This is a proof of concept for a barebones version of a Speech Assistant ("Ok Google", "Alexa" etc.) powered by GPT-3
Currently it works in a very simple way: It listens the microphone for 5 seconds after being run and then uses that snippet to query OpenAI's API, returning the result as speech
I covered the potential applications and risks of such an application in a blogpost that can be found here
For the application to work you will need to get and use your own API key from OpenAI's Beta
Once you have it, substitute <API_KEY> with your own key and then run the program like this
pip install -r requirements.txt
python okgpt.py