A few weeks ago I asked ChatGPT if it could transcribe an M4A audio file, the kind of file that Apple's voice memo app produces, and it said no, it couldn't do that. I gave it another try today and got a different answer. I uploaded the file, and it came close, it said it needed the file in WAV format. It gave me instructions for writing a desktop tool that would do the conversion. I expect in another couple of weeks they'll be able to deal with the M4A file directly.