[extropy-chat] Re: Structure of AI

[Damien Broderick]

At 06:48 PM 11/23/2004 +0100, Serafino wrote:
>From: "Russell Wallace"
>
> > Is a 'detailed enough' simulation of all the molecules
> > in a glass of water the same as a simulation 'deterministic
> > enough'?
>
>Yes, I suppose.
>
> > Because, in this case, the simulation would be too strong,
> > that is to say no 'wetness' (whatever it means) would
> > be allowed.
>
>'Wetness'? You mean those 'qualia'?

This seems to me one of those classic errors due to incompetent reframing 
(in this case failure to reframe appropriately). The standard skeptical 
move is to say `If the computer simulates a tornado, nobody gets blown away 
or drenched'. Of course not, not out here. But in a sufficiently complex 
simulation, the people inhabiting the sim surely *would* suffer both 
experiences. It seems arbitrary to declare that if those inhabitants were 
simulated down the level of responsive nervous systems they would *not* 
experience `wetness'. They would observe surface tension, sense coolness or 
warmth, respond to the flow of the liquid on their hands, etc. They would 
report all this in the words we have provided for them, the same words we 
use to annotate just those experiences. In what sense would they fail to 
find `wetness' in their world?

Damien Broderick