I used the whole sample, but something tells me I could have gotten better results faster if I’d cut out the lines she says outside.

I’d be interested in generating both of these lines again. While the first line sounds good, I think it would sound even better with different emphasis on certain words. And the second one is definitely not fully there, yet. Shouldn’t be an issue though. I only spent a few minutes trying with each (for reference, I normally spend 10-20 minutes to get ones that we actually use).