Nicolas Cole breaks down two types of context for AI prompting: information context (background data like client call details) and instruction context (explicit guidance on how the model should act). Combining both deliberately is essential to get the most from AI tools.
Chapters
Using AI, there are two different types of context. There is information context and there is instruction context. Information context is things that you want AI to know. It's information. We're going to do something together, but when we do it, I want you to know this information. Information could be, I'm creating content for a client as a ghostwriter, and I want you to have the context, the information of our last client call. This is different than instruction context. Instruction context is, this is how I want you to do something. It's not just information, it is instruction. Think of information as a big folder and think of instruction as prompts, skills, mini instruction manuals. You have to understand this distinction in order to get more and more out of some of these tools, because if you don't understand how these two pieces work together, you won't understand how to maximize your leverage.