Smoothmin and personal identity

I value ex­is­ten­tial risk re­duc­tion; I also value poverty re­duc­tion. Th­ese two things trade off against each other.

I value be­ing gen­er­ous; I also value read­ing in­ter­est­ing books. Th­ese two things trade off against each other.

But the way the two trade­offs work do not seem to be the same. For the first one, I feel com­fortable hav­ing a util­ity (for no ex­is­ten­tial risks) and (for no poverty), and then weight­ing them and max­imis­ing the sum:

  • .

If this ends up with me only re­duc­ing ex­is­ten­tial risks, or only re­duc­ing poverty, then that’s fine, I’m work­ing on the op­tion with the most marginal im­pact.

For the sec­ond one, I would not want to max­imise some sum , and would cer­tainly com­plain if I ended up never read­ing again, or never be­ing gen­er­ous again. I’d pre­fer to max­imise some­thing like the smooth min­i­mum of and , some­thing like:

  • .

And I’d want the weights to be cho­sen so that I am very likely to both be gen­er­ous and read, to some ex­tent, over longer pe­ri­ods of time.

World prefer­ences vs iden­tity preferences

Some time ago, I wrote a post about “prefer­ences over non-re­wards”. I’m plan­ning to col­lect most of these prefer­ences into the cat­e­gory of “per­sonal iden­tity”: the sort of be­ing you want to be.

The “You’re not the boss of me!” prefer­ence from that post—mean­ing you change your preferred ac­tion be­cause you were told to/​told not to—is very similar to the “4 Prob­lems with self-refer­en­tial ” from this post, and will be both be grouped un­der “per­sonal iden­tity”.

It’s my hope that all hu­man prefer­ences and meta-prefer­ences can be syn­the­sised into one of “world prefer­ences” or “per­sonal iden­tity prefer­ences”. As this post sug­gests, the meth­ods of ag­gre­ga­tion may be differ­ent for the two cat­e­gories.

No comments.