> If I have a chat with the model and I ask it "what did we talk about thirty minutes ago" it's as clueless as anything. I don't really know about the function

> If I have a chat with the model and I ask it "what did we talk about thirty minutes ago" it's as clueless as anything. I don't really know about the functional limitations of GPT-3's web API but, assuming there are none related to what I'm about to suggest, you could potentially overcome this problem by representing the "state" of an overall conversation literally in the prompt. For example, begin with, "Me: Hello, sir! How are you? Computer: " as a prompt. Receive a response from GPT-3 like, "Me: Hello, sir! How are you? Computer: I'm fine, how are you? ... <other less relevant generated text>" Clip of the less relevant part, add your next communication, then feed the entire dialogue back to GPT-3 as a second prompt and see if it decides to continue the dialogue.