Re: ESSAY: How to deter a rogue AI by using your first-mover advantage

From: Jeff L Jones (jeff@spoonless.net)
Date: Wed Aug 22 2007 - 13:46:51 MDT


I generally refrain from doing that, but it's mostly because I don't
like milk spilled everywhere... shutting down the simulation is a
secondary concern at best :)

On a more serious note, I don't think the fact that the Simulation
Argument could be used to explain a more general class of things than
what we current see is an argument against it (I'm not sure I agree
that it could be used to explain *anything*). Even if there is no way
of verifying the simulation argument, and even if the answer doesn't
affect the way we behave, it still might be true so it's worth
considering as a possibility.

Jeff

On 8/22/07, Tom McCabe <rocketjet314@yahoo.com> wrote:
> "It all might be a simulation" is a zero-knowledge
> hypothesis, since you can use it to explain any
> possible result. Do we refrain from eating cheesecake
> while balancing a glass of chocolate milk on our
> heads, because that might cause a shutdown of the
> simulation? No? Then why should the UFAI refrain from
> killing humans?
>
> - Tom
>
> --- Norman Noman <overturnedchair@gmail.com> wrote:
>
> > On 8/22/07, Tom McCabe <rocketjet314@yahoo.com>
> > wrote:
> > >
> > > Such an AI would easily be able to determine
> > whether
> > > it was RAI or RAI*, by simply checking the date.
> > Or
> > > checking whether an FAI had been created yet. Or
> > > checking the project history, etc.
> >
> >
> > All of these things could be perfectly faked inside
> > a simulation run by an
> > FAI
> >
>
>
>
>
> ____________________________________________________________________________________
> Take the Internet to Go: Yahoo!Go puts the Internet in your pocket: mail, news, photos & more.
> http://mobile.yahoo.com/go?refer=1GNXIC
>



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:58 MDT