On Wed, 31 May 2006 11:24:00 -500, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
Do you have a projection of the optomistic and pesimistic times of when NM
will achieve Trans-Human intelligence?
The pessimistic time-frame is never... the optimistic time-frame is 2
years from now.
More meaningful intervals are as follows (pasted out of a recent presentation):
These are what I think we may be able to achieve IF the project gets
funded at a level of $500K/year, enabling us to hire a team of ~10
full-time technical staff. This has not yet occurred. Lacking this
proper level of staffing, things may take much longer and the whole
project may not complete. I believe we have a workable AGI design but
there is plenty of work left to be done to get everything implemented,
get the "details" resolved and make it fully functional..
*****
2007: autonomous intelligent infant interacting spontaneously in a 3D
simulation world, communicating in simple English
2008-2009: young-child level (Piaget's concrete operational stage),
with the capability to carry out complex series of purposeful actions
and communicate in simple grammatical sentences
Toward the end of this stage, the spinning-off of a powerful,
profitable question-answering product becomes viable
2008-2012: embodied learning of complex language
2009-2014: Piaget's formal level, the ability to reason abstractly and
communicate in fully complex English
2010-?: Strongly self-modifying AI
******
To me the time estimates are NOT the important thing, though. The
important thing is that we have a well-articulated, scientifically and
engineeringly sensible plan for getting from here to the grand
end-goal.... Estimating what is possible is one thing, estimating
when is another...
-- Ben
How long before NM gets to first grade?
12th grade? collage?
Phd Status?
Dan Goe
----------------------------------------------------
From : Ben Goertzel <[EMAIL PROTECTED]>
To : agi@v2.listbox.com
Subject : Re: [agi] Largest test to date?.. Data there vs data not
there..
Date : Wed, 31 May 2006 12:14:13 -0400
> To answer your questions:
>
> Right now the most machines we have used for a single NM installation is
4
>
> However, scaling up to many machines is NOT our biggest issue by any
means...
>
> In 2000, we ran our Webmind AI Engine system (with a conceptually
> similar distributed processing infrastructure) on a network of 30+
> quad processor PC's ...
>
> This kind of infrastructural stuff is not the obstacle to AGI, the
> obstacle is in getting the learning/memory algorithms right.... I
> think the NM design contains a workable solution to this problem but
> there remain plenty of details to resolve via experimentation...
>
> NM is not at this point designing itself, no -- it is still at the
> "infantile" stage...
>
> ben
>
>
>
> On Wed, 31 May 2006 11:09:44 -500, [EMAIL PROTECTED] <[EMAIL PROTECTED]>
wrote:
> > What is the largest test to date of Novamate on a distributed network
of
> > machines?
> >
> > Is Novamate designing itself?
> >
> > Dan Goe
> >
> > ----------------------------------------------------
> > From : Ben Goertzel <[EMAIL PROTECTED]>
> > To : agi@v2.listbox.com
> > Subject : Re: [agi] Data there vs data not there, Limits to storage?
> > Date : Wed, 31 May 2006 11:51:45 -0400
> > > Novamente can run on a distributed network of machines, using both
RAM
> > > and disk for memory; hence in principle it has no "limits of
storage"
> > > except those imposed by cost...
> > >
> > > Of course there are limitations in terms of RAM access time and
> > > distributed processing overhead, but these are the familiar ones and
> > > not specific to Novamente
> > >
> > > ben
> > >
> >
> > -------
> > To unsubscribe, change your address, or temporarily deactivate your
subscription,
> > please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
> >
>
> -------
> To unsubscribe, change your address, or temporarily deactivate your
subscription,
> please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
-------
To unsubscribe, change your address, or temporarily deactivate your
subscription,
please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
-------
To unsubscribe, change your address, or temporarily deactivate your subscription,
please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
