> And the idea of a "patchability" is like many of your first intuitions an
> excellent and potentially true AGI concept (only then you always destroy
> them by  thinking of what your peers would do/accept).

Heh... no, I think I sometimes agree with your descriptions of subjectively
perceived mental processes...

However, I think that underlying algorithmic processes can give rise to these
subjectively perceived mental processes -- and the underlying algorithms may
be quite different than the subjective experiential perceptions

This is one place we seem to part ways.  You seem to take your introspective,
subjective feeling of what your mind does as the main guide to what is happening
in your brain....  But I think it's a very partial guide...

> Programs are in reality patched together rather than springing from fully
> formed prior programs (an endless regression).  So the idea of a program
> that would include "patchabilities" - a part that says "something to be
> patched in here later" is rather an interesting one

That wouldn't be hard to do within MOSES or other automated program
learning algorithms.  Basically you would have some MOSES program nodes
that were just specifications without any underlying realization.  Then you'd
count on the specifications somehow being filled in later.  This would also
be do-able in OpenCog just via interaction btw PLN and MOSES though, even
without specifically tweaking MOSES to enable it.  (Yes I know Mike T can't
understand these details, but maybe some other readers can ;)

-- ben g


-------------------------------------------
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