>Incidentally, John Leslie makes it clear that the extinction argument only
>goes through if there are objective reasons to expect doomsday - such as
>nuclear war threats, environmental hazards, black hole engulfment,
>metastable vacuum collapse on a cosmic scale, etc. It's a subtle point -
>in fact, it escapes me at the moment, but then I haven't had my dinner yet
>- - and lets him out of the implicit charge of quietism or defeatism. We can
>*change the odds* in our favour (or against them, as we're tending to do
>this century, alas).
The resultant probabilities depend both on the priors and on the strength
of the probability shift we make because of the DA. The priors are
determined by such things as how likely a nuclear war seems etc. If the
prior probability of doom is extremely small then even a rather large
shift won't make doom probable. By taking action to reduce the likelihood
of an all-out nuclear war, or the grey goo scenario, we change the odds of
doom. But it is not clear that even very good priors would suffice to make
doom not almost certain.
Nicholas Bostrom
http://www.hedweb.com/nickb