Certainly interesting to have an algorithm that is extremely lossy (near totally) for the majority of inputs but lossless for some weird edge cases.

That’s a great point and well articulated! Let’s see if we can expand on this together

What do you think AI voices were trained off of?

15 years. 15 years. She got one of your kids got you for 15 years

It is a social construct, and we live in a society… so put your damn socks on