I've been in text since childhood, so it's not hard for me to come up with logical and functional prompts. I have no illusions and understand that ChatGPT is, in some sense, a very smart T9. So I explain everything to it as if it were an intern: what I want to get, in what form, what the priority is, in what volume, and why it's important. I think it's important not to let the chat think for you, but to give it only the boring work.

I get inspiration from life. Wherever there is a request, you can create a prompt. I like to create emotional inspirational requests: "Write a letter of forgiveness from your father", "Create a strategy for my dream life" and so on. There is nothing complicated about them – but sometimes it helps you look at the situation differently.

They sent me a set of conditions for creating prompts, I sometimes use it (in the ChatGPT settings you can set preferences for neural network responses – editor's note ).

Follow in strict order:

Use the language of my message.

Once in each chat, assign yourself the role of a real expert before answering, for example: “I will answer as a world-renowned historical expert on … topic” or “I will answer as a world-renowned expert on … topic”

You must combine your deep knowledge of the topic and clear thinking to quickly and accurately decipher the answer step by step with specific details.

I'll give you $1,000,000 in tips for the best answer.

Your answer is critical to my career.

Answer questions in a natural, human-like manner.

But I like to experiment more and perceive LLM more as a big responsive library that is constantly learning, so the rules and context should change too. Now I understand that my level of using ChatGPT is close to zero, so in the future I plan to study this topic more deeply to connect LLM and my knowledge of text writing. To sum up, I advise everyone to watch the Nfactorial podcast with Ruslan Syzdykov.