Re: Signs of Intelligence (was: Axioms and Meaning)

From: Eliezer S. Yudkowsky (sentience@pobox.com)
Date: Mon May 28 2001 - 22:35:07 MDT


Emlyn wrote:
>
> Eliezer wrote:
> >
> > Spike Jones wrote:
> > >
> > > Now suppose one wishes to beam some kind of message
> > > analogous to a list of primes, except we want to demonstrate
> > > that the sender is *really* intelligent, not just a simple list of
> > > primes. What do we send? spike
> >
> > Send them the most impressive computer program humanity has been able to
> > write so far. Programming is the purest expression of intelligence.
>
> Well, I'd say that mathematics is probably a purer expression of
> intelligence.

Hehheh... I *knew* somebody was going to bring that up. Personally, I
disagree. The space of programs is wider and freer than the space of
mathematics; it is more purely creative. You can't say whatever you like
in a proof of Fermat's Last Theorem, but you can code just about anything.

> > Send them a generally intelligent AI, if you have one; this demonstrates
> > self-awareness and self-understanding as well. Progress in the cognitive
> > sciences is the purest expression of technological enlightenment. (But
> > don't send a *seed* AI, because then they don't know how much of the AI
> > *you* wrote and how much the *AI* wrote.)
>
> Careful... sending a general purpose intelligence might be seen as a direct
> threat.

So's any computer program.

-- -- -- -- --
Eliezer S. Yudkowsky http://singinst.org/
Research Fellow, Singularity Institute for Artificial Intelligence



This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 08:07:49 MST