I am really psyched about the tech to create voices for generated speech. I am really excited when in the future this tech might be small enough to deliver with a game or OS. Then much more interactive games can be built with generated text. It would be so cool to license voices for this kind of work.

It will probably end up with artists creating unique and interesting voices to allow game developers to pick and choose. So voice artists will be a thing as well as graphics artists. The tricky part will finding a way to add mood states to the generated voices. Right now this could be done with different voice profiles for different speech.

Right now the tech is "large", but this will rapidly become smaller and efficient as it gets developed more.