Something I want to experiment with is NVIDIA RTX Voice.
As I mentioned yesterday, an AI needs a well defined, precisely scoped problem and a lot of training data to be really useful. If you have all that, the result can really be impressive. I hope I can get this to work with my GTX 1650 GPU (officially not supported but there are some hacks).
Though this video is more hilarious (which is a good compensation for the very serious post yesterday) than technical, the real world application on video/voice conferencing can be really helpful.
I want to see the faces of people in a video conference when I'm using a hair dryer while talking: I imagine that they see it blowing and cannot hear it.