yeah but that doesn’t mean anything, does it? I don’t think they just tokenize the raw audio, that wouldn’t make sense, right?
Comment on Perhaps the only appropriate use of AI
sobchak@programming.dev 3 days agoAccording to their tech/marketing papers, it’s supposedly multi-modal, encoding audio to tokens.
anise@quokk.au 3 days ago
sobchak@programming.dev 3 days ago
I mean, you could. Just encode 100ms chunks or whatever into tokens then push them through the same model. I’m pretty sure that’s what the claim to do (though with MoE/routing now, maybe).
prole@lemmy.blahaj.zone 3 days ago
Jesus, what a complety fucking useless waste of resources