The reason this is a good discussion to have is because, at a superficial level, it looks like ORM's should be faster than writing SQL, but when you get down in

The reason this is a good discussion to have is because, at a superficial level, it looks like ORM's should be faster than writing SQL, but when you get down into the details you find that the complexities are a lot deeper than a superficial view suggests. "Do what is best for your project" is good advice, and the author states a good case for ORM's not being "what's best for your project".