From: Ben Goertzel (ben@goertzel.org)
Date: Fri Jan 20 2006 - 19:42:26 MST
Quite frankly, I doubt the AGI book release, in itself, is going to
make much of a difference. The reason is that none of the papers in
the book describe dramatically successful AGI research results.
But I think that this book, along with the various AGI-oriented
academic workshops and conference sessions that are increasingly
appearing, is going to help gradually and incrementally push the AGI
meme into the awareness of the academic and ultimately popular
community.
There is no doubt that AGI is on the rise, in the sense that its very
low profile is in a phase of gradual increase, both in terms of
popularity and in terms of being taken seriously.
If Novamente or some other AGI project progresses fast enough, then
the gradual increase in AGI's reputation will be irrelevant. But if
progress continues to be relatively slow (largely due to absence of
significant funding of those AGI projects with sound ideas underlying
them) then perhaps this reputation-building will bear fruit by making
funding more available 5-10 years down the road...
In short: Yeah, I think AGI's reputation is increasing exponentially.
But the exponent is still damn small at the moment, though positive,
and I don't see this changing anytime soon until someone announces a
dramatic breakthrough.
-- Ben
On 1/20/06, H C <lphege@hotmail.com> wrote:
> Richard I have to agree with you. AGI is kind of in a state where funding is
> incredibly hard to find, so flaunting some optimism can't exactly hurt you
> very much.
>
> Actually I'm really interested in the effect of the upcoming AGI book
> (goertzel, yudkowsky, et al).
>
> The release is probably a great time for maximizing an optimistic outlook in
> the AGI community, because, well, this book release (hopefully, if it does a
> good job) will be the starting point of a new "age" of Strong AI- called
> AGI, which is way better and worth way more funding than Strong AI (or
> whatever they used to call it).
>
> -hegem0n
>
>
> >From: Richard Loosemore <rpwl@lightlink.com>
> >Reply-To: sl4@sl4.org
> >To: sl4@sl4.org
> >Subject: Re: Why invest in AGI?
> >Date: Sat, 21 Jan 2006 07:56:52 -0500
> >
> >Ben,
> >
> >I absolutely agree with your comments.
> >
> >Investors do need to see something that is not open-ended research, but of
> >course *some* research is necessary in this arena.
> >
> >I omitted to mention that I can see some incremental returns using the path
> >that I am taking, so we are indeed thinking the same way. Without that, we
> >would be trying to sell a pure research project, which would be almost
> >impossible.
> >
> >The difference that I saw was in the emphasis on the potential speedup in
> >the pace. I think we have all become so depressed by the slow pace over
> >the last 30 years that we have come to believe that it will necessarily be
> >a long slow haul, and we want to avoid being tainted with the kind of
> >overinflated optimism that happened back in the 70s and 80s, when investors
> >were told that full intelligence was just around the corner. We don't want
> >to have that old mud thrown at us.
> >
> >My goal is to resurrect the old optimism, but for completely different
> >reasons.
> >
> >Much more that I could say about this, but I am going to save it for a few
> >weeks hence, when I will be putting a lot more ideas out into the public
> >domain.
> >
> >Richard Loosemore.
> >
> >Ben Goertzel wrote:
> >>Richard,
> >>
> >>I totally agree that AGI has a big potential to break through in the
> >>next few years and provide immense profits as well as immense
> >>scientific and humanitarian benefits. You are definitely preaching to
> >>the converted in this regard!!
> >>
> >>My point was a practical one: merely that nearly all business
> >>investors are interested in investment opportunities with less
> >>appearance of risk than AGI.
> >>
> >>Essentially, where AGI is concerned, what is needed is *research
> >>money*. Even for a project like Novamente, where we have a detailed
> >>and theoretically grounded software design that we are progressively
> >>implementing, there is *still* plenty of detail-level research to be
> >>done to make it all work. Business funding for research projects is
> >>hard to come by, unless (and even if) one can make a strong argument
> >>regarding the profitability of incremental results as well as the end
> >>goal. Not impossible of course -- just hard to come by.
> >>
> >>In the case of Novamente I do believe we can make profits via software
> >>systems embodying incremental progress toward our end goal of AGI (and
> >>we are doing so now, to a limited extent). So I think Novamente makes
> >>sense as a business opportunity even for fairly conventional investors
> >>-- but that is because we have concrete plans re how to make money
> >>from the incremental versions of our system along the path to general
> >>AI.
> >>
> >>
> >>-- Ben
> >>
> >>On 1/21/06, Richard Loosemore <rpwl@lightlink.com> wrote:
> >>
> >>>My take on your question is slightly different than Ben's because I have
> >>>different approach to the problem.
> >>>
> >>>My answer to "Why should I invest in AGI?":
> >>>
> >>>Because there is the potential for it to take off at a speed that would
> >>>supply the investor with her own personal starship within a decade or
> >>>two.
> >>>
> >>>Here is how.
> >>>
> >>>If the standard approach to AI turns out to have a Blind Spot Assumption
> >>>built into it - an assumption that is glaringly obvious when you know
> >>>it is there, but which otherwise is impossible to see - and if that
> >>>BSA turns out to a crucial blockage that has stopped progress all this
> >>>time, then we could open up the floodgates by dumping that assumption.
> >>>
> >>>At the risk of sounding like a wild optimist, I actually do believe that
> >>>we are in this situation.
> >>>
> >>>We could build a learning system that acquires concepts through
> >>>experience, and which also has the kind of motivation/emotion system
> >>>that would make it completely (and reliably) benign, we could find
> >>>ourselves surprised at how fast it could acquire knowledge and grow up
> >>>into a superintelligent AGI.
> >>>
> >>>This is the core of my approach, and I think it will work. It is so
> >>>radically different to the standard methods that, at the very least, it
> >>>might not have the same drawbacks.
> >>>
> >>>[More soon: I am trying to corral my ideas into a website.]
> >>>
> >>>Richard Loosemore.
> >>>
> >>>
> >>>
> >>>H C wrote:
> >>>
> >>>>I have really seen hardly any discussion or exposition directly related
> >>>>to the question of AGI investment, from a business standpoint (although
> >>>>I think the AGIRI forum is a step in the right direction).
> >>>>
> >>>>Say today is your lucky day, and you sat down next a millionaire
> >>>>businessman on the bus and he asked you "Why should I invest in AGI?".
> >>>>
> >>>>How would you respond? What makes a good response here? What are some
> >>>>major things you would bring up in your answer?
> >>>>
> >>>>
> >>>>
> >>>
> >>
> >>
>
>
>
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:55 MDT