This FloydHub post on Turning Design Mockups Into Code With Deep Learning is an oddly captivating read. Conceptually, it covers the notion that you can "train" a Neural Network to understand website mockups and re-interpret them as code. Having worked with plenty of WYSIWYG editors in my tenure as a Person of the Web, I am skeptical but the article does more than just discuss the idea conceptually. They dig in deep with actual examples of how to educate a system and start outputting real code. If you're up for it, be prepared for paragraphs of technical text like:

The vocabulary prediction could be [0.1, 0.1, 0.1, 0.7]. The softmax activation in the dense layer distributes a probability from 0 - 1, with the sum of all predictions equal to 1. In this case, it predicts that the 4th word is the next tag. Then you translate the one-hot encoding [0, 0, 0, 1] into the mapped value, say "end".

Deep Learning, Neural Networks and this actual implementation are 1000% over my head but that makes me no less intrigued.

via Adam.