[FoRK] Kudos for Jeff Hawkins' _On Intelligence_

Jeff Bone jbone at place.org
Thu May 26 07:56:16 PDT 2005


Per Eugen's comments...


Ah, now I get it.  Your general beef appears to be that because this  
isn't (yet, maybe ever) usable as a template for GAI, it's not  
interesting to you.  Well, I can't argue with that.  I'm more  
interested in general progress than in applications *at this point.*   
I don't even think Hawkins is doing computer science, really, though  
there will be applications in that direction.  He's researching  
something very different.  A secondary beef seems to be a gut  
reaction to the "hyped" / "dumbed-down" pop presentation.  And that's  
understandable, too, and I think I even mentioned that.  In fact the  
presentation is one reason I didn't read the book until now.  (I have  
a similar reaction to ANKOS, btw.)

Per your specific word salad, I guess the crux of the difference  
between your position and Hawkins is this:  maybe it's a lot simpler  
than you / we have been thinking.  Maybe all you need is a high-level  
functional simulation of a neuron;  take enough of those, wire 'em up  
right, and you've got something approximating a brain.  Maybe you  
don't even need a detailed model of neural function at the sub- 
cellular level;  maybe a simple approximation of electrical spiking  
and signal propagation are all that's needed.  Maybe it's not that  
simple, but wouldn't we all feel foolish if we continue perpetually  
tossing that idea out on the basis of some intution that "it just  
can't be that simple" --- and later learned that, in fact, it was?

jb





More information about the FoRK mailing list