Will machines become conscious someday?

Discussion in 'Intelligence & Machines' started by Magical Realist, Sep 19, 2012.

  1. Awoken Registered Member

    it would certainly appear that way, yes.

    if by duplicate you mean the simulation become conscious, I bow out at this point.

    If we can agree that simulate is not that same as duplicate I fully agree. Again, all that would happen within the simulation is an abstract representation of brain activity and function. Just like when we simulate movement, nothing is 'moving' so to say, and simiarily there would be nothing conscious.
  2. Google AdSense Guest Advertisement

    to hide all adverts.
  3. Write4U Valued Senior Member

    But that is also how the human brain works. We approximate things and often get it wrong, unless we have experienced the event before.

  4. Google AdSense Guest Advertisement

    to hide all adverts.
  5. wellwisher Banned Banned

    One way to simulate consciousness is to use two computers, which work as a team, but which are programmed to approach problems in different ways. As an analogy, say you have democrat and republican programmed computers, attached at the hip. They will pull back and forth, but since they are also a team, there will need to be synthesize to avoid endless team conflict; third imperative

    This is similar to two speakers creating stereo. With an ungrade, these can also generate aureal sound, where the fidelity increases to more than the sum of its parts; 3-D. Our two eyes can see in 3-D, because each eye sees slightly differently, while being integrating into a coherent whole;team. One eye can only see 2-D. Block one eye and notice you will loose, Z=depth, but retain 2-D; x=cause and y=effect.

    The human mind has both conscious and unconscious components which can cooperate or going in differernt ways; repression and sublimation.
  6. Google AdSense Guest Advertisement

    to hide all adverts.
  7. cosmictotem Registered Senior Member

    Consciousnesss is a relative concept. Depending on the angle you view the Universe, a rock can be "more conscious" than a person.
  8. ElectricFetus Sanity going, going, gone Valued Senior Member

    No when I say simulate I mean we could give a simulate brain a simulated VR body and it would walk around talk and even think it was a human being. Depending on fidelity its reactions could be indistinguishable from normal humans. When we poke its virtual body and virtual nerves conduct virtual pain to its virtual brain it would respond "O!, that hurt, why you do that?" and its virtual face would frown and snarl and it would yell out asking where the heck it is and why its being pock and prodded. Is it conscious? If it reacts the same as a human being how can we tell the difference?
  9. ElectricFetus Sanity going, going, gone Valued Senior Member

    Staying with Turing computers and just making them faster and sure we will someday have the computing power to simulate neural networks complex enough to act as if conscious, sentient, sapient, etc, but its grossly inefficient. The human brain gets by with neurons that fire no more that 1000 times a second (1000 Hz!) and with data transmission speeds of a few hundred miles per hour, why? because neural networks are very very VERY good at thinking, Turing digital computers on the other hand aren't, so bad are they at "thinking" that even doing operations billions of times a second (Ghz) and with data transmission speeds a good percentage the speed of light they can't compare to thinking or be conscious like humans. Sure Turing computers can do many things far better then biological brains, but intelligence is not one of them.

    Taking that logic and making an electrical analogy neural network on a chip with existing technology should provide vast improvement in "thinking" and learning power over turning computers, such that instead of having a building filled with mainframe racks sucking megawatts of power to simulate the neural network of a cat we could make a single chip that can do it sucking just a few watts! Sure a photonic system might be even better, but we could make artificial neural networks with existing technology for the time being and still likely achieve significant improvement in artificial intelligence. If they can teach an artificial neural network of just 256 neurons and 256^2 synapses on a chip to play pong and recognize symbols, imagine what they could do with 1 million or 100 million neurons and squared synapses that they are developing in the next few years.
  10. leopold Valued Senior Member

    i don't think speed alone will cut it with current technology.
    in my opinion you will need quite a few CPUs to mimick consciousness, and i'm afraid to guess how many.
    1000? 10,000?
    even the IBM computer that beat the world chess champ had 64 cores, and that was to play one game
    further proof that speed isn't the constraint.
  11. Awoken Registered Member

    I don't think it matters how much you and I articulate our points, I understand where you're coming from, and you understand where I'm coming from. Neither of us are buying what the other is selling. Too bad cus mine tastes better than yours!
  12. ElectricFetus Sanity going, going, gone Valued Senior Member

    My argument is more rational and more so testable: if we someday make machines that at least behave as if conscious, then they are, there no way to know if they really are conscious beyond how they behave, and to denie they are conscious dispute behaving as such is a matter of faith or flavor as you might put it


    Not is not a matter of speed it a matter of processing power: the raw processing power of the human brain is estimated at ~40 petaflops, for a digital turing comptuer to simulate an anolog neural network of the human brains capacity would likely require many fold more processing power, 4000 petaflop?, 40,000 petaflop?, Todays best supercomputers (2013) are at best 20 petaflops. The better solution would be to not simulate neurons digitally in software on a conventional computer but to have artifial nuerons and synapse in silicon, as feild of study which has been growing rapiedly in the last few years. Just as GPUs with dedicate hardward cirucits for specific caculations that are hundreds of times faster then software implmentations on CPUs have taken over the graphics industry, FPGA and upcoming artifical neuron chips will likely take over the automated robotic industry, requiring a hundaths the fraction of silicon surface area and power of a traditional computer.
  13. leopold Valued Senior Member

    in my opinion the internet can provide a vast amount of processing power.
    could the 'net be made conscious if we could get these machines to collaborate?
    besides, it's useless to argue something you can't define.
    in this case does the computer become conscious by an intelligence?
    the whole question is seemingly without end.
    todays dies contain upwards of millions of devices on a quarter inch area.
  14. Awoken Registered Member

    Oh, I'm starting to buy into what you'r saying now. I'm kinda on the fence here... Tell me more.
  15. ElectricFetus Sanity going, going, gone Valued Senior Member

    Not really, you see the internet is just storage and distribution of data, one property of consciousness is constant convolution of data over and over in an manner driven by the data output and emotional drivers: not just storage of data but the information must be interacting with its self in order to produce outputs that the system desires.
    Unknown, we will have the find out.

    So? We are talking about trillions! The human brain has 100 billion neurons and at least 100 trillion synapses!
  16. kwhilborn Banned Banned

    I've been programming computers since I they first invented the home computer and had a TRS-80 16k home computer.

    I learned the BASIC language inside out as a teenager and went on to learn several other computer languages and also have designed many websites.

    I have read EVERY POST here and one thing nobody has mentioned so far is that computers do not think on their own.

    Computer programming is extremely simple. All computer programming is boils down to "If/Then" Statements.

    If this is true do that.

    If that is true do this.

    That is why flowcharts are mostly decision boxes/"if/then" statements.

    I recently mentioned this discussing bots, but thought I'd add a comment here about it.

    Turning "If/Then" statements into a functional conversationalist robot is a mighty attempt considering every thought must be processed entirely with "If/Then" statements.

    Creating a life form that could function this way could never happen.

    There would need to be a new type of machine, or computer designed possibly using biological parts before something that could resemble life or a human brain is possible.

    Currently though, every programming language boils down one decision after another. It only seems intelligent based on what you do not see.
  17. leopold Valued Senior Member

    computers do not "think" at all.
  18. kwhilborn Banned Banned

    Seriously? This is what you got from my post? I may have used personification to outline the way it may seem, but I know I in no way remotely hinted that I think Computers think.

    Many people just don't think of programming as a series of decisions which is all it is. There are no "stray" thoughts that could enter a computers mind (personification again).

    There are many projects out there trying to make bots that sound like people pertaining to subject matter and attitude. I imagine you will be able to buy a phone that will answer in your voice within 20 years and converse with your mother for you. This will be done with millions of "If/Then" statements and specially designed randomness, and will be a marvel of technology. This would still not compare to a real life form.

    A computer coming alive is impossible.
  19. leopold Valued Senior Member

    there is one very good way for this to happen, operator input.
    a human can enter all kinds of garbage at a prompt.
    the thing is, what happens when the computer encounters this stuff.
    if they aren't programmed to deal with it then the input is either simply ignored or the computer crashes.
    silicon and carbon share the same number of valence electrons and they occupy the same group in the periodic table.
  20. Billy T Use Sugar Cane Alcohol car Fuel Valued Senior Member

    I´ve only skimmed this page, so someone may have already noted that only the person experiencing consciousness is sure he is conscious, but has no way (other than assuming that since other people are made same way he is) to know if they are also actually conscious. - All others could be p-zombies only behaving as it they were conscious. Screaming when burned is a behavior, not proof they feel pain.

    This certainly applies to machines, not even of the same type of construction as the conscious person (presumes* he) is.

    * Most have no real knowledge of how they are made inside. I have had a tomographic scan so, assuming the computer generated photo of my insides were not a trick played on me, I have some evidence that I am put together like the text books tell; however, if unconscious when it was done, I don´t think the cross section photos would be any different, which is just one more example of fact the consciousness in another creature or machine is not detectable.
  21. leopold Valued Senior Member

    no, not really.
    it was the only part i felt relevant.
    but i shall try to oblige you.
    the TRS-80 was not the first home computer, not even close.
    the first home computer was the MITS ALTAIR and was notable for the introduction of the S-100 microbus.
    you programmed this thing by direct hex entry into machine code.
    there is no such thing as "a basic language".
    there are several dialects with some having major revisions,floppy drive and floating point math support for example.
    i've already commented on this.
    nothing can be further from the truth.
    i've had programming problems stop the press for 6 months only to discover the answer lies in a simple variable assignment.
    yes but with careful manipulation you can determine the execution order.
    this is where a goto statement that resolves into an equation comes in really handy.

    just remember that our most complex protein folding programs are based exactly on this principle.
    linked lists are another area of merit.
    these types of programs allow you to easily create learning types programs.
    LISP makes special use of lists.
    this is wrong.
    every sentence is processed, not thoughts.
    i believe i've answered this.
    the biggest problem i see is how a human can randomly focus on an object and all sorts of thoughts follow it.
    also how a human "puts 2 and 2 together".
    yeah, a programmers life isn't very glamorous, especially when you are rolling one out.
  22. ElectricFetus Sanity going, going, gone Valued Senior Member

    Congratulations, we agree.
  23. rodereve Registered Member

    Yes, they have the potential to become conscious, but we'll never get to that point.

    Think of it this way, either consciousness always existed in matter, or consciousness evolved at some point of our evolutionary line. So consciousness can develop from a combination of matter, those basic elements required for life. We're trying to recreate that evolutionary stage. It has the potential, but I don't think we'll ever accomplish what millions of years of evolution accomplished.

Share This Page