I think a core problem that the author identifies is that it's very easy to find markers of programming ability, and those are the things we've optimized for, w

I think a core problem that the author identifies is that it's very easy to find markers of programming ability, and those are the things we've optimized for, while its much more difficult and time consuming to determine if some one has good judgement. You can be an amazing engineer, but being an ok engineer with good judgment ("do we REALLY need to do this?") generates more value.