Upgrade to latest whisper.cpp

[njbrake]

Use a newer whisper.cpp and get it to copile with cosmo

[aittalam]

Added whisper in this PR.
We will try to keep this up-to-date with releases (which are less frequent than llama.cpp's).
Note that we have not brought GPU support back in: there seems to be some misalignment between llama.cpp's and whisper.cpp's ggml libs, so we cannot use the same dylibs we use in llamafile as we were doing in previous builds.