Ben, thank you for clarifying.

Ben Goertzel wrote:
 I like the marketing technique at this mailing list. AGI "developers"
 are claiming that they are building "AGI" but they are just building
 narrow programs.

Personally I am working on both -- the former for R&D purposes and the
latter to make a living ;-p

 The term "artificial general intelligence" is an oxymoron. That term is
 metaphysical, since there is no such thing as "general".

This point is well-understood already.

Hutter's theoretical analyses of AIXI make pretty clear the meaning of
"fully general intelligence" -- which is unambiguous but pragmatically
useless, as it requires infinite computational power.

So in reality one is talking about "degrees of generality" ... absolute
generality being impossible w/in finite computational resources...

I prefer
 something like "human-like intelligence" or the more common term
 "human-level intelligence." Intelligence requires human-like perception.

Actually, I think human-like intelligence is a clear notion -- but places
very narrow constraints on AGI, which not all approaches adhere to
(mine doesn't)

OTOH, human-level intelligence is an even slipperier term than AGI

Most likely the first AGIs will be superhuman in some regards and subhuman
in others, so that whether they're "human level" will be basically a meaningless
question

 What is intelligence? It is the interaction of human-like perception
 with respect to episodic memory.

Well, no.

The interaction of perception and episodic memory is one aspect of
intelligence, not the whole thing.

And human-like perception is not the only feasible kind.

 In the past, I was ignorant about cognitive science and rejected
 cognitive science for AGI on the "airplane and birds" analogy. However,
 when I learned more about it, I was discovering how our perception, such
 as synesthesia-like neurons, spatial cells, and episodic intelligence,
 are almost absolutely required for AGI. I would never have been inspired
 about this if I was ignorant about cognitive science.

I taught cog sci for a couple years ... I think the brain's functions
are fascinating,
but I have no idea why you think this particular evolved system is the ONLY
sort of intelligence that is possible for us to create...

-- Ben G

-------------------------------------------
agi
Archives: http://www.listbox.com/member/archive/303/=now
RSS Feed: http://www.listbox.com/member/archive/rss/303/
Modify Your Subscription: http://www.listbox.com/member/?&;
Powered by Listbox: http://www.listbox.com


-------------------------------------------
agi
Archives: http://www.listbox.com/member/archive/303/=now
RSS Feed: http://www.listbox.com/member/archive/rss/303/
Modify Your Subscription: 
http://www.listbox.com/member/?member_id=8660244&id_secret=95818715-a78a9b
Powered by Listbox: http://www.listbox.com

Reply via email to