3 ms·

It becomes much easier to understand the AGI threat if you consider whole brain emulations -- human minds running at superhuman speed on silicon. Hollywood has

It becomes much easier to understand the AGI threat if you consider whole brain emulations -- human minds running at superhuman speed on silicon. Hollywood has made it obvious, if it wasn't already, the serious implications of that type of development. Whole brain emulations also make it clear that the issue has nothing at all to do with human values: a human superintelligence would be a catastrophic risk to humanity.

rwallace 12y ago
Actually, a whole brain emulation would, at least for some generations after the first working model, run slower, not faster, than an organic brain. Bear in mind that without whole brain emulation (or nanomedicine sufficiently advanced to confer biological immortality, which is an even considerably more difficult engineering challenge), you're going to die. So are all your family and friends. Everyone you ever cared about. Being confused about the difference between Hollywood and real life is not harmless.
- foobarqux 12y ago
  The point was that everyone seems to have trouble understanding why software AGI might be dangerous but a much easier time why whole brain emulation is, not that Hollywood is a representation of reality.
Balgair 12y ago
Whole Brain Emulation is far off, even putting in moore's law, because we still are far off from understanding the brain. Yes, optogenetics and the connectome are going to bring us a long way there, but at the end, we have to understand that the brain is far to plastic to really know at any time point what is going on. Experiments with macaque motor cortex (M1) provide a great example. When you cut off a finger or toe and record from M1 at the same time, you can watch the synapses change in real time. The remapping of M1 in the monkey starts the second that the nerves are severed. This remapping is poorly understood as it stands. Typically, the brain will allow innervation of the areas the control the other digits into the now useless area where the missing digit is. Even more interesting is when a pair of monkeys were stolen by anti-vivisectionists (The silver spring incident, I think [0]). When returned, the monkey's M1 had totally remapped from it's original state. Hippocampal neurogenesis [1] is also very poorly understood as it is today. Why our brains make new neurons and mostly for the more short-term memory centers in the hippocampus is a mystery. What I am getting at is that we do not understand our own intelligence from the standpoint of it's mechanics enough to emulate a brain near at all. It is a looooooong ways off. To truly emulate it, you have to know that it is a moving target, and therefore any emulation is a moving one as well. One that has to live in a body that you emulate in a world that impinges on it. Trying to take the mind out of the body is impossible as well as trying to take both then out of the world. The results would have no meaning. [0]https://en.wikipedia.org/wiki/Silver_Spring_monkeys https://en.wikipedia.org/wiki/Silver_Spring_monkeys [1]https://en.wikipedia.org/wiki/Neurogenesis https://en.wikipedia.org/wiki/Neurogenesis
- foobarqux 12y ago
  The point is that the danger of AGI is much easier to understand through whole brain emulation, not that whole brain emulation is necessarily imminent. That said, the most pernicious risk in AGI is that it will happen "slowly, then all at once"; that it will always be a long way away until the day it isn't. That said, it isn't obvious to what extent we need to understand the brain in order to transfer it to another substrate: much of engineering relies on a simplified understanding of the world.

rwallace 12y ago

Actually, a whole brain emulation would, at least for some generations after the first working model, run slower, not faster, than an organic brain. Bear in mind that without whole brain emulation (or nanomedicine sufficiently advanced to confer biological immortality, which is an even considerably more difficult engineering challenge), you're going to die. So are all your family and friends. Everyone you ever cared about. Being confused about the difference between Hollywood and real life is not harmless.

foobarqux 12y ago
The point was that everyone seems to have trouble understanding why software AGI might be dangerous but a much easier time why whole brain emulation is, not that Hollywood is a representation of reality.

Balgair 12y ago

Whole Brain Emulation is far off, even putting in moore's law, because we still are far off from understanding the brain. Yes, optogenetics and the connectome are going to bring us a long way there, but at the end, we have to understand that the brain is far to plastic to really know at any time point what is going on. Experiments with macaque motor cortex (M1) provide a great example. When you cut off a finger or toe and record from M1 at the same time, you can watch the synapses change in real time. The remapping of M1 in the monkey starts the second that the nerves are severed. This remapping is poorly understood as it stands. Typically, the brain will allow innervation of the areas the control the other digits into the now useless area where the missing digit is. Even more interesting is when a pair of monkeys were stolen by anti-vivisectionists (The silver spring incident, I think [0]). When returned, the monkey's M1 had totally remapped from it's original state. Hippocampal neurogenesis [1] is also very poorly understood as it is today. Why our brains make new neurons and mostly for the more short-term memory centers in the hippocampus is a mystery. What I am getting at is that we do not understand our own intelligence from the standpoint of it's mechanics enough to emulate a brain near at all. It is a looooooong ways off. To truly emulate it, you have to know that it is a moving target, and therefore any emulation is a moving one as well. One that has to live in a body that you emulate in a world that impinges on it. Trying to take the mind out of the body is impossible as well as trying to take both then out of the world. The results would have no meaning. [0]https://en.wikipedia.org/wiki/Silver_Spring_monkeys https://en.wikipedia.org/wiki/Silver_Spring_monkeys [1]https://en.wikipedia.org/wiki/Neurogenesis https://en.wikipedia.org/wiki/Neurogenesis

foobarqux 12y ago
The point is that the danger of AGI is much easier to understand through whole brain emulation, not that whole brain emulation is necessarily imminent. That said, the most pernicious risk in AGI is that it will happen "slowly, then all at once"; that it will always be a long way away until the day it isn't. That said, it isn't obvious to what extent we need to understand the brain in order to transfer it to another substrate: much of engineering relies on a simplified understanding of the world.