THE PROMPT I CANNOT READ

Summary from the TLDR AI newsletter

An LLM's output is shaped by an all-encompassing prompt, which includes system instructions, conversation history, and context files, forming the model's entire environment.