From: Tyler Emerson (emerson@intelligence.org)
Date: Wed Feb 15 2006 - 04:26:05 MST
Olie:
You should *ask* before you forward a private message to a public list. My
message was not intended for SL4. The email to you was written in haste - it
was blunt and will be misconstrued without context. My disagreements or
agreements with AGI researchers are expressed privately, not publicly. My
position with the Institute clarifies implicitly my views on how to approach
AGI, but there is need and room for respectful dialogue with others. For the
record: I believe the AGI researchers on this list are clearly aware of the
Friendliness Challenge, regardless of any disagreement on how to handle it.
> -----Original Message-----
> From: owner-sl4@sl4.org [mailto:owner-sl4@sl4.org] On Behalf Of Olie L
> Sent: Wednesday, February 15, 2006 2:44 AM
> To: sl4@sl4.org
> Subject: FW: RE: Singularity Institute: Likely to win the race to build
> GAI?
>
>
>
>
> >From: "Tyler Emerson" <emerson@intelligence.org>
> >To: "'pdugan'" <pdugan@vt.edu>, "'Olie L'" <neomorphy@hotmail.com>
> >Subject: RE: Singularity Institute: Likely to win the race to build GAI?
> >Date: Tue, 14 Feb 2006 19:57:36 -0800
> >
> >Our chief goal is to create Friendly AI through our own project.
> Acquiring
> >the funding and researchers to sustain an eight- to ten-person team is
> >challenging but sufficiently achievable. I am not against influencing
> other
> >projects, but that is much less optimal, based on my present assessment.
> I
> >don't see enough appreciation from other projects on how *hard* it will
> be
> >to achieve Friendly AI, and how critical it is to have a mathematical
> >understanding of Friendly AI *before* building AGI. I don't know why this
> >is
> >so hard for most projects to understand. Based on my present
> understanding,
> >AGI projects are playing with fire the likes of which the world has never
> >seen, and I haven't seen a sufficient appreciation of this. The Institute
> >must find and develop brilliant researchers one individual at a time.
> We're
> >presently looking for our second full-time Research Fellow. If and when
> we
> >find that person, we'll be in a stronger position to find the 3rd and
> 4th.
> >
> >TE
> >
> > > -----Original Message-----
> > > From: pdugan [mailto:pdugan@vt.edu]
> > > Sent: Tuesday, February 14, 2006 7:31 PM
> > > To: emerson; Olie L
> > > Cc: pdugan@vt.edu
> > > Subject: RE: Singularity Institute: Likely to win the race to build
> GAI?
> > >
> > > I think you have a good idea of what SIAI's role could be, though I
> > > suppose
> > > Tyler should corroborate (I'm only affilied as a volunteer and not
> much
> >of
> > > an
> > > authority). I remember Goertzel saying something about Eliezer's
> >writings,
> > > how
> > > they made him take the Friendliness problem more seriously. I think
> the
> > > Institute could operate in the mediatory way you describe without
> > > requiring
> > > the fifty million dollar budget needed for build an AGI themselves. I
> > > think it
> > > would be most benificial for SIAI to gear themselves toward fostering
> an
> > > underlying forum for ensuring Friendliness in the AGI community.
> > >
> > > Now, though this should definetly be a role the institute plays, I
> can't
> > > say
> > > whether it would be a primary or secondary role, Eliezer seems very
> > > commited
> > > toward engineering an AGI himself.
> > >
> > > Patrick
> > >
> > > >===== Original Message From Olie L <neomorphy@hotmail.com> =====
> > > >Hi Patrick, Tyler
> > > >
> > > >I'd like to bounce this off you, first - Could you check and verify
> >that
> > > I'm
> > > >not just reiterating lame info or stepping over any lines of
> > > >appropriateness? Thankye...
> > > >
> > > >As I see it, the SIAI's stated goals of
> > > >
> > > > the "advancement of beneficial artificial intelligence and ethical
> > > >cognitive enhancement"
> > > >
> > > >can be perfectly well achieved by having other institutions "win the
> > > race"
> > > >to AGI.
> > > >
> > > >Their role is not necessarily to be the most advanced group on the
> path
> > > >toward an AGI implementation. Their role -as I see it - is to work
> > > towards
> > > >the creation of beneficial AI. That is different from creating
> > > beneficial
> > > >AI themselves.
> > > >
> > > >Many other NGOs have created powerful positions for themselves, where
> > > they
> > > >can work with commercial institutions to achieve their stated goals.
> A
> > > good
> > > >example is the RSPCA (Royal Soc. for Prevention of Cruelty to
> Animals)
> >-
> > > >which has become (1) a de-facto enforcer of government legislation
> (2)
> >An
> > > >powerful lobby for creating government legislation (3) an operation
> >that
> > > >directly provides shelter to some animals (4) An organisation that
> >works
> > > >collaboratively with many businesses.
> > > >
> > > >Not only do businesses provide the RSPCA with resources and money,
> they
> > > also
> > > >engage in joint projects and give them unusual access to commercially
> > > >sensitive information.
> > > >
> > > >As a non-profit organisation, the Institute may often be given far
> more
> > > >access to proprietary information than other groups - such as
> > > universities
> > > >or even investors. Such access relies on the Institute developing an
> > > >appropriate reputation, including having the right skills to be able
> to
> > > >provide consultative services.
> > > >
> > > >No commercial AGI project would want to create an unfriendly AI. It
> is
> > > >against their interests to do so.
> > > >
> > > >If a business believes that the Institute can provide a service -
> such
> >as
> > > >improving the friendliness of the business's AI project - there is a
> > > strong
> > > >incentive to work with the Institute, advancing the Institute's
> goals.
> > > >
> > > >As far as I can see, by carving out a niche for itself - FAI theory -
> >the
> > > >Institute has already done much to advance its reputation. Even if
> its
> > > own
> > > >projects are not the furthest towards demonstrating AGI potential,
> any
> > > >efforts will hopefully assist in improving the Institute's FAI
> >expertise.
> > > >If there are demonstrable successes, these will also greatly advance
> >the
> > > >Institute's reputation.
> > > >
> > > >I would like to see the Institute expand its capacity to provide
> > > >consultative services. This is only my opinion. But it has already
> >had
> > > >substantial influence on a number of projects other than those of its
> > > staff.
> > > > Let us hope that more AGI projects will take their advice.
> > > >
> > > >-- Olie
> > > >
> > > >
> > > >
> > > >>From: pdugan <pdugan@vt.edu>
> > > >>Reply-To: sl4@sl4.org
> > > >>To: sl4 <sl4@sl4.org>
> > > >>CC: pdugan@vt.edu
> > > >>Subject: RE: Singularity Institute: Likely to win the race to build
> >GAI?
> > > >>Date: Tue, 14 Feb 2006 18:25:04 -0500
> > > >>
> > > >>Well I'd say its worth evaluating the prospective Friendliness of
> >these
> > > >>systems, for the obvious reasons. This is probably fairly difficult
> to
> > > do,
> > > >>particularly for projects based on proprietary information. I think
> a
> > > >>useful
> > > >>hueristic when gauging the risks associated with an AGI is to
> evaluate
> > > the
> > > >>likelyhood of a hard take-off. From what I gather about Novaemente,
> >you
> > > >>seem
> > > >>to see soft take-off as much more likely. If Novamente does prove
> >robust
> > > >>enough to be deemed a "general intelligence" would it possible for
> > > someone
> > > >>else, possibly SIAI, to conceive of a more "powerful" system that
> > > enganges
> > > >>in
> > > >>hard take-off while Novamente spends its "childhood"? Or one the
> other
> > > >>hand,
> > > >>what sort of Friendliness constraints does Novamente possess?
> > > >>
> > > >> Patrick
> > > >>
> > > >> >===== Original Message From ben@goertzel.org =====
> > > >> >In fact I know of a number of individuals/groups in addition to
> >myself
> > > >> >who fall into this category (significant progress made toward
> > > >> >realizing a software implementation whose design has apparent AGI
> > > >> >potential), though I'm not sure which of them are list members.
> > > >> >
> > > >> >In addition to my Novamente project (www.novamente.net), I would
> > > >> >mention Steve Omohundro
> > > >> >
> > > >> >http://home.att.net/~om3/selfawaresystems.html
> > > >> >
> > > >> >(who is working on a self-modifying AI system using his own
> variant
> >of
> > > >> >Bayesian learning) and James Rogers with his
> > > >> >algorithmic-information-theory related AGI design (James is a list
> > > >> >member, but his work has been kept sufficiently proprietary that I
> > > >> >can't say much about it). There are many others as well...
> > > >> >
> > > >> >Based on crude considerations, it would seem SIAI is nowhere near
> >the
> > > >> >most advanced group on the path toward an AGI implementation. On
> >the
> > > >> >other hand, it's of course possible that those of us who are
> >"further
> > > >> >along" all have wrong ideas (though I doubt it!) and SIAI will
> come
> >up
> > > >> >with the right idea in 2008 or whenever and then proceed rapidly
> > > >> >toward the end goal.
> > > >> >
> > > >> >ben
> > > >> >ben
> > > >> >
> > > >> >On 2/14/06, pdugan <pdugan@vt.edu> wrote:
> > > >> >> There is a certain list member who already has an AGI model more
> > > than
> > > >>half
> > > >> >> implemented, making it a few years from testablility to see if
> it
> > > >>classifies
> > > >> >> as a genuine AGI, and if so then maybe another half a decade
> >before
> > > >>something
> > > >> >> like recursive self-improvement becomes possible.
> > > >> >>
> > > >> >> Patrick
> > > >> >>
> > > >> >> >===== Original Message From P K <kpete1@hotmail.com> =====
> > > >> >> >>Yes, I know that they are working on _Friendly_ GAI. But my
> > > question
> > > >>is:
> > > >> >> >>What reason is there to think that the Institute has any real
> > > chance
> > > >>of
> > > >> >> >>winning the race to General Artificial Intelligence of any
> sort,
> > > >>beating
> > > >> >> >>out those thousands of very smart GAI researchers?
> > > >> >> >>
> > > >> >> >There is no particular reason(s) I can think of that make the
> > > >>Institute
> > > >>more
> > > >> >> >likely to develop AGI than any other organization with skilled
> > > >>developers.
> > > >> >> >It's all a fog. The only way to see if their ideas have any
> merit
> > > is
> > > >>to
> > > >>try
> > > >> >> >them out. Also, I suspect their donations would increase if
> they
> > > >>showed
> > > >>some
> > > >> >> >proofs of concept. It's all speculative at this point.
> > > >> >> >
> > > >> >> >As for predicting success or failure, the best calibrated
> answer
> >is
> > > to
> > > >> >> >predict failure to anyone attempting to build a GAI. You would
> be
> > > >>right
> > > >>most
> > > >> >> >of the time and wrong probably only once or right all the time
> (o
> > > >>dear,
> > > >> >> >heresy).
> > > >> >> >
> > > >> >> >That doesn't mean it isn't worth trying. By analogy, think of
> AGI
> > > >>developers
> > > >> >> >as individual sperm trying to reach the egg. The odds of any
> > > >>individual
> > > >>are
> > > >> >> >incredibly small but the reward is so good it would be a shame
> >not
> > > to
> > > >>try.
> > > >> >> >Also, FAI has to be developed only once for all to benefit.
> > > >> >> >
> > > >> >>
> >_________________________________________________________________
> > > >> >> >MSN(r) Calendar keeps you organized and takes the effort out of
> > > >>scheduling
> > > >> >> >get-togethers.
> > > >> >>
> > > >>
> > > >http://join.msn.com/?pgmarket=en-
> > > ca&page=byoa/prem&xAPID=1994&DI=1034&SU=http
> > > >> >> ://hotmail.com/enca&HL=Market_MSNIS_Taglines
> > > >> >> > Start enjoying all the benefits of MSN(r) Premium right now
> and
> > > get
> > > >>the
> > > >> >> >first two months FREE*.
> > > >> >>
> > > >> >>
> > > >> >>
> > > >>
> > > >>
> >
> >
> >
This archive was generated by hypermail 2.1.5 : Tue Feb 21 2006 - 04:23:29 MST