From: Ben Goertzel (ben@goertzel.org)
Date: Mon Jul 31 2006 - 19:08:34 MDT
>
> The amount of computer power does not matter, sure. What matters, though, is the
> amount of resources needed to create AGI software. Google (or IBM or Microsoft) has
> enough resources to enter the AGI race late and still win easily.
Yeah -- *if* they can solve the (hard) problem of keeping their
innovative edge... and if they decide to focus some of their resources
on AGI instead of search and such...
In recent decades, not that many big software innovations have come
out of big company research labs. Startups and universities have been
the innovators, and then big companies have bought their
innovations...
Google started with some great innovation but they were a startup
then. Their recent business success has been based on clever product
introductions not deep technical innovation. It remains to be seen if
they can escape the Big Company Curse (to which M$ seems to currently
be succumbing)... even if they get interested in AGI in a more
fundamental way than they are now, this "curse" (which is really a
complex of psychosocial factors) may prevent them from hosting the
fundamental breakthroughs...
Digging a little deeper...
As many on this list know, I believe I have a viable design for an AGI
(the Novamente AGI system) --- and this design tells me several
things, such as:
a) there are probably a lot of other viable AGI designs too
b) however, creating a viable AGI design is quite hard
c) creating a viable AGI design requires a lot of work that is
different from work on narrow AI problems (such as NLP)
d) implementing and testing and teaching an AGI system is a lot of
work separate from narrow AI work (such as NLP)
Whether Google is going to put $$ into AGI research, and provide the
right environment for AGI innovation, remains to be seen. If I am
right about points b-d, AGI is not just going to emerge from their
work in information retrieval and natural language processing.
-- Ben G
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:56 MDT