Restless Souls/Technology: Difference between revisions

m
no edit summary
mNo edit summary
mNo edit summary
Line 153: Line 153:
: The Blue Brain Project ignored glia cells and similar supporting elements. The BRAIN Initiative [http://www.nature.com/news/neuroscience-map-the-other-brain-1.13654 did better] but still had no real neuronal input and no information from cells that messure concentration of oxygen, sugar, etc.
: The Blue Brain Project ignored glia cells and similar supporting elements. The BRAIN Initiative [http://www.nature.com/news/neuroscience-map-the-other-brain-1.13654 did better] but still had no real neuronal input and no information from cells that messure concentration of oxygen, sugar, etc.


: And there are a lot [[wikipedia:Autonomic_nervous_system#Sensory_neurons|neuros in the human body]]. So it became more and more clear that reducing a brain to its own neurons was far from enough.
: And there are a lot [[wikipedia:Autonomic_nervous_system#Sensory_neurons|neurons in the human body]]. So it became more and more clear that reducing a brain to its own neurons was far from enough.


: After they had a full body simulation up and running the next step was to educatate the AI and let it control a body in the physical world.
: After they had a full body simulation up and running the next step was to educatate the AI and let it control a body in the physical world.
Line 1,153: Line 1,153:
Right now it looks like [http://phys.org/news/2014-08-optalysys-prototype-optical-processor.html optical] and [http://phys.org/news/2014-08-tiny-chip-mimics-brain-supercomputer.html#ajTabs neurosynaptic] processors will hit the market before memristor- and spintronic-based processors do.
Right now it looks like [http://phys.org/news/2014-08-optalysys-prototype-optical-processor.html optical] and [http://phys.org/news/2014-08-tiny-chip-mimics-brain-supercomputer.html#ajTabs neurosynaptic] processors will hit the market before memristor- and spintronic-based processors do.


If we assume a human neocortex with 1 million cortical columns each with 60.000 neurons, it would take 60.000 of IBM's new neurosynaptic processors to do the job.
If we assume a human neocortex with 1 million cortical columns each with 60.000 neurons, it would take 60.000 of IBM's new neurosynaptic processors each with 1.000.000 programmable neurons to do the job.


A SLD brain with components based on these technologies might be possible but would be still far to big in size. It might fill a little building and remote control its body. So that's generation 0.
A SLD brain with components based on these technologies might be possible but would be still far to big in size. It might fill a little building and remote control its body. So that's generation 0.
8,452

edits