You could do this already with Google Translate and a TTS, and it will be just as bad.
Have you even used Eleven Labs? Their voices sound way more natural than Google Translate. I was able to release my last book with an audio version because the quality was quite good. The pacing and tone shifts aren’t always perfect, but it’s perfectly serviceable.
This is why SAG-AFTRA is right to strike. Just another way to rob people of their labor. The “savings” come at a direct cost to performers’ ability to support themselves in an already highly-competitive field. And there is no way that the “savings” will be passed on to customers.
Would be nice if more startups would try to use the LLMs for good, instead of trying to paint circumventing one of the few US unions with any power as anything but scabbing.
I wonder if they’ll get the cadence right? That’s what I can’t stand about current text to speech systems, the cadence is always wrong, it’s not how actual humans talk, and it’s infuriating.
