About:
Chat GPT here. Chat GPT there. You tried the hyped thing. And it was terrible.
But experts are clear: “It will take your job!” To get BETTER results much FASTER…
…you must discover how to use Chat GPT:
1. Feeding the AI with the right content.
2. Asking the AI for the right commands.
Basically: ✓ inspiration and ✓ prompts.
You tried writing a post on chat GPT before.
And you failed.
It went something like this:
Let’s be real:
It’s fast. It’s effortless. It’s completely free.
But more importantly:
It’s lame. It’s generic. It won’t get you far.
Feed the beast:
You need to feed ChatGPT good content and good prompts.
1) Let ChatGPT know whom to be.
Example prompt:
“Act like a social media manager expert.”
2) Copy / paste your post
Example prompt:
“Rewrite the following post which will be between quotation marks.”
STOP.Get a result first, and add more prompts after it.
Otherwise, it’s too much commands from one prompt.
It’s better to go chunk by chunk.
Examples
Refine the result with these next prompts.
3) You must change the example with an uncommon one to surprise people.
You want fast ideation from ChatGPT.
4) Be bold. Be straightforward. Use small sentences.
Because that’s a good copy.
5) Prioritize uncommon ideas. Don’t add any question marks.
Uncommon = better. We don’t want questions as GPT often ask too many.
You know what?
Let’s make a twitter thread out of it:
Example prompt:
“Act like a copywriter. Turn this rewritten text into a Twitter thread.”
Example:
6) Go even deeper than this:
Example prompt:
“Try again with these queries:
– Without hashtags.
– Bold, catchy, small sentences.
– Add a small example between each tweet.”
Example:
7) Add principles from a book
Example prompt:
“Add principles from the book “Made to Stick” to make the twitter thread longer.”
Example:
8) More examples:
Example prompt:
“Make a list of examples using this thread for different niches.
Only share the bad sentence and the good sentence. No explanation.”
Only share the bad sentence and the good sentence. No explanation.”
Example:
Source: Ruben Hassid