I've since looked over my tokens usage and realized that had I been saving them properly in a local vector database I would have spent less than $8 yesterday for all of my testing. At least I learn quickly! #

I have been blindly exploring OpenAI's API for both chat and embeddings to allow me to "chat" with my own content. One experiment was pulling data from a local MySQL server and creating embeddings to use with chatGPT. What I wasn't paying attention to is the cost! I spent almost $15 yesterday on embedding experiments not realizing how many tokens I was using! Really shocked me! My biggest mistake was not saving the vector database locally and loading it instead of recreating the embeddings every time. Live and learn! 😂 #