The probability that the species will become extinct because every individual human will die of old age is negligible compared the the extinction risk of insufficiently-careful AGI research.
I’m not talking about old age, I’m talking about death. This includes death from plague, asteroid, LHC mishap, or paperclip maximizer. I didn’t say “cure death” or “cure old age” but “[solve] the problem of death”. And for the record, to my mind, the likeliest solution involves AGI, developed extremely carefully—but as quickly as possible under that condition.
Having refreshed, I see you’ve changed the course of your reply to some degree. I’d like to respond further but I don’t have time to think it through right now. I will just add that while I don’t assign intrinsic value to individuals not yet born, I do intrinsically value the human species as a present and future entity—but not as much as I value individuals currently alive. That said, I need to spend some time thinking about this before I add to my answer. I may have been too hasty and accidentally weakened the implication of “extinction” through a poor turn of phrase.
I don’t assign intrinsic value to individuals not yet born
Note that this is dynamically inconsistent: given the opportunity, this value implies that at time T, you would want to bind yourself so that at all times greater than T, you would still only intrinsically care about people who were alive at time T. (Unless you have ‘overriding’ values of not modifying yourself, or of your intrinsic valuations changing in certain ways, etc., but that sounds awfully messy and possibly unstable.)
(Also, that’s assuming causal decision theory. TDT/UDT probably gives a different result due to negotiations with similar agents binding themselves at different times, but I don’t want to work that out right now.)
The probability that the species will become extinct because every individual human will die of old age is negligible compared the the extinction risk of insufficiently-careful AGI research.
I’m not talking about old age, I’m talking about death. This includes death from plague, asteroid, LHC mishap, or paperclip maximizer.
. . .
Having refreshed, I see you’ve changed the course of your reply to some degree.
I did, when I realized my first reply was vulnerable to the response which you in fact made and which I quote above. (I should probably let my replies sit for 15 minutes before submitting/uploading them to reduce the probability of situations like this one, which can get confusing.)
(And thank you for your reply to my question about your values.)
I’m not talking about old age, I’m talking about death. This includes death from plague, asteroid, LHC mishap, or paperclip maximizer. I didn’t say “cure death” or “cure old age” but “[solve] the problem of death”. And for the record, to my mind, the likeliest solution involves AGI, developed extremely carefully—but as quickly as possible under that condition.
Having refreshed, I see you’ve changed the course of your reply to some degree. I’d like to respond further but I don’t have time to think it through right now. I will just add that while I don’t assign intrinsic value to individuals not yet born, I do intrinsically value the human species as a present and future entity—but not as much as I value individuals currently alive. That said, I need to spend some time thinking about this before I add to my answer. I may have been too hasty and accidentally weakened the implication of “extinction” through a poor turn of phrase.
Note that this is dynamically inconsistent: given the opportunity, this value implies that at time T, you would want to bind yourself so that at all times greater than T, you would still only intrinsically care about people who were alive at time T. (Unless you have ‘overriding’ values of not modifying yourself, or of your intrinsic valuations changing in certain ways, etc., but that sounds awfully messy and possibly unstable.)
(Also, that’s assuming causal decision theory. TDT/UDT probably gives a different result due to negotiations with similar agents binding themselves at different times, but I don’t want to work that out right now.)
I did, when I realized my first reply was vulnerable to the response which you in fact made and which I quote above. (I should probably let my replies sit for 15 minutes before submitting/uploading them to reduce the probability of situations like this one, which can get confusing.)
(And thank you for your reply to my question about your values.)