"J. R. Molloy" wrote:
>
> Michael S. Lorrey has written,
>
> > I don't know if that were comparable. Now, if we had, say, a celebrity death
> > match between Eli and Eugene at the next Extro... that might be fun to
> watch...
>
> Which side would you bet on (if you were to bet on one side or the other).
>
> --J. R.
>
> PS: I'd favor Eliezer in the first rounds, but I think Eugene would pull ahead
> in the middle of the bout. Then Eliezer would flounder unless he could come up
> with a working AI. If someone injected evolvable algorithms into Eliezer's Code,
> that would cinch his final victory, if the match were allowed to go that far.
> But depending on where the match were held, local officials might stop the fight
> before useful AI could emerge. So Eugene would win in that case by default.
Nah, give each a 128 node Beowulf system (each node is a 1ghz Athlon) to run on.
Each gets x time to load and construct/evolve their AI's, then they compete in a
virtual landscape constructed and operated by a third system. One key component
is while competing head to head, each has a virtual tamagotchi they have to care
for simultaneously. Competitions would cover the whole range, from combat, to
puzzle solving, art & music to free form essay questions judged by human judges
and 'compassion contests'.
What I'm most interested in seeing is some AI module for generating original
initiative.
This archive was generated by hypermail 2b30 : Mon May 28 2001 - 09:50:14 MDT