10 ideas about using so-called "neural networks" to improve one's (creative) workflow

I'm not deep into neural networks (yet). I trained a neural network on text based works (Jules Verne) the last days (used https://github.com/karpathy/char-rnn).

1. Assist in writing - trained NN suggests sentences in the style of e.g. a certain author
2. Create dialogue, names, graphics, sounds for game and other digital media
3. Use as creative writing prompt - create summary to base text on/ or rewrite generated text
4. Use NN to exploit or close bugs in games and other software
5. Train NN on own coding style - suggest fitting code
6. Use well-trained NN to assist in decision-finding
7. Use NN to suggest related art for a given keyword as inspiration - trained on the things the user "favorites"
8. Train NN on pieces of artwork with desired style -> recreate similar artwork for personal projects
9. NN trained on artwork can also be used to input random inspiration, e.g. stuck with painting, NN adds something to it
10. If need to create a big, similar set of assets by hand, can train NN to do it for you

I hope this makes sense .. a very interesting topic. I don't get it completely yet, but I think it'll increase in use and influence in near future.
I attached my first test of training a RNN on 4 books of Jules Verne for a day. It would need a bit more time and data, but it already yields pretty interesting results.