3 ms·

Or apply modus tollens and infer our consciousness is not reducible to our meat machines.

by yters 7y ago

coldtea 7y ago
Nobody said it is: it's in the interconnections and stored info inside the meat machines. A T-bone is also meat but without consciousness. Except if you think it's in an external, not meat/neurons/info soul.
- TheOtherHobbes 7y ago
  Substrate independence - whether it's connections or something else - is a very strong claim that should be subjected to empirical testing before being accepted. (And IMO is almost certainly wrong for good practical reasons that have nothing to do with souls.) Of course you can't test it empirically, because there's no empirical test for subjective experience. You can only test for the appearance of subjective experience - specifically for external behaviours, loosely defined, which might be considered to imply subjective experience to you as an observer. This isn't a solvable problem, and Bostrom is begging the question by trying to hand wave it away with a sneaky "If we just accept this premise..."
  - coldtea 7y ago
    >Substrate independence - whether it's connections or something else - is a very strong claim that should be subjected to empirical testing before being accepted. (And IMO is almost certainly wrong for good practical reasons that have nothing to do with souls.) Care to mention any of those "good practical reasons"? Penrose gave it a try (though even there, there was nothing about how it can't be modelled with code for example, in which case, rocks are fine too for a simulation a la XKCD, just not as a direct substrate for an Android). >Of course you can't test it empirically, because there's no empirical test for subjective experience. You can only test for the appearance of subjective experience - specifically for external behaviours, loosely defined, which might be considered to imply subjective experience to you as an observer. It's not really that we can't test it, it's more that we can't measure it. But we can, and do, test it, e.g. with psychological tests, the same kind we give to people to understand different aspects of their personality (which we also can't just measure with some equipment or simple measurement). If it quacks like a duck, and looks like a duck, and cries like a duck, and has other sentimental duck qualities, it's a duck. We only accept the subjective experience of other people from its appearance (their actions etc) too, anyway.
  - krageon 7y ago
    > good practical reasons The only reasoning I have seen on this subject so far has been "people are special" or "God made me". Sometimes the arguments are dressed up more. Do you have a good source that has a different type of reasoning? Edit: > You can only test for the appearance of subjective experience Yes, and we solve this in people by assuming that if they sound, act and look like a person they are a person. I've never seen a compelling argument why it is somehow wrong to apply this principle to everything else, except of course the huge trouble philosophers go to to then argue what exactly needs to be looked at before we are satisfied. For some reason the Turing test is not adequate, but to my mind it's only not accurate if you believe that you are somehow special and that there is more to you than what can be observed.
    - Wowfunhappy 7y ago
      > For some reason the Turing test is not adequate Let's say I build a chatbot with a predetermined database of perfect responses to every possible string of words. This is quite clearly possible in theory, given a database with infinite room. This chatbot would pass the touring test, even though all it's doing is retrieving prewritten sentences. Is it conscious? Ergo, it's quite clearly possible to build an automaton which acts like a human, but actually has no clue what it's saying. Which means the touring test isn't adequate. --- Here's another take: You have no proof that anyone in the universe is conscious other than yourself. However, given that you were born in the same way as everyone else, and you're conscious, everyone else probably is too. Why would you be so special? Well, a computer program was not born in the same way that you were, so that assumption goes out the window.
      - coldtea 7y ago
        >Let's say I build a chatbot with a predetermined database of perfect responses to every possible string of words. This is quite clearly possible in theory, given a database with infinite room. This chatbot would pass the touring test, even though all it's doing is retrieving prewritten sentences. Is it conscious? For all intents and purposes, yes. Though without actual consciousness it would be impossible to build it. >Well, a computer program was not born in the same way that you were, so that assumption goes out the window. Well, a computer airflow program was also not created in the same way as actual air, but it can simulate it to the degree that we use it to build actual planes. A computer program might not have been 'born' in the same way that we are, but we can make it simulate the "way that we are" (giving enough computer power, down to the atom or quantum element). So?

Or apply modus tollens and infer our consciousness is not reducible to our meat machines.

by yters 7y ago

Or apply modus tollens and infer our consciousness is not reducible to our meat machines.

Nobody said it is: it's in the interconnections and stored info inside the meat machines. A T-bone is also meat but without consciousness. Except if you think it's in an external, not meat/neurons/info soul.

TheOtherHobbes 7y ago
Substrate independence - whether it's connections or something else - is a very strong claim that should be subjected to empirical testing before being accepted. (And IMO is almost certainly wrong for good practical reasons that have nothing to do with souls.) Of course you can't test it empirically, because there's no empirical test for subjective experience. You can only test for the appearance of subjective experience - specifically for external behaviours, loosely defined, which might be considered to imply subjective experience to you as an observer. This isn't a solvable problem, and Bostrom is begging the question by trying to hand wave it away with a sneaky "If we just accept this premise..."
- coldtea 7y ago
  >Substrate independence - whether it's connections or something else - is a very strong claim that should be subjected to empirical testing before being accepted. (And IMO is almost certainly wrong for good practical reasons that have nothing to do with souls.) Care to mention any of those "good practical reasons"? Penrose gave it a try (though even there, there was nothing about how it can't be modelled with code for example, in which case, rocks are fine too for a simulation a la XKCD, just not as a direct substrate for an Android). >Of course you can't test it empirically, because there's no empirical test for subjective experience. You can only test for the appearance of subjective experience - specifically for external behaviours, loosely defined, which might be considered to imply subjective experience to you as an observer. It's not really that we can't test it, it's more that we can't measure it. But we can, and do, test it, e.g. with psychological tests, the same kind we give to people to understand different aspects of their personality (which we also can't just measure with some equipment or simple measurement). If it quacks like a duck, and looks like a duck, and cries like a duck, and has other sentimental duck qualities, it's a duck. We only accept the subjective experience of other people from its appearance (their actions etc) too, anyway.
- krageon 7y ago
  > good practical reasons The only reasoning I have seen on this subject so far has been "people are special" or "God made me". Sometimes the arguments are dressed up more. Do you have a good source that has a different type of reasoning? Edit: > You can only test for the appearance of subjective experience Yes, and we solve this in people by assuming that if they sound, act and look like a person they are a person. I've never seen a compelling argument why it is somehow wrong to apply this principle to everything else, except of course the huge trouble philosophers go to to then argue what exactly needs to be looked at before we are satisfied. For some reason the Turing test is not adequate, but to my mind it's only not accurate if you believe that you are somehow special and that there is more to you than what can be observed.
  - Wowfunhappy 7y ago
    > For some reason the Turing test is not adequate Let's say I build a chatbot with a predetermined database of perfect responses to every possible string of words. This is quite clearly possible in theory, given a database with infinite room. This chatbot would pass the touring test, even though all it's doing is retrieving prewritten sentences. Is it conscious? Ergo, it's quite clearly possible to build an automaton which acts like a human, but actually has no clue what it's saying. Which means the touring test isn't adequate. --- Here's another take: You have no proof that anyone in the universe is conscious other than yourself. However, given that you were born in the same way as everyone else, and you're conscious, everyone else probably is too. Why would you be so special? Well, a computer program was not born in the same way that you were, so that assumption goes out the window.
    - coldtea 7y ago
      >Let's say I build a chatbot with a predetermined database of perfect responses to every possible string of words. This is quite clearly possible in theory, given a database with infinite room. This chatbot would pass the touring test, even though all it's doing is retrieving prewritten sentences. Is it conscious? For all intents and purposes, yes. Though without actual consciousness it would be impossible to build it. >Well, a computer program was not born in the same way that you were, so that assumption goes out the window. Well, a computer airflow program was also not created in the same way as actual air, but it can simulate it to the degree that we use it to build actual planes. A computer program might not have been 'born' in the same way that we are, but we can make it simulate the "way that we are" (giving enough computer power, down to the atom or quantum element). So?