ChatGPT’s custom instructions allow you to add preferences or requirements that you’d like ChatGPT to consider when generating its responses. It’s super useful to include this information in the context frame so that you don’t have to keep telling the model what that context is.
“ChatGPT will consider your custom instructions for every conversation going forward. The model will consider the instructions every time it responds, so you won’t have to repeat your preferences or information in every conversation.”
For example, ChatGPT knows that I’m the Associate Professor for Digital Innovation in a School of Health and Social Care, so it includes that context in it’s responses to me.
Another great use-case for custom instructions is to tell ChatGPT what your weaknesses are, and ask it to remember that in your interactions with it. It will then look for those weaknesses in your thinking.
I can imagine a scenario where the memory for custom instructions increases to not only include static information that I specify, but where it can incorporate dynamic information from other services. For example, an integration between ChatGPT and my email client, calendar, and personal diary, is going to give me very specific responses to my prompts. This starts getting us very close to a personal AI assistant.