Interesting idea and execution thus far is impressive!
I have one question regarding privacy.
>. For your privacy, we store all of the recordings locally on your Mac. Only you have access to them. Compression and Automated Speech Recognition (ASR) is all done locally. Recording data is never sent off of your machine.
Although recording data might not be sent off the machine, usage of GPT-4 does imply you're sending sensitive data to OpenAI (and thus off the machine).
Presumably you're embedding user's screen and audio data locally, which means that potentially the embedding process and definitely the response generation process both send data to OpenAI, which is a privacy concern.
Do you plan on implement on-device models in the future, as way to mitigate this?