[daghamm]

Sorry for a possibly stupid question, but I assume a lot of prompts will be duplicate of questions other people have asked.

Can one construct an AI cache system to minimise the questions sent to the model? Can one use a "cheaper" model to find similar prompts?

[PaulScotti]

Yeah nice idea -- does sound plausible and would make things much cheaper and faster.