**From:** Joshua Fox (*joshua@joshuafox.com*)

**Date:** Thu Dec 04 2008 - 04:05:21 MST

**Next message:**Stuart Armstrong: "Re: [sl4] AI's behaving badly"**Previous message:**BillK: "Re: [sl4] Join me on Bebo"**In reply to:**Stefan Pernar: "[sl4] Convergence of Expected Utilities with Algorithmic Probability Distributions - uh?"**Next in thread:**Peter de Blanc: "Re: [sl4] Convergence of Expected Utilities with Algorithmic Probability Distributions - uh?"**Reply:**Peter de Blanc: "Re: [sl4] Convergence of Expected Utilities with Algorithmic Probability Distributions - uh?"**Messages sorted by:**[ date ] [ thread ] [ subject ] [ author ] [ attachment ]

Commonly people feel that "every human is equally valuable" (in terms of

moral utility).

Thus, the utility of your actions scales linearly with the number of human

lives you save, or the number of humans you benefit (with constant benefit

per person).

However, the question is whether this intuition continues, so that the limit

of the utility function, as the number of humans approaches infinity, is

infinite.

In fact, De Blanc says (to simplify greatly), your utility function must be

bounded from above.

See this related piece by Eliezer, which takes another perspective on

utility for massive numbers of people (although here discussing specific

finite numbers) http://www.overcomingbias.com/2007/10/torture-vs-dust.html

This is also related:

http://www.overcomingbias.com/2007/05/scope_insensiti.html

"Infinite set atheism" is another related concept

http://www.overcomingbias.com/2008/01/trust-in-bayes.html

Various moral calculations break down as you take the limit to infinity.

Joshua

On Thu, Dec 4, 2008 at 9:03 AM, Stefan Pernar <stefan.pernar@gmail.com>wrote:

*> About two months back Peter De Blanc posted a paper called "Convergence of
*

*> Expected Utilities with Algorithmic Probability Distributions<http://arxiv.org/PS_cache/arxiv/pdf/0712/0712.4318v1.pdf>."
*

*> The Singularity Institute blogged about it but I cannot for the life of me
*

*> figure out what the significance of the paper is.
*

*>
*

*> Would anyone mind shedding some light on this apparently important paper
*

*> for me so I can put it in context?
*

*>
*

*> Much appreciated.
*

*>
*

*> Stefan
*

*>
*

*> --
*

*> Stefan Pernar
*

*> Unit 06-01 Platinum Residence
*

*> Nr. 10 Lorong 31 Geylang
*

*> 388039 Singapore
*

*> SINGAPORE
*

*> Mobile: +65 9755 7590
*

*> Skype: Stefan.Pernar
*

*> __________________________________________________
*

*> D O T E A S Y - "Join the web hosting revolution!"
*

*> http://www.doteasy.com
*

**Next message:**Stuart Armstrong: "Re: [sl4] AI's behaving badly"**Previous message:**BillK: "Re: [sl4] Join me on Bebo"**In reply to:**Stefan Pernar: "[sl4] Convergence of Expected Utilities with Algorithmic Probability Distributions - uh?"**Next in thread:**Peter de Blanc: "Re: [sl4] Convergence of Expected Utilities with Algorithmic Probability Distributions - uh?"**Reply:**Peter de Blanc: "Re: [sl4] Convergence of Expected Utilities with Algorithmic Probability Distributions - uh?"**Messages sorted by:**[ date ] [ thread ] [ subject ] [ author ] [ attachment ]

*
This archive was generated by hypermail 2.1.5
: Wed Jul 17 2013 - 04:01:03 MDT
*