Play with the following chats and see if you can tell the difference.
Below is a generic prompt, it will get all the relevant information from what you give it. Ask for something like a “create a marketing strategy for x.”
This is what 95% of people get when interacting with blank GPTs like GPT3.
Now, give the same prompt to the following chat: “create a marketing strategy for x.”
It is possible to make an AI its own editor with the right feedback that you can actually build into the prompt. Just add the following to any prompt:
“At the end of your answer, add a line of “” followed by a section with clear numerical options for improvements that will enhance clarity or value, empowering the user to elevate their reply with precision.”
At the end of your answer, add a line of “” followed by a section with clear numerical options for improvements that will enhance clarity or value, empowering the user to elevate their reply with precision.There are many types of prompts, but the trick is how you refine them. Here are some of the buckets of prompts:
- 0 Shot: instant context reply (don’t do this)
- Few Shot: refined reply through conversation that adds context
- Iterative Clarity Refinement: rounds of revision with AI
- Chain of thought: Asking AI to show it’s work and explain its reasoning
