Ben,
If I am beating a dead horse, please feel free to ignore this, but I'm
imagining a prototype that shows glimmerings of AGI. Such a system, though
not useful or commercially viable, would sometimes act in interesting, even
creepy, ways. It might be inconsistent and buggy, and work in a
Yes, this is one of the things we are working towards with Novamente.
Unfortunately, meeting this low barrier based on a genuine AGI
architecture is a lot more work than doing so in a more bogus way
based on an architecture without growth potential...
ben
On 12/20/06, Joshua Fox [EMAIL
I'm also surprised there aren't more programmers or AGI enthusiasts who
aren't willing to work for beans to further this goal. We're just two
students in Arizona, but we'd both gladly give up our current lives to work
for 15-20G's a year and pull 80 hour weeks eating this stuff up. Having a
Well, the requirements to **design** an AGI on the high level are much
steeper than the requirements to contribute (as part of a team) to the
**implementation** (and working out of design details) of AGI.
I dare say that anyone with a good knowledge of C++, Linux, and
undergraduate computer
Ben,
The question which I would ask, were I a potential funder How soon can I
see something that, though not true AGI, makes me say 'Wow, I've never seen
anything like that before.' ?
I appreciate that this is an incredibly challenging project, and that in
some cases investors will accept a
Hi,
You mention intermediate steps to AI, but the question is whether these
are narrow-AI applications (the bane of AGI projects) or some sort of
(incomplete) AGI.
According the approach I have charted out (the only one I understand),
the true path to AGI does not really involve commercially
What kind of numbers are we talking here to fund a single AGI project
like Novamente? If I could, I'd instantly dedicate all my time and
resources to developing AI, but because most of my knowledge is auto
didactic, I don't get considered for any jobs. So for now, I'm stuck
in the drudgery of
BTW Ben, for the love of God, can you please tell me when your AGI book is
coming out? It's been in my Amazon shopping cart for 6 months now!
The publisher finally mailed me a copy of the book last week!
Ben
-
This list is sponsored by AGIRI: http://www.agiri.org/email
To unsubscribe or
Hi Joshua,
Thanks for the comments
Indeed, the creation of a thinking machine is not a typical VC type
project. I know a few VC's personally and am well aware of their way
of thinking and the way thir businesses operate. There is a lot of
technology risk in the creation of an AGI, as
Ben,
My A.I. group of friends (was: CommonSense Computing Group, and is now
more scattered) has been trying to do an open-source development for a set
of programs that are working toward human-scale intelligence. For
example, Hugo Liu's commonsense reasoning toolkit, ConceptNet, was ported
My main reason for resisting the urge to open-source Novamente is AGI
safety concerns.
At the moment Novamente is no danger to anyone, but once it gets more
advanced, I worry about irresponsible people forking the codebase
privately and creating an AGI customized for malicious purposes...
This
Regarding de Garis' graph of the number of people who've died in
different wars throughout history, are the numbers raw or divided by
the population size?
-Chuck
On 12/11/06, Ben Goertzel [EMAIL PROTECTED] wrote:
Hi,
For anyone who is curious about the talk Ten Years to the Singularity
(if we
The exponential growth pattern holds regardless of whether you
normalize by global population size or not...
-- Ben
On 12/11/06, Chuck Esterbrook [EMAIL PROTECTED] wrote:
Regarding de Garis' graph of the number of people who've died in
different wars throughout history, are the numbers raw or
Darn.
On 12/11/06, Ben Goertzel [EMAIL PROTECTED] wrote:
The exponential growth pattern holds regardless of whether you
normalize by global population size or not...
-- Ben
On 12/11/06, Chuck Esterbrook [EMAIL PROTECTED] wrote:
Regarding de Garis' graph of the number of people who've died in
14 matches
Mail list logo