The most useful feature of ChatGPT is little publicised, it is absent from popular YouTube tutorials or any official training manuals—indeed, there is no official manual for ChatGPT! This feature, subtle yet powerful, significantly enhances user interaction with ChatGPT, effectively giving it “wings.”
Officially, ChatGPT doesn’t have any memory of its interactions with a user, each session is a one off. However, at the conclusion of a task, one does have a choice as to where to start the next session––whether to add it to an existing thread or to start completely anew. For related tasks there are often huge advantages in continuing with a previous thread. For example, you may wish Chat GPT to adopt a particular voice, such as when drafting an academic paper. That voice––the style and vocabulary of the user––can be gradually built up over a number sessions, provided each task is part of the same thread.
To facilitate such an approach ChatGPT indexes each thread, making it retrievable in future sessions. This contrasts sharply with single shot focus of numerous YouTube videos that promote the techniques of “Prompt Engineering” or offer extensive lists of essential prompts. Although such resources can be helpful, they do not represent the optimal way to engage with ChatGPT.
A more effective approach is to treat ChatGPT as a conversational partner—a friend or tutor. This interaction starts with a question, which leads to an answer, and this in turn sparks further questions or observations. Through this dynamic, Chat GPT gradually increases its knowledge of what you require and can offer richer and more appropriate outputs.
ChatGPT functions on two levels: tasks and threads. A task might be a straightforward query, such as requesting a brief summary of the capital of Pakistan. Alternatively, it could be more complex, like calculating the potential CO2 savings if 10% of Britain’s freight traffic shifted from road to rail. However, the full potential of ChatGPT is realized not through isolated tasks but through connected threads of inquiry.
For example, one might begin a thread by asking about contemporary psychotherapy techniques. Following up with more specific questions—such as which techniques could be of assistance to someone struggling with challenges in organising work. Then, prompt Chat GPT to ask more detailed questions and finally to act as a therapist helping you to deal with your procrastination. This way ChatGPT builds a broader context. Each question and answer adds a layer of data, enhancing ChatGPT’s ability to generate relevant and nuanced responses.
This method of engaging in a sustained, evolving dialogue with ChatGPT allows for a richer interaction, moving beyond single prompts to achieve deeper insights and more practical solutions.
Really Useful AI 
Leave a comment