1. Post #81
    Gekkosan's Avatar
    October 2010
    5,668 Posts
    I thought that we established that free will doesn't exist
    Humans still are more free to think and act and do stuff than any other animal that cannot even comprehend any of this. It is our brain that gives us this almost like free will that we have.

    Like, surely there's nothing chemically significant going on in my brains when I choose to write down "herpity derp"

  2. Post #82
    Bawbag's Avatar
    December 2011
    530 Posts
    Humans still are more free to think and act and do stuff than any other animal that cannot even comprehend any of this. It is our brain that gives us this almost like free will that we have.

    Like, surely there's nothing chemically significant going on in my brains when I choose to write down "herpity derp"
    It is however, fully predictable.

  3. Post #83
    Gekkosan's Avatar
    October 2010
    5,668 Posts
    It is however, fully predictable.
    What is? Man's actions? Not exactly "fully predictable"

  4. Post #84
    Gold Member
    Secrios's Avatar
    March 2011
    1,040 Posts
    Is it ethical to sleep with your artificial organism?

  5. Post #85
    What is? Man's actions? Not exactly "fully predictable"
    In principle, not in practice.

  6. Post #86
    What is? Man's actions? Not exactly "fully predictable"
    The Universe is deterministic, but that doesn't mean it's predictable.

  7. Post #87
    The Universe is deterministic, but that doesn't mean it's predictable.
    Eh, what? Thanks to quantum mechanics, as far as we know the universe is not deterministic.

  8. Post #88
    Gold Member
    chonks's Avatar
    April 2009
    1,138 Posts
    I don't think that human thought is compatible with binary so as long as we're using conventional computers, human-like sentience isn't likely.

    Eh, what? Thanks to quantum mechanics, as far as we know the universe is not deterministic.
    But that's not for sure, it could just be progressively more unobservable things influencing our environment.

  9. Post #89
    Eh, what? Thanks to quantum mechanics, as far as we know the universe is not deterministic.
    many worlds

  10. Post #90
    Gold Member
    Dennab
    June 2009
    17,111 Posts
    If we create a human being in a virtual world, like the human works exacly like us and has feelings and all. If we cause it pain for no reason, would that be immoral?

  11. Post #91
    let's stick to interpretations which physicists take seriously shall we

    Edited:

    it's at best conjectural still and at worst fundamentally unfalsifiable (and therefore completely unscientific)

    Edited:

    But that's not for sure, it could just be progressively more unobservable things influencing our environment.
    Bell's theorem shows that no local hidden variable theory can accurately account for all of the phenomena of quantum mechanics.

  12. Post #92
    let's stick to interpretations which physicists take seriously shall we
    okay

    many worlds it is.

    it's at best conjectural still and at worst fundamentally unfalsifiable (and therefore completely unscientific)
    don't the other interpretations propose ridiculous things as well?

  13. Post #93
    okay

    many worlds it is.
    Yeah no. MWI is not accepted by the majority of physicists. Copenhagen has prevailed and still does for decades. (along with decoherence) MWI is basically little more than an interesting thought at the moment.

    Anything Michio Kaku takes seriously is probably considered at least a tad "out there" by the rest of the physics community

    don't the other interpretations propose ridiculous things as well?
    No. Not nearly by comparison anyway. Got any examples of what you mean?

  14. Post #94
    No. Not nearly by comparison anyway. Got any examples of what you mean?
    collapse postulates assume more than branching.

    oh and for what it's worth, there are plenty of physicists that take MWI seriously, but that's argument from authority/bandwagon

  15. Post #95
    collapse postulates assume more than branching.
    How do you claim that

    oh and for what it's worth, there are plenty of physicists that take MWI seriously, but that's argument from authority/bandwagon
    Sure there are, but they're not in the majority. What concerns me most is that you just said, "Reality is deterministic," with no explanation that that was working under a certain assumption with which the majority of physicists would disagree, as though it were matter of fact.

    Edited:

    Also, wavefunction collapse doesn't really assume anything at all. It is an observed phenomenon. The way MWI differs is that it claims the collapse is only apparent, and there is more at work.

  16. Post #96
    Also, wavefunction collapse doesn't really assume anything at all. It is an observed phenomenon. The way MWI differs is that it claims the collapse is only apparent, and there is more at work.
    if we went by that logic we'd still use epicycles to predict planet movements

  17. Post #97
    Never mind this is getting too off-topic.

  18. Post #98
    alright I admit I was rude, I'm sorry.

    I said about epicycles because at the time, the new heliocentric model was criticised because it didn't really describe or predict anything new. It just predicted the existing data, perhaps to a greater degree of accuracy, but nothing really revolutionary. I think "saved appearances" was the phrase used at the time.

    In the epicyclic model, the apparently zigzaggy motions of the planets in the night sky exist explicitly in the model, while in the heliocentric model, the zigzaggy motions are the result of us being on one orbit while observing others. As such they are only there implicitly.

    Even if the heliocentric model offered no practical advantage in making predictions - if we lived in a world in which the Ptolemaic model described everything just as well as heliocentrism, it would still make sense to consider the latter more correct, because it assumes less.

    In this vein, I will just say that the relative advantages of MWI are implicit in my metaphor.

  19. Post #99
    Let's take this to the science thread, this is too tangentially related to AI to be arguing here anymore

  20. Post #100
    Let's take this to the science thread, this is too tangentially related to AI to be arguing here anymore
    good idea

  21. Post #101
    matsta's Avatar
    September 2009
    347 Posts
    alright I admit I was rude, I'm sorry.

    I said about epicycles because at the time, the new heliocentric model was criticised because it didn't really describe or predict anything new. It just predicted the existing data, perhaps to a greater degree of accuracy, but nothing really revolutionary. I think "saved appearances" was the phrase used at the time.

    In the epicyclic model, the apparently zigzaggy motions of the planets in the night sky exist explicitly in the model, while in the heliocentric model, the zigzaggy motions are the result of us being on one orbit while observing others. As such they are only there implicitly.

    Even if the heliocentric model offered no practical advantage in making predictions - if we lived in a world in which the Ptolemaic model described everything just as well as heliocentrism, it would still make sense to consider the latter more correct, because it assumes less.

    In this vein, I will just say that the relative advantages of MWI are implicit in my metaphor.
    OF COURSE that must be true for scientific models. They are designed to explain all existing phenomena at the least expense, and that is ALL they should be designed for.

    But if you're expecting a scientific model to predict/explain transcendent things then you will have trouble even proving they do.

  22. Post #102
    OF COURSE that must be true for scientific models. They are designed to explain all existing phenomena at the least expense, and that is ALL they should be designed for.

    But if you're expecting a scientific model to predict/explain transcendent things then you will have trouble even proving they do.
    What do you mean by "transcendent"

  23. Post #103
    matsta's Avatar
    September 2009
    347 Posts
    What do you mean by "transcendent"
    Things that are not phenomena (things as they appear to us) but are things in-themselves

  24. Post #104
    give an example, that's hella vague

  25. Post #105
    Gold Member
    FlakAttack's Avatar
    November 2006
    5,822 Posts
    Why is everyone here so focused on a so-called human-level AI?

    What the fuck kind of AI researcher wants to focus on something so pointless as a machine that can think like a human? If you want something that can think like a human, get laid and wait 9 months.

    I don't think people really understand the sheer danger of making an AI with a greater-than-human intelligence, which is much more likely to come about.
    You think it's aiming low but actually, designing a computer to emulate the human brain is likely more difficult than continuing the standard tree of computer development, even past a better-than-human level of reasoning.

    This development works with two key outcomes in mind: the first to explore new technological options, and the second to better understand our own brains and how they work when it comes to learning and decision making.

  26. Post #106
    Impact1986's Avatar
    December 2011
    1,662 Posts
    It all depends what your definition of artificial intelligence is. Is it just some scripts slapped together to simulate certain behaviours? Or are you talking about sentient beings with full conscience of themselves?

  27. Post #107
    Bat-shit's Avatar
    October 2010
    13,316 Posts
    A pretty much fully aware bio-"synths" that remember and learn and could assume a role in our world with human-like legs and body would be a.. great fucking view. And probably possible, and definitely ethic because it would be fun, hopefully for the robots themselves too.

    But I think for now there's just those massive assembly robots writing circuit boards or assembling car parts at a faster pace, scripted and coded to do just that or it would be a useless piece of metal.

    So, we have artificial intelligence in use and it's getting better and better, isn't it?

  28. Post #108
    matsta's Avatar
    September 2009
    347 Posts
    give an example, that's hella vague
    Ok, phenomena are the appearances of the objects, for example, the colour 'red' is a phenomenon, because it doesn't really exist outside of our experience. Actually, everything concerning our experience of the world (or 'the world as we experience it') is phenomena. Science also falls into this category because it's studies are based on empiric evidence (evidence of experience also includes measuring).

    Noumena, on the other side, are things on-themselves, existing without us perceiving them. Most philosophers agree that, if noumena exists we may not be able to know its nature at all. Think about it this way: there needs to be an observer to know noumena, but everything that is observed is (sort of) "made up" by the nature of the observer, and is phenomena, so we can't really 'observe' noumena.

  29. Post #109
    so phenomena are a subset of noumena?

    Edited:

    unless you're postulating souls

  30. Post #110
    matsta's Avatar
    September 2009
    347 Posts
    so phenomena are a subset of noumena?

    Edited:

    unless you're postulating souls
    No, you must think about them as being "different".

    Edited:

    Although it is understood that noumena is the cause of phenomena.

  31. Post #111
    matsta's Avatar
    September 2009
    347 Posts
    And if I keep explaining we're going to go off topic, so I'll leave it like that and will start discussing about it when Facepunch starts having a philosophical discussion (never, lol).

    And about machines having "better reasoning" than humans, I think we've already done that. Actually, some of the most elementary computers can perform complex logic operations in a much shorter time than a human would take to perform them.

    Of course, that doesn't count as an AI machine. But some sort of AI machines have been created, they just lack consciousness (lol).

  32. Post #112
    Bat-shit's Avatar
    October 2010
    13,316 Posts
    snip(oh god why i dont know)

  33. Post #113
    Gold Member
    Mindtwistah's Avatar
    July 2008
    3,753 Posts
    Why is it impossible? All our brain is is a powerful and evolved carbon based computer with a series of inputs and specialised regions, exactly like a computer is. The only real difference is that we are carbon based and highly refined thanks to a few hundred thousand years of evolutionary pressures.

    And if the AI isn't computer based then it'll be alive and thus just intelligence and not artificial intelligence.
    Actually, the brain isn't a computer, it's a neural network. There's a big difference.


    Can you justify this post further
    I believe he means (and if he does, I agree with him) that you can not create artificial intelligence with traditional programming, it has to be done with neural networks. Programming is purely instruction while neural networks are made to adapt and form patterns according to the environment and input, which is what we know as intelligence.
    You could however simulate a highly advanced neural network like they're doing in the Blue Brain project right now.

  34. Post #114
    just because neural networks were the design that evolution happened to come up with doesn't give them some kind of magical power that other architectures lack

    Edited:

    No, you must think about them as being "different".

    Edited:

    Although it is understood that noumena is the cause of phenomena.
    the hell do you mean

    minds aren't magical, experienced phenomena are entirely physical

  35. Post #115
    Gold Member
    Mindtwistah's Avatar
    July 2008
    3,753 Posts
    just because neural networks were the design that evolution happened to come up with doesn't give them some kind of magical power that other architectures lack
    I agree, there might be other architectures that were either not evolutionary advantageous compared to a neural network or just never happened/were not possible to make with the materials given.
    I suppose I expressed myself wrong in the previous post. What I meant was that you can't make a "real" intelligence with traditional programming.

  36. Post #116
    I agree, there might be other architectures that were either not evolutionary advantageous compared to a neural network or just never happened/were not possible to make with the materials given.
    I suppose I expressed myself wrong in the previous post. What I meant was that you can't make a "real" intelligence with traditional programming.
    how do you know

    right now, nobody has made "real" intelligence with traditional programming or neural nets. what makes you think neural nets will fare better? remember, we've only been at this for about 40 years now.

  37. Post #117
    flyguy88's Avatar
    July 2007
    336 Posts
    I think this could be ethical if you made it so the 'AI' did not have sensations or negative mental states.

  38. Post #118
    Gold Member
    Mindtwistah's Avatar
    July 2008
    3,753 Posts
    how do you know

    right now, nobody has made "real" intelligence with traditional programming or neural nets. what makes you think neural nets will fare better? remember, we've only been at this for about 40 years now.
    Because, as I said earlier, traditional programs can not change themselves and adapt, they can only follow the set-out instructions of their programming. You'd need to program every single behavior and action, and it would never be able to go outside of its programming. It would not be able to think, feel or learn new things, it would just follow the code.
    With neural nets we've already made great progress. We've created several minor neural networks that can adapt, remember, learn e.t.c.

  39. Post #119
    uhm not really there are people working on those problems right now. granted they're not solved yet, but still neural nets aren't immune to them either

    you forget that the vast majority of the animal kingdom use neural networks in their brains, and they can barely learn anything either. you can count on one hand the number of animal species that can think creatively

    Edited:

    human minds are the exception rather than the rule

  40. Post #120
    matsta's Avatar
    September 2009
    347 Posts
    just because neural networks were the design that evolution happened to come up with doesn't give them some kind of magical power that other architectures lack

    Edited:



    the hell do you mean

    minds aren't magical, experienced phenomena are entirely physical
    Actually, nothing in your experience is outside of it (nothing you are experiencing now exists in the allegedly "real"/outer/"physical" world as you are experiencing it), so you should really reconsider the word "physical".