Article View: sci.econ
Article #2096Re: Technological Singularity
From: Patrick Juola
Date: Mon, 08 Jun 1998 00:00
Date: Mon, 08 Jun 1998 00:00
43 lines
3015 bytes
3015 bytes
In article <357BD9B4.72E0@remove.muenchen.this.org.junk> Bernd Paysan <bernd.paysan@remove.muenchen.this.org.junk> writes: >James Sharman wrote: >> For >> example, their is a reasnable chance that the first truely inteligent >> A.I.'s will be created for the military and then the technology will be >> applied to the dometic market. > >Do you really think so? One of the best scenes of "Forrest Gump" was >when the conversion with the drill sergeant: > >"What are you supposed to do here, Gump?" >"Obey your order, sir!" >"Gump, you are a genius!". > >Creating something that mindlessly obeys orders isn't what I would call >AI. We already have these devices. The difference between a Kamikaze >pilot and a cruise missile control computer is that you don't need to >put shit in the brain of the latter before it does it's job - and that >it is less likely to miss the target. However, that's exactly what the military doesn't want in its AI applications. The whole *problem* they're trying to solve is to reduce the load on the person giving the orders, by providing initiative and problem-solving capacity in the equipment, so that the personnel can concentrate on the hard problems. For example, one of the hard problems (towards which large steps have been taken, but which has not been solved) is the identify-friend-or-foe (IFF) problem. In a nutshell, "friendly fire isn't" -- and you don't want to shoot at your friends. So it would be useful to install cutouts on your weapons so that you don't shoot friendlies by mistake. And if the cutouts worked, then you don't need to detail as many men to sentry duty; a bunch of automated machine guns could be set up to shoot at any hostiles crossing an area. This makes this whole "Halt! Who goes there? Pass, friend!" schtick a thing of the past -- or would make it if the IFF stuff worked. The military does *NOT* want mindless order-following machines; they want machines that follow orders when given and otherwise do something "intelligent." Hence the interest in AI. -kitten
Technical Details
Message-ID:
<6lgpfa$hfa$1@news.ox.ac.uk>Path:
arc1-ca.pugleaf.net!unknown!not-for-mailReferences:
<357857B7.5244@sigmatechinc.com> <897299396.23952.0.nnrp-02.c2d9a433@news.demon.co.uk> <357BD9B4.72E0@remove.muenchen.this.org.junk>Show Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: fd5f7,3de6aaf66239df89 X-Google-Attributes: gidfd5f7,public X-Google-Thread: fd6dd,3de6aaf66239df89 X-Google-Attributes: gidfd6dd,public X-Google-Thread: 1045ed,3de6aaf66239df89 X-Google-Attributes: gid1045ed,public X-Google-Thread: 102b75,3de6aaf66239df89 X-Google-Attributes: gid102b75,public X-Google-Thread: 10461e,3de6aaf66239df89 X-Google-Attributes: gid10461e,public From: patrick@gryphon.psych.ox.ac.uk (Patrick Juola) Subject: Re: Technological Singularity Date: 1998/06/08 Message-ID: <6lgpfa$hfa$1@news.ox.ac.uk> X-Deja-AN: 360671234 References: <357857B7.5244@sigmatechinc.com> <897299396.23952.0.nnrp-02.c2d9a433@news.demon.co.uk> <357BD9B4.72E0@remove.muenchen.this.org.junk> Organization: M.R.C. Research Centre in Brain and Behaviour, Oxford University Newsgroups: sci.econ,bionet.neuroscience,comp.ai,comp.arch,comp.robotics.misc