For business or other functional writing, I think Google translations are more than adequate most of the time, especially between fairly similar languages (like English and Swedish).
The problem with this text is that it is very emotional, and I think a lot of the impact gets lost in translation.
I agree with both of you, for technical language a technical approach to machine translation works.
It is an open question whether this is a fundamental problem with the approach, or simply a lack of relevant data for such translations. I suspect the former, but that's just a feeling. Perhaps with a sizeable corpus of literary translations, the system wouldn't do so badly.