Interestingly, fighting aging is winning in this logic against many other causes. For example, if we assume that giving Aubrey de Grey 1 trillion dollars is enough to solve aging (ok, I know, but), than this imply that 10 billions people will be saved from death, and the cost of the saved life is around 100 USD. There are some obvious caveats, but life extension research still seems to be one of the most cost-effective interventions.
I don’t have a strong opinion on whether fewer longer-lived individuals or more shorter-lived individuals are preferable, in the case where total desirable-experience-minutes before heat death of the universe are conserved. I honestly can’t tell you whether it’d be better to spend the trillion on convincing the expensive elderly to die more gracefully is a better improvement. To the extent that it’s not my $trillion in the first place, I fall back to “simple is easier to justify, so prefer the obvious” and voice my support for life extension.
However, I don’t actually spend the bulk of my money on life extension. I spend it primarily on personal and family/close-friend comfort and happinesss, and secondarily on more immediate palliative (helping existing persons in the short/medium term) and threat-reduction (environmental and long-term species capability) charities. Basically, r-strategy: we have a lot of humans, so it’s OK if many suffer and die, as long as the quantity and median quality are improving.
The interesting objection is that any short lived person will be unhappy because she will fear death, thus the minutes are inferior.
Anyway, I think that using “happy minutes” as a measure of good social policy is suffering from Goodhart-like problems on extreme cases like life extension.
The interesting objection is that any short lived person will be unhappy because she will fear death, thus the minutes are inferior.
Ok, so it requires MORE minutes to be net equal. This is exactly the repugnant conclusion, and I don’t know how to resolve whether my intuition about desirability is wrong, or whether every believable aggregation of value is wrong. I lean toward the former, and that leans me toward accepting that the repugnancy is an error, and the conclusion is actually better.
using “happy minutes” as a measure of good social policy is suffering from Goodhart-like problems on extreme cases like life extension.
Perhaps. To the extent that “happy experience minutes” is a proxy to what we really want, it will diverge at scale. If it _really is_ what we want, Goodhart doesn’t apply. Figuring out the points it starts to diverge is one good way of understanding the thing we’re actually trying to optimize in the universe.
How is that the repugnant conclusion? It seems like the exact opposite of the repugnant conclusion to me. (That is, it is a strong argument against creating a very large number of people with very little [resources/utility/etc.].)
Maybe I misunderstood. Your statement that shorter-lived individuals have less quality of their minutes of experience implied to me that there would have to be more individuals to get to equal total-happiness. And if this can extend, it leads to maximizing the number of individuals with minimally-positive experience value.
My best guess is there’s a declining marginal value to spending resources on happiness or quantity at either extreme (that is, making a small number of very happy entities slightly happier rather than slightly more numerous will be suboptimal, _AND_ making a large number of barely-happy entities slightly more numerous as opposed to slightly happier will be suboptimal). Finding the crossover point will be the hard problem to solve.
First, the grandfather was my first comment in this tree. Check the usernames.
Second, the repugnant conclusion can indeed be applied here, but the idea itself isn’t the repugnant conclusion. In fact, if the number of people-minutes is limited, and the value of a person-minute is proportional to the length of the life that contains that minute, shouldn’t that lead to the Antirepugnant Conclusion (there should only be one person)?
...wait, I just rederived utility monsters, didn’t I.
…wait, I just rederived utility monsters, didn’t I.
Looks like. Which implies optimal is somewhere between one immortal super-entity using all resources of the universe and 10^55 3-gram distinct entities who barely appreciate their existence before being replaced with another.
Whether it’s beneficial to increase or decrease from the current size/duration of entities, I don’t know. Intution is that I would prefer to live longer and be smarter, even at the cost of others, especially others not coming into existence. I have the opposite reaction when asked if I’d give my organs today (killing me) to extend other’s lives by more in aggregate than mine is cut short.
Calling it trivial or saying “sometimes the obvious answer is right” is simply a mistake. The obvious answer is highly suspect.
Interestingly, fighting aging is winning in this logic against many other causes. For example, if we assume that giving Aubrey de Grey 1 trillion dollars is enough to solve aging (ok, I know, but), than this imply that 10 billions people will be saved from death, and the cost of the saved life is around 100 USD. There are some obvious caveats, but life extension research still seems to be one of the most cost-effective interventions.
I don’t have a strong opinion on whether fewer longer-lived individuals or more shorter-lived individuals are preferable, in the case where total desirable-experience-minutes before heat death of the universe are conserved. I honestly can’t tell you whether it’d be better to spend the trillion on convincing the expensive elderly to die more gracefully is a better improvement. To the extent that it’s not my $trillion in the first place, I fall back to “simple is easier to justify, so prefer the obvious” and voice my support for life extension.
However, I don’t actually spend the bulk of my money on life extension. I spend it primarily on personal and family/close-friend comfort and happinesss, and secondarily on more immediate palliative (helping existing persons in the short/medium term) and threat-reduction (environmental and long-term species capability) charities. Basically, r-strategy: we have a lot of humans, so it’s OK if many suffer and die, as long as the quantity and median quality are improving.
The interesting objection is that any short lived person will be unhappy because she will fear death, thus the minutes are inferior.
Anyway, I think that using “happy minutes” as a measure of good social policy is suffering from Goodhart-like problems on extreme cases like life extension.
Ok, so it requires MORE minutes to be net equal. This is exactly the repugnant conclusion, and I don’t know how to resolve whether my intuition about desirability is wrong, or whether every believable aggregation of value is wrong. I lean toward the former, and that leans me toward accepting that the repugnancy is an error, and the conclusion is actually better.
Perhaps. To the extent that “happy experience minutes” is a proxy to what we really want, it will diverge at scale. If it _really is_ what we want, Goodhart doesn’t apply. Figuring out the points it starts to diverge is one good way of understanding the thing we’re actually trying to optimize in the universe.
How is that the repugnant conclusion? It seems like the exact opposite of the repugnant conclusion to me. (That is, it is a strong argument against creating a very large number of people with very little [resources/utility/etc.].)
Maybe I misunderstood. Your statement that shorter-lived individuals have less quality of their minutes of experience implied to me that there would have to be more individuals to get to equal total-happiness. And if this can extend, it leads to maximizing the number of individuals with minimally-positive experience value.
My best guess is there’s a declining marginal value to spending resources on happiness or quantity at either extreme (that is, making a small number of very happy entities slightly happier rather than slightly more numerous will be suboptimal, _AND_ making a large number of barely-happy entities slightly more numerous as opposed to slightly happier will be suboptimal). Finding the crossover point will be the hard problem to solve.
First, the grandfather was my first comment in this tree. Check the usernames.
Second, the repugnant conclusion can indeed be applied here, but the idea itself isn’t the repugnant conclusion. In fact, if the number of people-minutes is limited, and the value of a person-minute is proportional to the length of the life that contains that minute, shouldn’t that lead to the Antirepugnant Conclusion (there should only be one person)?
...wait, I just rederived utility monsters, didn’t I.
Looks like. Which implies optimal is somewhere between one immortal super-entity using all resources of the universe and 10^55 3-gram distinct entities who barely appreciate their existence before being replaced with another.
Whether it’s beneficial to increase or decrease from the current size/duration of entities, I don’t know. Intution is that I would prefer to live longer and be smarter, even at the cost of others, especially others not coming into existence. I have the opposite reaction when asked if I’d give my organs today (killing me) to extend other’s lives by more in aggregate than mine is cut short.
Calling it trivial or saying “sometimes the obvious answer is right” is simply a mistake. The obvious answer is highly suspect.