Part I · Chapter 17
17. Reversed meta-prompting
Major question
Can AI reconstruct a reusable prompt from an example of the result I want?
Description
Reversed meta-prompting moves from output to prompt. The user provides a desirable result and asks the model to infer the context, instructions, structure, constraints, voice, and evidence requirements that might have produced it.
Below is an example of the kind of paper I would like to obtain. Analyse its structure, depth, argumentation, perspective, terminology, use of evidence, and writing style. Based on these characteristics, reconstruct a detailed prompt that could have produced a paper like this. Do not reproduce the paper. Give me a reusable prompt for another paper of comparable quality and structure on the economic value of beauty in tourism.
[Insert example paper here.]
Example files to download
Expected impact
The reverse meta-prompt inverts the direction of work: from observable features to a generative specification. You get a testable, refinable prompt - not merely a commentary on the text.
Desired result -> analyse characteristics -> infer requirements -> reconstruct prompt -> test.