I think Option 2 comes closest, even if you throw anthropics and many-worlds completely out the window (so let’s just hypothesize that there is a 99% probability that all life will die during any given second)
Once you are dead, you will cease to care about your utility function. It won’t matter how much fun you had in the moments before you died, nor will it matter what happens after, since everything that matters is gone. Your last words to your loved ones will not matter either, because they will be gone too. There will be nothing.
On the off chance that all life does not end, however, the future continues to matter. You’ll care what happens to you in the future, or, if dead, you’ll still care about what happens to the people who come after you.
By this logic, you also have to take the “stupid” low collapse bets (if you lose, it’s all ending soon so who cares?)
Practically speaking, if you embody this logic, you will not emphasize short term pleasures in the moments before you die. (Stuff like the “last meal” on death row, looking at the sky one last time, drugs, etc). You will only care about long term stuff, like what happens to your loved ones after you die and what message you leave them with.
Actually, I think that’s a pretty accurate description of the stuff I imagine myself caring about pre-death. I don’t feel that the part of my utility function which only cares about the present moment suddenly gets more weight prior to impending death. So yeah, I’ll definitely be doing a modified version of option 2, in which you carry on as normal.
If I know with 100% chance of all life will die, however...well, them I’m actually stumped on how I’m supposed to maximize utility. U = a.short term utility + b.long term utility = a.short term utility + b.null, unable to complete operation… but this is not as pathological as you might imagine, since one never reaches 100% certainty.
I think Option 2 comes closest, even if you throw anthropics and many-worlds completely out the window (so let’s just hypothesize that there is a 99% probability that all life will die during any given second)
Once you are dead, you will cease to care about your utility function. It won’t matter how much fun you had in the moments before you died, nor will it matter what happens after, since everything that matters is gone. Your last words to your loved ones will not matter either, because they will be gone too. There will be nothing.
On the off chance that all life does not end, however, the future continues to matter. You’ll care what happens to you in the future, or, if dead, you’ll still care about what happens to the people who come after you.
By this logic, you also have to take the “stupid” low collapse bets (if you lose, it’s all ending soon so who cares?)
Practically speaking, if you embody this logic, you will not emphasize short term pleasures in the moments before you die. (Stuff like the “last meal” on death row, looking at the sky one last time, drugs, etc). You will only care about long term stuff, like what happens to your loved ones after you die and what message you leave them with.
Actually, I think that’s a pretty accurate description of the stuff I imagine myself caring about pre-death. I don’t feel that the part of my utility function which only cares about the present moment suddenly gets more weight prior to impending death. So yeah, I’ll definitely be doing a modified version of option 2, in which you carry on as normal.
If I know with 100% chance of all life will die, however...well, them I’m actually stumped on how I’m supposed to maximize utility. U = a.short term utility + b.long term utility = a.short term utility + b.null, unable to complete operation… but this is not as pathological as you might imagine, since one never reaches 100% certainty.