A 4-step framework for better ChatGPT prompts: assign a role, give context, issue a command, and specify the output format. After you get a result you like, ask the model to reverse-engineer the prompt that would have produced it faster.
I'm going to take you from ChatGPT beginner to expert using a 4-step prompting process that I created. Number 1 is the role. When you're chatting with AI, you got to tell it how to think. If it's a marketing tweet, you got to say, as a marketer, because then it uses that information to go look around its neural network for information that matches to that. Number 2 is the context. You have to tell it what you're trying to do. If you're writing a better blog, tell it, I'm trying to write a better blog, or I'm trying to create 10 awesome novel tweets for my client. Number 3, give it the command for what you want. So if you You want 10 tweets, tell it, I need you to write me 10 tweets. And then the fourth is the format. The format is how you want to consume it. So you might want it in a spreadsheet, you might want it in a bulleted list. That'll give you the output. Once you get that output dialed in and it's awesome, ask it to write the command or the prompt that would've got you to that place faster.
