You can do amazing things with artificial intelligence — if you know how to write (and rewrite) great AI prompts.
Large language models (LLMs) such as Notion AI, ChatGPT, and Llama use huge amounts of language data, equivalent to millions of books and web pages. They break down each word or part of a word into tokens and rank them based on how often they appear together in the dataset.
When you give an instruction to an AI model, it uses these rankings to analyze your request and provide a response that it thinks is best. This process is relatively simple for straightforward instructions.
Speak normally
Generative AI models are different from Siri or Google Assistant. They can effectively respond to more than just exact phrases. These models have been trained on a large amount of conversational dialogue, so they understand the nuances of how people talk and text each other. When interacting with the model, try to speak to it like you would speak to a human, and you will receive a more natural response.
Be concise
Make your prompt as simple as you can while still explaining your request in all relevant detail (more on that later). The clearer your language, the less likely it is that the model will misinterpret your words (more on that later, too).
Don’t use negative phrases like "Don't use negative phrases"
When you say, “Do not...”, an LLM might focus on “Do” while ignoring the “not,” and thus take the exact action you think you’ve instructed it to avoid. So:
BAD: Do not include incomplete lists
GOOD: Only include complete lists
Now that we've discussed how to communicate with AI, let's move on to the topic of discussion. I have chosen a research project that a typical market analyst might seek assistance with, but feel free to ask AI about anything, whether it's related to schoolwork or how to create a fantastic menu for a New Year's Eve dinner party. The same principles apply.