding technique, but a place to learn discipline and concepts.
Did anyone else just envision a future Slashdot, except with fuzzy, warm hugs instead of mod points? Now I can see what this site has been lacking all this time... Thanks, Microsoft!
It's funny how Slashdot, which is so against computer-automated HFT, is equally in favor of the self-driving car. Consider what would happen if, due to some unrecognized design flaw, all the self-driving cars of the world were to crash at once. The resulting mash-up would put stock market flash crashes to shame... Anything can be automated and scaled up, including disasters. At least with HFTs, it's just someone's retirement savings that is getting crushed.
We already have completely automated production of software in the IT industry, i.e., we effortlessly create millions of copies of our product with the push of a button and no manual labor required. And yet there is still full employment of computer programmers, in design rather than in production. Why should the hardware industry be any different? Isn't hardware design an endlessly difficult and varied problem, just like software?
Your comments remind me of the famous quote by George Bernard Shaw: "The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore, all progress depends on the unreasonable man." Or as I would put it, if RMS is the backbone of the free software world, it is more important that he be stiff than that he be flexible.
Clearly, once we accept the premise that it is in fact a suicide mission, it would be far more efficient to kill all the astronauts BEFORE launch and THEN shoot their dead bodies to Mars, leaving off all those unnecessary expenses for both heat and water...
Having considered the options presented here, I just thought of an alternative to college. Just imagine if the big tech companies got together and started their own online university solely devoted to STEM majors. That would include Microsoft, Google, Apple, IBM, Facebook, Yahoo!, etc. In between them they are sitting on more than $100 billion of cash, which is enough to start an Ivy-league university of their own. They could disrupt higher education and supply themselves with plenty of STEM graduates that meet their needs as employees, in addition to funding research through a college-like organization. They could work their way through the accreditation process so that it is a "real" degree and not just a collection of online classes. And they could make use of it for their own ongoing employee education. Just think of the possibilities...
Slashdot Mirror
There are two main problems with the Chinese room experiment. 1. Thought experiments don't really prove things, only real experiments do. 2. John Searle ignored the person who created the room, who did most of the intellectual work, making the rest derivative and predictable. It is the person who created the room who is the real intellect there, not the room itself or the person manipulating symbols inside the room. It is the room's creator who solved the general class of problem of language translation in a manner superior to the present entire output of humankind, and who therefore deserves the label "intelligent" more than the room itself. It's as though I were having a discussing with Searle and I say, "You should read this book, it's intelligent." and he replies, "What's so intelligent about it? Ink? Glue? Dead tree matter? The book-binder or person who sold you the book, perhaps?" "No," I would reply, "when I say a book is intelligent, I mean it's author is intelligent, not the book itself!" Likewise, the creator of the Chinese room, were such a person to exist, would be more intelligent than the room itself, although the room itself would admittedly be impressive in its own right. The room would not need to display "understanding", because that need would have been eliminated by its creator, who somehow reduced the problem of general understanding to one of rote mechanization.
So... now every time I hear one human being describe another as "tasty" or "quite the dish" I'm going to remember your unfortunate remarks. Thanks a lot for that mental image!
Possible reasons for their failure: 1. Fizzbuzz contains a hidden trap. The cases are given to you in a different order than in what order you are supposed to apply them. If you apply the cases as given, the 15 test comes last and fails, because it is silently passed over due to being redundant to the other cases. 2. The applicants probably thought the problem looked simple and didn't test their code. They assumed that it worked. Or the tester took the first invocation of their program as their final answer, depriving them of the opportunity for trial and error that professionals do in practice. 3. Most of the applicants were probably college grads, not professional programmers. Classes probably accept their code if it is close enough and don't ask them to fix it. Anyway, college is not a coding factory, to drill them in basic coding technique, but a place to learn discipline and concepts.
Did anyone else just envision a future Slashdot, except with fuzzy, warm hugs instead of mod points? Now I can see what this site has been lacking all this time... Thanks, Microsoft!
It's funny how Slashdot, which is so against computer-automated HFT, is equally in favor of the self-driving car. Consider what would happen if, due to some unrecognized design flaw, all the self-driving cars of the world were to crash at once. The resulting mash-up would put stock market flash crashes to shame... Anything can be automated and scaled up, including disasters. At least with HFTs, it's just someone's retirement savings that is getting crushed.
We already have completely automated production of software in the IT industry, i.e., we effortlessly create millions of copies of our product with the push of a button and no manual labor required. And yet there is still full employment of computer programmers, in design rather than in production. Why should the hardware industry be any different? Isn't hardware design an endlessly difficult and varied problem, just like software?
Out of curiosity, exactly what would you say about GW "deniers" if their skepticism eventually proves correct?
Your comments remind me of the famous quote by George Bernard Shaw: "The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore, all progress depends on the unreasonable man." Or as I would put it, if RMS is the backbone of the free software world, it is more important that he be stiff than that he be flexible.
I agree. What you are suggesting seems to be a variation on the law of diminishing returns (http://en.wikipedia.org/wiki/Diminishing_returns).
Clearly, once we accept the premise that it is in fact a suicide mission, it would be far more efficient to kill all the astronauts BEFORE launch and THEN shoot their dead bodies to Mars, leaving off all those unnecessary expenses for both heat and water...
Having considered the options presented here, I just thought of an alternative to college. Just imagine if the big tech companies got together and started their own online university solely devoted to STEM majors. That would include Microsoft, Google, Apple, IBM, Facebook, Yahoo!, etc. In between them they are sitting on more than $100 billion of cash, which is enough to start an Ivy-league university of their own. They could disrupt higher education and supply themselves with plenty of STEM graduates that meet their needs as employees, in addition to funding research through a college-like organization. They could work their way through the accreditation process so that it is a "real" degree and not just a collection of online classes. And they could make use of it for their own ongoing employee education. Just think of the possibilities...