Lumifer comments on Probabilities Small Enough To Ignore: An attack on Pascal’s Mugging

Lumifer 21 Sep 2015 15:13 UTC
0 points

If there are infinitely many possibilities, then the generalized theorem does require bounded utility.

I am not sure I understand. Link?
- entirelyuseless 21 Sep 2015 19:54 UTC
  0 points
  Parent
  http://arxiv.org/pdf/0907.5598.pdf
  - Lumifer 21 Sep 2015 20:09 UTC
    1 point
    Parent
    
    Our main result implies that if you have an unbounded, perception determined, computable utility function, and you use a Solomonoff-like prior (Solomonoff, 1964), then you have no way to choose between policies using expected utility.
    
    So, it’s within the AIXI context and you feed your utility function infinite (!) sequences of “perceptions”.
    
    We’re not in VNM land any more.