Ben,

On Tue, Sep 4, 2012 at 2:07 PM, Ben Goertzel <[email protected]> wrote:

>
>
> My point was that it is damn near impossible to even make a case for
>> designing in emergent properties.
>>
>> *How would YOU do such a thing*, other than just trying zillions of
>> semi-random experiments?
>>
>
> Via theoretical understanding of the type of system one is building...
>

But, for emergent properties, this is equivalent to having done the math to
do this, which we have previously agreed is a BIG job probably requiring
years of work at a research institute.

>
>> Have you thought about a GA "supervisor" for OpenCog, to replace current
>> developers?!!! If the prospect of an OpenCog-descendant AGI is workable
>> (you already know my opinion), this might provide your only realistic shot
>> at it. As in the previous paragraph, you will first need the measure for
>> intelligence as part of this.
>>
>
> Currently, automated program learning works OK for LISP programs
> containing up to a few dozen nodes in their corresponding program trees.
> That's way too small to use GP or any other automated program learning
> methods to evolve AGIs.
>

Agreed. We are still a LONG way from being able to do this.

>
> One would need an AGI evolutionary program learning system, to evolve
> programs big enough to be AGIs ;)
>

My (and I think your) hope is to be able to develop tiny systems with a
hint of what you want on large systems. The trick would be in even
recognizing <1% of an AGI.

Steve



-------------------------------------------
AGI
Archives: https://www.listbox.com/member/archive/303/=now
RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-c97d2393
Modify Your Subscription: 
https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-2484a968
Powered by Listbox: http://www.listbox.com

Reply via email to