This strikes me as an extremely bad idea. Putting aside all the issues about whether it is even possible to select genetic traits that will predictably result i

This strikes me as an extremely bad idea. Putting aside all the issues about whether it is even possible to select genetic traits that will predictably result in higher 'intelligence', how do you even define this outcome? You can only optimise what you can measure, and it seems that this process will just produce people that are good at intelligence tests. At what cost to all other aspects and dimensions of human experience and achievement? How do you prove that this is a good idea for a particular individual? My opinion: Bostrom has some interesting perspectives, but has always tended to preach pseudoscience to an audience that is entirely non-critical.