Hal Finney wrote:
>Robin also suggests that a simulation may be more likely to end sooner
>than actual human history, hence you should care less about tomorrow.
>However, if we are not in a simulation, actual human history may end soon.
>If we are in a simulation (assuming humans or their descendants run
>the sim) then this tells us that human history extends long enough for
>simulations to be common, causing us to actually be more optimistic that
>we could live a long time.
For simulations to be common it is not necessary that the ration of human
species that become posthuman is large. Each civilization that makes it
through may run a vast number of ancestor-simulations. And they may
terminate them when they are about to become posthuman, since the
computational costs of running a simulation of a posthuman civilization is
much larger than that of running a simulation of a human civilization.
BTW, note that if the simulators are nice, they might let us live in a
relatively low-complexity "heaven" after our simulation is finished. Maybe
they'll even reward us on the basis of how ethically good we have been.
There are many humans who would do that if they got to run
ancestor-simulations, so if posthumans inherit human values, this is what
we should expect.
Nick Bostrom
Department of Philosophy
Yale University
Homepage: http://www.nickbostrom.com
This archive was generated by hypermail 2b30 : Mon May 28 2001 - 09:59:40 MDT