In this post, I go into hopefully better output parsing.