Thanks ADS. I’m pretty close to agreeing with all those bullet points actually?
I wonder if, to really get to the crux, we need to outline what are the specific steps, actions, programs, investments, etc. that EA/XR and PS would disagree on. “Develop safe AI” seems totally consistent with PS, as does “be cautious of specific types of development”, although both of those formulations are vague/general.
Re Bostrom:
a single percentage point of reduction of existential risks would be worth (from a utilitarian expected utility point-of-view) a delay of over 10 million years.
By the same logic, would a 0.001% reduction in XR be worth a delay of 10,000 years? Because that seems like the kind of Pascal’s Mugging I was talking about.
(Also for what it’s worth, I think I’m more sympathetic to the “person-affecting utilitarian” view that Bostrom outlines in the last section of that paper—which may be why I learn more towards speed on the speed/safety tradeoff, and why my view might change if we already had immortality. I wonder if this is the crux?)
Side note: Bostrom does not hold or argue for 100% weight on total utilitarianism such as to take overwhelming losses on other views for tiny gains on total utilitarian stances. In Superintelligence he specifically rejects an example extreme tradeoff of that magnitude (not reserving one galaxy’s worth of resources out of millions for humanity/existing beings even if posthumans would derive more wellbeing from a given unit of resources).
I also wouldn’t actually accept a 10 million year delay in tech progress (and the death of all existing beings who would otherwise have enjoyed extended lives from advanced tech, etc) for a 0.001% reduction in existential risk.
In the abstract, yes, I would trade 10,000 years for 0.001% reduction in XR.
In practice, I think the problem with this kind of Pascal Mugging argument is that it’s really hard to know what a 0.001% reduction looks like, and really easy to do some fuzzy Fermi estimate math. If someone were to say “please give me one billion dollars, I have this really good idea to prevent XR by pursuing Strategy X”, they could probably convince me that they have at least a 0.001% chance of succeeding. So my objections to really small probabilities are mostly practical.
Thanks ADS. I’m pretty close to agreeing with all those bullet points actually?
I wonder if, to really get to the crux, we need to outline what are the specific steps, actions, programs, investments, etc. that EA/XR and PS would disagree on. “Develop safe AI” seems totally consistent with PS, as does “be cautious of specific types of development”, although both of those formulations are vague/general.
Re Bostrom:
By the same logic, would a 0.001% reduction in XR be worth a delay of 10,000 years? Because that seems like the kind of Pascal’s Mugging I was talking about.
(Also for what it’s worth, I think I’m more sympathetic to the “person-affecting utilitarian” view that Bostrom outlines in the last section of that paper—which may be why I learn more towards speed on the speed/safety tradeoff, and why my view might change if we already had immortality. I wonder if this is the crux?)
Side note: Bostrom does not hold or argue for 100% weight on total utilitarianism such as to take overwhelming losses on other views for tiny gains on total utilitarian stances. In Superintelligence he specifically rejects an example extreme tradeoff of that magnitude (not reserving one galaxy’s worth of resources out of millions for humanity/existing beings even if posthumans would derive more wellbeing from a given unit of resources).
I also wouldn’t actually accept a 10 million year delay in tech progress (and the death of all existing beings who would otherwise have enjoyed extended lives from advanced tech, etc) for a 0.001% reduction in existential risk.
Good to hear!
In the abstract, yes, I would trade 10,000 years for 0.001% reduction in XR.
In practice, I think the problem with this kind of Pascal Mugging argument is that it’s really hard to know what a 0.001% reduction looks like, and really easy to do some fuzzy Fermi estimate math. If someone were to say “please give me one billion dollars, I have this really good idea to prevent XR by pursuing Strategy X”, they could probably convince me that they have at least a 0.001% chance of succeeding. So my objections to really small probabilities are mostly practical.