Yeah don't get me wrong, I think large language models are amazing, and the whole personal assistants thing could work if there was a lot of back end work to build something useful rather than just Siri/Alexa + Pippa's therapist. But then focusing on the tech part, the whole "oh everyone will want to use AI-created vtuber avatars", just like, what? What's the utility? How do you connect to a need? There's so much of this in AI-related startups right now.Neuro only works because Vedal's full time job, now, is data management on the project. Yet, even still, the way Neuro works is to be the insane responder to someone else talking, which requires another human interacting.
Everyone is slapping AI on things looking for VC funds. There's a couple of places where this tech has utility, but we're already see a big rejection in automated systems in places they don't work well generally. Where a human to human interface is needed, AI simply won't work.
Hell, one of those companies is still talking about vtubing needing "fussy set-up of motion capture suits". Like it's 2018.