From: Ben Goertzel (ben@goertzel.org)
Date: Thu Jun 27 2002 - 10:01:34 MDT
Hi Stephen,
> 2. I predict that regardless of sentiment expressed on this list,
> (perhaps indicative of popular sentiment worldwide), that as evidence
> mounts that an AGI is possible, then government institutions,
> especially the US
> military research organizations will be the first to fund, and fund with
> the most money - thus sponsoring the progress of Seed AI to AGI and
> perhaps beyond. If I am right, then you will see increasingly more
> ambitious Darpa AI project announcements in years 2003, 2004 and 2005.
>
> Currently Cycorp the company is behaving as though my prediction is on
> track.
I actually agree that there will be increasingly more ambitious DARPA AI
project announcements in the future...
However, at this point, our differing intuitions about AI become relevant.
I have a feeling that DARPA will continue to pour the vast majority of its
funds into fairly "conservative" AI approaches like Cyc ;> Thus, as has
often been the case, even if the gov't spends a vast amount of money on AI,
it will almost entirely be wasted!
It's not that I think Cyc is worthless, or that Cyc could *never* be
transformed in to an AGI. But I think that the mainstream AI community is
hung up on some fairly unproductive approaches, and I include Cyc in that
category. So I think that, with DARPA money focused on conventional
approaches, we're looking at decades to AGI -- whereas I think that with
more innovative approaches (which DARPA will likely not be willing to fund
in the near future), progress can and will be made much faster.
I know that DARPA's proposals often involve a language of "innovation." But
very often in the past, I've seen this used as a kind of feel-good
smokescreen for funding small various on the same old stuff...
> If I am wrong, then Seed AI and AGI will be developed some other way and
> we can observe/discuss events as they unfold.
Indeed...
ben g
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:39 MDT