You link to :
Jeff Hawkins - On Intelligence.pdf 1.5 MB 174 pages !
…will read it and come back much later (many times) to update this comment.
.1) earliest date in that document is 2003, this is already somewhat outdated !
.2) entrepreneur’s & pragmatic approach to the question of creating G.A.i.
.3) Mountcastle’s paper : Rosetta stone of neuroscience (exactly the same 6 layer neurone structure in every part of the cortex would produce a single workflow)
My search: “Redwood Neuroscience Institute” ==> humm, they are keeping silent for 20 years, meaning they might be (must be) with the military now. Quite worrying is this ?
It’s been a while since I read it. The main parts that stuck with me were about the approach of trying to map out brain structures and figure out what they do. Then implement these in a neural network and see if they produce similar results. This seems like the right approach for trying to figure out how to implement a biologically inspired AI. And the part about the 6 layered structure being repeated in the cortex is very encouraging. If it is a single repeating structure, then there’s a good chance we’ll be able to figure out the algorithm there.
There is some evidence that the brain has a common recurrent structure to it. On Intelligence is a really fun read about this.
You link to :
Jeff Hawkins - On Intelligence.pdf 1.5 MB 174 pages !
…will read it and come back much later (many times) to update this comment.
.1) earliest date in that document is 2003, this is already somewhat outdated !
.2) entrepreneur’s & pragmatic approach to the question of creating G.A.i.
.3) Mountcastle’s paper : Rosetta stone of neuroscience (exactly the same 6 layer neurone structure in every part of the cortex would produce a single workflow)
My search: “Redwood Neuroscience Institute” ==> humm, they are keeping silent for 20 years, meaning they might be (must be) with the military now. Quite worrying is this ?
It’s been a while since I read it. The main parts that stuck with me were about the approach of trying to map out brain structures and figure out what they do. Then implement these in a neural network and see if they produce similar results. This seems like the right approach for trying to figure out how to implement a biologically inspired AI. And the part about the 6 layered structure being repeated in the cortex is very encouraging. If it is a single repeating structure, then there’s a good chance we’ll be able to figure out the algorithm there.
i agree & many thanks again 👍