Argon Gruber's Shortform 2024-09-08T21:06:23.526Z

Comment by Argon Gruber (argon-gruber) on Argon Gruber's Shortform · 2024-09-08T22:25:10.835Z · LW · GW

No..? I called it "Pascal's Demon"? The hypothetical being is called "Pascal's Demon".

Comment by Argon Gruber (argon-gruber) on Argon Gruber's Shortform · 2024-09-08T21:06:23.648Z · LW · GW

Do counterfactual mugging scenarios violate the 2nd Kolmgorov Axiom, or Normalization?

I'm thinking about, eg, this situation:
Imagine a superintelligence (like a version of Pascal’s demon) approaches you and tells you it has flipped a coin. If the coin landed heads, the superintelligence would never interact with you. If it landed tails, it will ask you to give $100. If you agree to give $100 in the tails scenario, the superintelligence will retroactively reward you with $10,000 if the coin landed heads.

There are 4 possible outcomes, right? {Heads+Would Pay, Heads+Would Not Pay, Tails+Would Pay, Tails+Would Not Pay} To obey normalization, P(H+WP, H+WNP, T+WP, T+WNT) = 1, right?

Comment by Argon Gruber (argon-gruber) on Am I confused about the "malign universal prior" argument? · 2024-08-29T19:55:16.800Z · LW · GW

What is a "universal prior"?