On 4/18/07, Russell Wallace <[EMAIL PROTECTED]> wrote:
On 4/17/07, YKY (Yan King Yin) <[EMAIL PROTECTED]> wrote:
> I think the goal of the project is to be able to answer queries about
the stories etc.  Explicit reconstruction of the videos is not essential, as
we can be sure that the AGI actually understands things by asking it
questions that require spatio-temporal reasoning.

Suppose we postulate that for the sake of argument - why would we want to
avoid it? The biggest pitfall of the entire enterprise for the last yea
decades has been flinching away from what needs to be done because we fear
it will be too hard. Given that the wherewithal to generate the video will
be needed anyway, why not turn our bow into the gale and say, let's actually
do the video generation and make sure. Worst case some cool demos we
wouldn't otherwise have had; best case, we save ourselves a flinch and a
terminal fall.

From my perspective, video generation is not required for AGI.  It could be
an impressive demo, but that's all.  Am I missing your point?  What kind of
application can you think of for video generation?  How can it "save us from
a terminal fall"?

OTOH answering commonsense queries may be much more useful, especially when
the KB reaches a certain level of mass / sophistication.  In a commercial
project the top priority is to produce something useful, rather than
impressive demos.

YKY

-----
This list is sponsored by AGIRI: http://www.agiri.org/email
To unsubscribe or change your options, please go to:
http://v2.listbox.com/member/?member_id=231415&user_secret=fabd7936

Reply via email to