Re: reasoning under computational limitations

Nick Bostrom (
Tue, 30 Mar 1999 14:13:09 +0000

Lee Daniel Crocker wrote:

> In the DA,
> the universe of events is "all possible physical universes"--a
> concept so ill-defined that one could write pages just on what
> that means.

Actually the situation is no worse than for any other method you have of trying to predict any future event. Specify two hypotheses--say extinction after 200 billion humans vs. extinction after 200 trillion humans--and you can apply to the DA to see how the probability shifts after taking into account the fact that you live at a time when less than 200 billion humans have been born.

> The second assumption, of course, is self-sampling. That too has
> more problems in the DA than here, because in the DA we must make
> some assumption about just what we are a sample of, whereas in the
> problem above that's a given.

That is true, but those assumptions affect only the scope of the DA, not its validity.

Nick Bostrom Department of Philosophy, Logic and Scientific Method London School of Economics