Subject: "I give up, I think" 28 January 2013
> You know, I'm starting to suspect I should just "assume" (choose actions conditional on the hypothesis that) that our species is "already" dead, and we're "mostly" just here because Friendly AI is humanly impossible and we're living in an unFriendly AI's ancestor simulation and/or some form of the anthropic doomsday argument goes through. This, because the only other alternatives I can think of right now are (A) arbitrarily rejecting some part of the "superintelligence is plausible and human values are arbitrary" thesis even though there seem to be extremely strong arguments for it, or (B) embracing a style of thought that caused me an unsustainable amount of emotional distress the other day: specifically, I lost most of a night's sleep being mildly terrified of "near-miss attempted Friendly AIs" that pay attention to humans but aren't actually nice, wondering under what conditions it would be appropriate to commit suicide in advance of being captured by one. Of course, the mere fact that I can't contemplate a hypothesis while remaining emotionally stable shouldn't make it less likely to be true out there in the real world, but in this kind of circumstance, one really must consider the outside view, which insists: "When a human with a history of mental illness invents a seemingly plausible argument in favor of suicide, it is far more likely that they've made a disastrous mistake somewhere, then that committing suicide is actually the right thing to do."
Subject: "I give up, I think" 28 January 2013
> You know, I'm starting to suspect I should just "assume" (choose actions conditional on the hypothesis that) that our species is "already" dead, and we're "mostly" just here because Friendly AI is humanly impossible and we're living in an unFriendly AI's ancestor simulation and/or some form of the anthropic doomsday argument goes through. This, because the only other alternatives I can think of right now are (A) arbitrarily rejecting some part of the "superintelligence is plausible and human values are arbitrary" thesis even though there seem to be extremely strong arguments for it, or (B) embracing a style of thought that caused me an unsustainable amount of emotional distress the other day: specifically, I lost most of a night's sleep being mildly terrified of "near-miss attempted Friendly AIs" that pay attention to humans but aren't actually nice, wondering under what conditions it would be appropriate to commit suicide in advance of being captured by one. Of course, the mere fact that I can't contemplate a hypothesis while remaining emotionally stable shouldn't make it less likely to be true out there in the real world, but in this kind of circumstance, one really must consider the outside view, which insists: "When a human with a history of mental illness invents a seemingly plausible argument in favor of suicide, it is far more likely that they've made a disastrous mistake somewhere, then that committing suicide is actually the right thing to do."