I am repeating myself, but just in case you missed something.
1. I think that it would be easy to teach a text-only AI program (that
I created) a rudimentary programming language (like a simple database
language) through a trial and error method. It would be very
rudimentary.
2. If it was possible to teach the program a Simple Context-Free
programming language that way, then it would be possible to teach it a
rudimentary Context-Sensitive language (a non Context-Free language)
through a trial and error method.
3. If that were so then I believe I could use the rudimentary human
language to teach it something about the world. That would be AGI. AGI
does not have to be human level AGI. However, it does have to show
that it can learn new things and build on that learning.
4. If that were so then I believe that the programming model could be
adapted for other IO modalities. Therefore the program would be AGI
even if it is extremely limited (and somewhat tricky to use).

I am interested in studying how a computer might be capable of doing
some genuine learning. That emphasis on learning, the capability of
learning about things that it had it had not known before and the
ability to build on that learning, is what I am interested in.


-------------------------------------------
AGI
Archives: https://www.listbox.com/member/archive/303/=now
RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424
Modify Your Subscription: 
https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657
Powered by Listbox: http://www.listbox.com

Reply via email to