I’m experimenting with AI-based live transcription (eww, AI, slop, I get it) but it only works real time with English at the moment.
the real problem there is not AI slop, but if you are basically uploading their voice to a company. if you are doing it locally, I would say fine. LLMs are better (not flawless) with translation than facts.
the real problem there is not AI slop, but if you are basically uploading their voice to a company. if you are doing it locally, I would say fine. LLMs are better (not flawless) with translation than facts.
It runs locally on my laptop, no upload anywhere, no recording either. I would be fired in an instant otherwise.
EDIT: It’s a forked version of this project, in case anyone is interested: https://github.com/pasrom/meeting-transcriber