On the one hand, I honestly agree that good procedures for decision making under more uncertainty than is usually considered in Bayesianism and LW is valuable, and this is an okay attempt at it.
On the other hand, I disagree with this post somewhat primarily because I think that the case where people have differing values is a case where unfortunately, it’s probable that you can’t do anything about it, and in particular I believe that some restriction of the value space is necessary in order to do anything at all. In essence, some values will probably have to lose out, so utility aggregation is a reasonably okay choice.
On the one hand, I honestly agree that good procedures for decision making under more uncertainty than is usually considered in Bayesianism and LW is valuable, and this is an okay attempt at it.
On the other hand, I disagree with this post somewhat primarily because I think that the case where people have differing values is a case where unfortunately, it’s probable that you can’t do anything about it, and in particular I believe that some restriction of the value space is necessary in order to do anything at all. In essence, some values will probably have to lose out, so utility aggregation is a reasonably okay choice.