Good stuff to know. Here's a sort of upside-down trick I use to create useful/reusable/customizable prompts ... Whenever you find a particular article, post, paper, explanation — any piece of writing, really — that you think is written particularly well on a given topic (movies to watch, explanations of particular pieces of legislation, etc.), copy and paste (or link to) it into ChatGPT / Claude / Gemini / Perplexity and preface it with this statement:
“Reverse engineer the prompt that could have been used to create this text:”
You can then re-use the resulting prompt — modified as needed — to create other similar-quality material.
Something worth considering is to add the following to your prompt suggestion. Given each LLM has some form of context bias it's going to use the reverse engineer content into a prompt for you. You could add the following the end of your prompt:
---
"Before you start, please ask me any questions you have about this so I can give you more context. Be extremely comprehensive"
---
This way you might be able to drive the results into some particular direction (bias) that fits your needs better. I have not tried this yet but it seems like it would be useful. I'm adding it to my list of GOOD TIPS. Again, thank you.
Bill, I took your prompt and asked PromptVibes to create a version for us. Seemed like the right thing to do given we're talking about the post on PromptVibes. Here is it's answer:
---
Act as a reverse engineering expert specializing in text analysis. I will provide you with a body of text. Your task is to analyze the text and deduce the most likely prompt that could have led to its creation. Focus on identifying key themes, purposes, and any specific language or terminology used that could indicate the original prompt's intent. Provide a prompt that succinctly encapsulates these elements. Ensure that your suggested prompt is broad enough to have plausibly resulted in the provided text, but specific enough to reflect the nuances and particularities of the content. If the text is ambiguous or lacks clear indicators of its origin, request additional context to better inform your analysis
I’ve recently been introduced to CitizenAI. It combs local legislation (video and audio) and transcribes text for you. I research Bitcoin at the local level and this helps me decide on who to vote for.
Tom:
Good stuff to know. Here's a sort of upside-down trick I use to create useful/reusable/customizable prompts ... Whenever you find a particular article, post, paper, explanation — any piece of writing, really — that you think is written particularly well on a given topic (movies to watch, explanations of particular pieces of legislation, etc.), copy and paste (or link to) it into ChatGPT / Claude / Gemini / Perplexity and preface it with this statement:
“Reverse engineer the prompt that could have been used to create this text:”
You can then re-use the resulting prompt — modified as needed — to create other similar-quality material.
BILL - THIS IS GREAT thank you. And it's exactly the sort of thing of thing is useful to share.
Smart!
Something worth considering is to add the following to your prompt suggestion. Given each LLM has some form of context bias it's going to use the reverse engineer content into a prompt for you. You could add the following the end of your prompt:
---
"Before you start, please ask me any questions you have about this so I can give you more context. Be extremely comprehensive"
---
This way you might be able to drive the results into some particular direction (bias) that fits your needs better. I have not tried this yet but it seems like it would be useful. I'm adding it to my list of GOOD TIPS. Again, thank you.
Bill, I took your prompt and asked PromptVibes to create a version for us. Seemed like the right thing to do given we're talking about the post on PromptVibes. Here is it's answer:
---
Act as a reverse engineering expert specializing in text analysis. I will provide you with a body of text. Your task is to analyze the text and deduce the most likely prompt that could have led to its creation. Focus on identifying key themes, purposes, and any specific language or terminology used that could indicate the original prompt's intent. Provide a prompt that succinctly encapsulates these elements. Ensure that your suggested prompt is broad enough to have plausibly resulted in the provided text, but specific enough to reflect the nuances and particularities of the content. If the text is ambiguous or lacks clear indicators of its origin, request additional context to better inform your analysis
___
I’ve recently been introduced to CitizenAI. It combs local legislation (video and audio) and transcribes text for you. I research Bitcoin at the local level and this helps me decide on who to vote for.
Oh cool. I’m going to check that out! Thank you Larry.
Link:
https://citizenportal.ai/
This is such an interesting side. Thank you for forwarding this along. There is quite a lot of value in using the site.