This post is contributing to some of the most important questions around AI risk
Sets a good example of having a high-level conversation that takes into account various real objections
Has been written in a clear and engaging way
The following seem to me to the most promising avenues for improvement / future work:
It is quite long, and there are probably some ways to increase the quality density of the post
I definitely have a sense that some aspects of this post and the overall conversation could be improved by trying to formalize things more (though I recognize that formalizing things is hard, and trying to do so might increase the effort required to write this post by a factor of 2 or so)
Curated for the following reasons:
This post is contributing to some of the most important questions around AI risk
Sets a good example of having a high-level conversation that takes into account various real objections
Has been written in a clear and engaging way
The following seem to me to the most promising avenues for improvement / future work:
It is quite long, and there are probably some ways to increase the quality density of the post
I definitely have a sense that some aspects of this post and the overall conversation could be improved by trying to formalize things more (though I recognize that formalizing things is hard, and trying to do so might increase the effort required to write this post by a factor of 2 or so)