I hold the same view towards “non-naive” maximization being suboptimal for some people. Further clarification in my other comment.
I have concerns about the idea that a healthy-seeming maximizer can prove the point that maximization is safe. In mental health, we often come across “ticking time bomb” scenarios that I’m using as a sort of Pascal’s mugging (except that there’s plenty of knowledge and evidence that this mugging does in fact take place, and not uncommonly). What if someone just appears to be healthy and this appearance of being healthy is simply concealing and contributing to a serious emotional breakdown later in their life, potentially decades on? This process isn’t a mysterious thing that comes without obvious signs, but what may be obvious to mental health professionals may not be obvious to EAs.
I don’t reject the possibility that healthy maximizers can exist. (Potentially there is a common ground where a rationalist may describe a plausible strategy as maximization, and I, as a mental health advocate, would say it’s not, and our disagreement in terminology is actually consistent with both our frameworks.) If EA continues to endorse maximizing, how about we at least do it in a way that doesn’t directly align with known risks of ticking time bombs?
I hold the same view towards “non-naive” maximization being suboptimal for some people. Further clarification in my other comment.
I have concerns about the idea that a healthy-seeming maximizer can prove the point that maximization is safe. In mental health, we often come across “ticking time bomb” scenarios that I’m using as a sort of Pascal’s mugging (except that there’s plenty of knowledge and evidence that this mugging does in fact take place, and not uncommonly). What if someone just appears to be healthy and this appearance of being healthy is simply concealing and contributing to a serious emotional breakdown later in their life, potentially decades on? This process isn’t a mysterious thing that comes without obvious signs, but what may be obvious to mental health professionals may not be obvious to EAs.
I don’t reject the possibility that healthy maximizers can exist. (Potentially there is a common ground where a rationalist may describe a plausible strategy as maximization, and I, as a mental health advocate, would say it’s not, and our disagreement in terminology is actually consistent with both our frameworks.) If EA continues to endorse maximizing, how about we at least do it in a way that doesn’t directly align with known risks of ticking time bombs?