Tips to Make ChatGPT Work Harder for You
As we begin to experiment with this new AI tech for the masses, we begin to see a potential that can help us. Responses cover a wide range of topics from poems, code, philosophy and more. To get what you want you’ll need to improve your skills by refining your prompts (telling GPT to focus harder on what you want). Just a few more words or extra line of instructions leads to improved results.
Get your answers in tabular form if you ask. This works well for lists of things. It can produce them for use in spreadsheet programs. Refine your response with more prompts (requests/instructions) in natural language.
Ask for a response in the style of your favorite author.
Set limits such as the number of words or paragraphs or say “condense” or “summarize in 100 words”.
Tell who your audience is so it will provide a response for luddites, children or scientists.
Let it produce output to be used in other AI engines. Maybe even ask for prompts from ChatGPT itself. Be as detailed and specific as you can.
Ask for the results to be in ASCII Art. The response won’t be art but symbols and characters will represent the response.
Copy and paste from other sources and ask for it to be refined. Usually there is a word limit (around 4,000). If you work is larger, submit in sections. You could even get the results in a different language or different style.
Give it examples to work with or tell it what things you enjoy and ask for suggestions.
Act out a role play with examples: a frustrated salesman, an excited teenager and so on.
Get answers that are more than the sum of the parts by giving it a list of ingredients to see what a new recipe might be. Provide a list of characters who will appear in a murder mystery at a small seaside town in the 1940s.
See and hear both sides of a debate by asking for a middle ground between the extremes with pros and cons so one can see an argument from multiple perspectives.
Download your own copy of ChatGPT at https://chat.openai.com/auth/login
From an article in Wired magazine, March 26, 2023 by David Nield