Raw LLM Responses

I think this objection will be entirely forgotten when you get a director or a small team and a director turning out quality films without any actors in them. They can already do this today of course - that's basically animation. The problem is people get hung up on the "stealing likeness's part" - hell the current round of idiot Hollywood execs did that with the writer's strike, and it's an example of really not understanding the actual paradigm change: the reason you hire voice actors is for consistency, and acting talent. Consistency is that they'll deliver clear lines, keep the volume level, and avoid blowing on the mic and other things. Acting talent (and voice actors are always trained actors) is so you can get direct them to get the right emotional delivery for a script - which you do because you want to get this done quickly, consistently and to a budget. But the thing *we* the audience remember tends to be the voice itself, and how it sounds: we substitute to some extent the quality of the performance for a familiar voice. The thing is, AI lets us separate all those components out: the big focus of a lot of generative work was "style transfer" where you copy the higher-order stylistic elements of one work to another. So what's the biggest problem with VA? It's that there's a real limit on how different any one VA can sound. AI can fix that though: one VA will be able to be the entire cast of a production, providing the emotional content but not the sound. That's good news for the one guy or girl who gets cast: that's a lot of VA hours in a block, which is a great guaranteed income. It's good news for directors too. In fact that's great for everyone except whoever *didn't* get hired because the volume of work just vanished. But there's no real reason it should stop there: because how many types of emotional delivery are there? Probably a lot, but any one person may have quirks or a limited range or a habit. But AI can fix that too: because a big librar