I’ve read it, and it doesn’t cite Bostrom; as one would expect, this means it’s pretty useless and a retread of Bostrom’s paper. The main contribution of the paper, for me, is that it includes one or two useful examples I hadn’t covered, and it includes some simple math models showing how U-shaped curves can fall out of optimizing for multiple properties.
EDIT: I emailed a link to Bostrom to the main author, who replied:
I think the arguments are actually quite similar, but from a slightly different perspective. We’re both arguing that enhancement is possible, but that an understanding of the evolutionary and cognitive constraints is needed. We further add a bit on the kinds of domains where such trade-offs are most likely.
Copy jailbreaked: http://www2.warwick.ac.uk/fac/sci/psych/people/academic/thills/thills/hillspublications/hillshertwig2011cdps.pdf
I’ve read it, and it doesn’t cite Bostrom; as one would expect, this means it’s pretty useless and a retread of Bostrom’s paper. The main contribution of the paper, for me, is that it includes one or two useful examples I hadn’t covered, and it includes some simple math models showing how U-shaped curves can fall out of optimizing for multiple properties.
EDIT: I emailed a link to Bostrom to the main author, who replied: