Have you been using Github copilot and ChatGPT wrong!

As an engineer who was surprised by AI, LLM and GPTs

Who found himself in front of multiple tools provide AI assistant for writing code

I jumped on the moving train to catch up with every one who have been advertising how great those tools are.

While the different tools try to provide very intuitive experience. it is just a text box to add my input, question or simply copy and past code right into it

I found out the hard way that this is not the correct way of using tools like ChatGPT or GitHub Copilot

There are more into how we write our questions to the answers we get than what I thought

Here is a list of tips to consider in your next message

1. The list of tabs you have open in your editor affect the response you get from GitHub copilot, it act as context and input for your questions
2. Do not just write instructions in comments wait for the copilot to auto generate the implementation, consider chatting with copilot or use CMD + I to open the prompts input to give more clear instructions and steps
3. Delete old questions from your chat history with GitHub Copilot if they do not add value or you mistyped something because that history still act as context to your latest question
4. learn the difference between pasting an error log into ChatGPT vs starting your message with investigate this vs starting with solve this because the investigating prompt will lead to the LLM model suggesting multiple solutions vs asking it to solve could lead to you getting just one option
5. Understand and be sure that till now all LLMs make mistakes when coding and you are the one who is writing the code and they here to assist you.

Did you learn any tips or trikes the hard way too?

It would be great if you share it in a comment to help others