We will soon have a generative model for speech that is as good as GPT-3 for text generation.<p>We already have a lot of data (from youtube) and computing power.<p>All we need is a good model.
Hey, I'm interested in what you're doing. I've been trying to do the same for music generation, but I don't have enough compute resources. Is there a way to get in touch? My email hncomments5@gmail.com