Doing hypothesis testing with the Whisper OpenAI API.
Hypothesis: Whisper will create high-enough quality text transcripts (with some prompting) to automate the act of converting podcast recordings into text.
What I want to learn:
- how large a model will be needed for english to text translation.