Why vague prompts produce vague answers

The model predicts what a good answer usually looks like. When your prompt is vague, 'usually' means the statistical middle — generic, hedged, everyone-shaped. Every specific you add moves the answer away from the middle and toward your situation.

The three most expensive missing pieces

• Missing audience — the model writes for everyone, which means no one.
• Missing length — you get a wall of text when you wanted a paragraph, or vice versa.
• Missing example — one example of the style you want is worth ten lines of description.

Show, don't only tell: paste one example of the output you like (even a rough one) and ask for 'the same style'. Models are extremely good at mimicry and merely okay at guessing.