Re: "true" AI Hardware Development


Curt Welch wrote:
jalegris@xxxxxxxxxxxx wrote:
jj wrote:

Are you interested in developing true AI or are you just content with
simulating it on a PC??

Most branches of science are quite happy to achieve successful computer
simulations, but none of them hopes to get the *real* thing, except AI.

Yeah, except AI isn't a science. It's an engineering project. The study
of human intelligence is science, but that's not AI (and the reason that
many scientists feel like the field of AI is lost - because they tend to
not do enough science to make the scientists happy).

Physicist don't expect simulations of particle interactions to produce
actual gamma rays; meteorologists don't keep umbrellas in their desk
drawers and biologists are not required to do simulations in
containment facilities. Why then should AI expect to achieve anything
other than a simulation of intelligence?

For the same reason that car manufactures expect to get real cars coming
out of their plant and not simulations of cars.

A machine implementaion can
only be truly equivalent to another machine. True AI is truly
artificial.

Only if you choose for some odd reason to define "intelligence" as only
being a human brain.

A real car is any machine that acts like a car, and true intelligence is
any machine that acts truly intelligent. It's artificial only in the fact
that it is created by man.

But cars are artifacts to begin with, so coming up with a functional
equivalent is straightforward. Natural phenomena resist functional
interpretations. Try making a car that is functionally equivalent to a
horse.

But intelligence is assumed to be different - it is the only natural
phenomenon that is assumed to be equivalent to its own simulation. Why
should this be?

I think it's because AI buffs implicitly accept the Computational
Theory of Mind - a Turing machine is the only thing that actually is
functionally equivalent to its own simulation. But brains don't just
process information - they also process matter. A "truly intelligent"
simulation is going to need some real meat.

--
Joe Legris

.

Relevant Pages

  • Re: The perfect racing simulation oriented game?
    ... inputs for a particular car and track combination, much more so that would be ... limited real life racing experience. ... of yaw response (in most games, the rate of yaw response is faster than real life, ... All proper for a simulation IMO. ...
    (rec.autos.simulators)
  • Re: iRacing.com and Skip Barber Announce Strategic Alliance
    ... That's me in the R/T cockpit in the wet while at Lime Rock. ... race car experience and infact very rarely drive on every day roads - I ... the company, I'd have wanted realistic tracks in their simulation, I'm ... find yourself in one of a series of 1-2" deep puddles at 120+MPH.. ...
    (rec.autos.simulators)
  • Re: Isnt it a WEE bit soon...?
    ... substitute for track time and not at all on a par of several ... I doubt it can accurately reproduce the G forces on the driver or ... You have zero appreciation of what it is like to drive an F1 car. ... That's stimulation NOT simulation. ...
    (rec.autos.sport.f1)
  • 3D Visuals using VR Toolbox - Help!
    ... I am trying to create a 3D visualisation to complement a car ... simulation based on both existing and real-time input data. ... The problem I am having is controlling the simulation using the data, ... The profiler shows that the vrdrawnow command is where ...
    (comp.soft-sys.matlab)
  • Re: GTR: You must be kidding me
    ... plenty of GPL laps with "rapid inputs". ... Also, in a real race car, countersteering ...
    (rec.autos.simulators)