I usually read quite a bit about the area of artificial intelligence, machine learning, NLP, image processing and others. These days, talking to some friends, I had a crazy thought of joining several state-of-the-art algorithms that already exist, to create an idea of the entertainment in the future. Today through recurrent neural networks (LSTM) some crazy people decide to generate books or scripts of movies, as seen in these posts:

Movie written by algorithm turns out to be hilarious and intense

Harry Potter: Written by Artificial Intelligence

A while ago, I had also seen a guy who generated a comic book and sent for a contest as if he had written, and the HQ was published and they neither suspected that it was written by a machine. Unfortunately, I did not find the link again to share here. :(

Here we have another algorithm that can generate images from texts:

Image Synthesis From Text With Deep Learning - Two Minute Papers

Or, generate videos from images:

DeepStereo: Learning to Predict New Views from the World’s Imagery

And yet, if all this is generated with the low quality it is possible to improve it:

Enhance! Super Resolution From Google - Two Minute Papers

Now imagine all this interconnected inside a television, acting as an on-demand entertainment unit. Where you sit to watch television, you choose the style of movie you want to watch, the television then generates the movie in that style, and also through a camera observe your reactions to adjust the film while you watch, If you appear to be bored a comic scene appears or something like that. Your remote control instead of switching channels buttons has an “I’m not liking” button. What do you think? Too crazy?