6 ms·

Will Uploaded Minds in Machines be Alive?

9 points by r11t 17y ago

jacquesm 17y ago
That makes the very large assumption that such a thing will be possible in the first place, which is by no means a certainty.
- thwarted 17y ago
  Do all base assumptions need to be met in order to exercise a thought experiment? Sometimes it's best to explore these philosophical questions before answering them is forced upon us.
- rbanffy 17y ago
  Since there are no physical impossibilities involved, it's just a question of time until it is possible to completely acquire and simulate an organic brain on a computer model.
  - jacquesm 17y ago
    How much time though ? It's only a matter of time before we make contact with an alien civilization. But that doesn't mean we need to go and wonder overly long about how we'd dress for the occasion, it is simply too far off for that. What kind of an upper limit would you put on this kind of thing becoming a reality?
    - rbanffy 17y ago
      A hundred years, at most. The problem is not the computers to run the simulation, but capturing enough of the state of an organic brain to run a consistent simulation. That is the tricky part. The brains we simulate today are currently running on clusters of machines designed for word processing. If and when we decide to take brain simulation seriously, we will probably do so on machines designed for that purpose.
      - jacquesm 17y ago
        That's outside the betting territory, too bad! I'd gladly taken you up on anything less than 20 years, a hundred is far out enough that we might run in to some kind of spectacular development. 10^14th connections is quite an impressive number to map, especially if it turns out that the state is stored at the molecular level (we have DNA as a hint of how dense nature tends to pack information). Henry Makram is on the record as saying it will take the blue brain project less than 10 years, in a TED talk given the 22nd of July 09.
    - wmf 17y ago
      Here's a detailed roadmap: http://www.fhi.ox.ac.uk/selected_outputs/fohi_publications/brain_emulation_roadmap http://www.fhi.ox.ac.uk/selected_outputs/fohi_publications/b... It supports your claim that uploads are pretty far in the future.
  - scott_s 17y ago
    We still don't know what consciousness is. So I think it's naive to state that it's "just a question of time" before we can simulate it.
    - rbanffy 17y ago
      Now knowing exactly what it is does not preclude one from speculating whether it could be simulated on a computer. Unless consciousness involves some magical phenomena, it's all physics and physics can be simulated on a computer.
      - scott_s 17y ago
        Saying "it's just a matter of time" is not speculating whether it could be simulated. It's assuming that it can, and saying that it's inevitable. Consciousness is probably an emergent phenomenon of some kind. But until we know exactly what physical processes cause the emergence, then we have no way of knowing if our simulation will encompass it. Without that knowledge it's just a guessing game if our simulation is sufficient. My own personal guess is that attaining that knowledge will require an upheaval somewhere in science orthodoxy - much like solving the black body radiation problem required an upheaval in physics in the form of quantum mechanics. Such disruptive discoveries are inherently unpredictable. Hence, I think it's naive to claim that simulating consciousness is inevitable since we can't predict when we'll even understand it.
        rbanffy 17y ago
        But if it's only physics, then it can be simulated and if it can be simulated, then it is a matter of time until someone cracks it and makes it work. Your mind is already running on your meatware brain. There are no magical properties to it. We may not understand them all fully, but that too is a matter of time, since there are no magical entities at work within it either.
        scott_s 17y ago
        We're going in circles. My point, stated as simply as I can: since understanding what exactly consciousness is will probably require an upheaval in scientific knowledge, it is naive to say it is "simply a matter of time." It's possible that we will never figure it out, that the necessary scientific upheaval will never happen. Further, since what is required is unknown, it's possible that nothing we are ever capable of building will be up to the task. I am only willing to use the phrase "it's a matter of time" on scientific and technological improvements that are iterative. This is not iterative. It is disruptive. Also, it's not a given that all physics can be simulated. We have never encountered a non-computable physical phenomenon, but we also can not prove that all physical phenomenon is computable.
pwmanagerdied 17y ago
Though the topic is interesting, this article is rambling and senseless. For example it states that life, by reducing local entropy, violates the second laws of thermodynamics.

9 points by r11t 17y ago

That makes the very large assumption that such a thing will be possible in the first place, which is by no means a certainty.

thwarted 17y ago
Do all base assumptions need to be met in order to exercise a thought experiment? Sometimes it's best to explore these philosophical questions before answering them is forced upon us.
rbanffy 17y ago
Since there are no physical impossibilities involved, it's just a question of time until it is possible to completely acquire and simulate an organic brain on a computer model.
- jacquesm 17y ago
  How much time though ? It's only a matter of time before we make contact with an alien civilization. But that doesn't mean we need to go and wonder overly long about how we'd dress for the occasion, it is simply too far off for that. What kind of an upper limit would you put on this kind of thing becoming a reality?
  - rbanffy 17y ago
    A hundred years, at most. The problem is not the computers to run the simulation, but capturing enough of the state of an organic brain to run a consistent simulation. That is the tricky part. The brains we simulate today are currently running on clusters of machines designed for word processing. If and when we decide to take brain simulation seriously, we will probably do so on machines designed for that purpose.
    - jacquesm 17y ago
      That's outside the betting territory, too bad! I'd gladly taken you up on anything less than 20 years, a hundred is far out enough that we might run in to some kind of spectacular development. 10^14th connections is quite an impressive number to map, especially if it turns out that the state is stored at the molecular level (we have DNA as a hint of how dense nature tends to pack information). Henry Makram is on the record as saying it will take the blue brain project less than 10 years, in a TED talk given the 22nd of July 09.
  - wmf 17y ago
    Here's a detailed roadmap: http://www.fhi.ox.ac.uk/selected_outputs/fohi_publications/brain_emulation_roadmap http://www.fhi.ox.ac.uk/selected_outputs/fohi_publications/b... It supports your claim that uploads are pretty far in the future.
- scott_s 17y ago
  We still don't know what consciousness is. So I think it's naive to state that it's "just a question of time" before we can simulate it.
  - rbanffy 17y ago
    Now knowing exactly what it is does not preclude one from speculating whether it could be simulated on a computer. Unless consciousness involves some magical phenomena, it's all physics and physics can be simulated on a computer.
    - scott_s 17y ago
      Saying "it's just a matter of time" is not speculating whether it could be simulated. It's assuming that it can, and saying that it's inevitable. Consciousness is probably an emergent phenomenon of some kind. But until we know exactly what physical processes cause the emergence, then we have no way of knowing if our simulation will encompass it. Without that knowledge it's just a guessing game if our simulation is sufficient. My own personal guess is that attaining that knowledge will require an upheaval somewhere in science orthodoxy - much like solving the black body radiation problem required an upheaval in physics in the form of quantum mechanics. Such disruptive discoveries are inherently unpredictable. Hence, I think it's naive to claim that simulating consciousness is inevitable since we can't predict when we'll even understand it.
      - rbanffy 17y ago
        But if it's only physics, then it can be simulated and if it can be simulated, then it is a matter of time until someone cracks it and makes it work. Your mind is already running on your meatware brain. There are no magical properties to it. We may not understand them all fully, but that too is a matter of time, since there are no magical entities at work within it either.
        scott_s 17y ago
        We're going in circles. My point, stated as simply as I can: since understanding what exactly consciousness is will probably require an upheaval in scientific knowledge, it is naive to say it is "simply a matter of time." It's possible that we will never figure it out, that the necessary scientific upheaval will never happen. Further, since what is required is unknown, it's possible that nothing we are ever capable of building will be up to the task. I am only willing to use the phrase "it's a matter of time" on scientific and technological improvements that are iterative. This is not iterative. It is disruptive. Also, it's not a given that all physics can be simulated. We have never encountered a non-computable physical phenomenon, but we also can not prove that all physical phenomenon is computable.

pwmanagerdied 17y ago

Though the topic is interesting, this article is rambling and senseless. For example it states that life, by reducing local entropy, violates the second laws of thermodynamics.