Don't follow LLM research and development closely, but today it's first time Im hearing OAI talk aboit prompt caching.

Imo, one of those techniques, that has a step change in efficiency.
#AI