From: Russell Wallace (russell.wallace@gmail.com)
Date: Fri Aug 25 2006 - 15:57:33 MDT
On 8/25/06, Richard Loosemore <rpwl@lightlink.com> wrote:
>
> I can't say I agree with your position.
I can. I think Charles's suggestion looks like the right way to do it.
I just do not think that lots of little, random projects are going to do
> it.
Bear in mind that he didn't _just_ say support lots of little random
projects, but also some larger, carefully selected ones. How much would it
take to fully fund Novamente, for example? Not much. While I don't think it
has enough depth to be a potential AGI, I do think it's heading in at least
the right compass direction, and maybe Novamente 3 or 4 somewhere down the
line might become an AGI.
I see the need, for example, of sophisticated software tools that
> would transform the field: and if I am right about that, then it will
> never happen because nobody is going to be disciplined enough to do
> that. Too big; too structured.
Existing tools, qua tools, strike me as sufficiently adequate to not be a
major limiting factor (or perhaps it's more that there isn't much I think
tools could do here in the first place, compared to what you think they
could do)... I vaguely think I asked you once before just what you'd want
said tools to do and you said something like a full reply would take more
time than you had to spare, which is fair enough - correct me if I'm wrong
there; but if you do have any particular ideas on that and time to write
them up, I'd be interested in them.
Lone hackers we have plenty of. People with a smattering of knowledge
> and a passion for coding, we have plenty of.
Well we also have a few people with a lot of knowledge, intelligence and
bright ideas - and sod all money.
Government does not only fund big outfits with lots of constraints. They
> also give money to little ones with few constraints. But they have to
> perceive a direct *need* for the product. AGI has no purpose except to
> take over the world, so they wouldn't really , uh, be, uh...
> interested.... 8-;
This is one of the reasons I think we should move on from the fantasy of AGI
conquering the world :P Seriously, AGI has a great many purposes - look at
all the things computers are used for. Now imagine _smart_ computers. That's
why the American government _did_ fund AGI, back in the 1980s. The reason
they're not doing it anymore is because they no longer regard it as feasible
on a reasonable timescale. Changing that opinion will require proof, proof
will require running code, code will require a lot of work, work will
require money to live on in the meantime. Solutions to this problem, I alas
do not have up my sleeve.
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:57 MDT