Explain it for a specific reader

Why it works

Most explanations from AI sit in the awkward middle: too technical for a beginner, too shallow for an expert. This prompt forces a single audience target and a length cap, which collapses the model's default toward something usable. The named-audience pattern matters - 'explain this to a smart 12-year-old' produces a measurably different output than 'explain this simply,' because the model can simulate a specific reader. The 100-word cap is the second forcing function: without it, the model pads. With it, you get the version you'd actually share with someone.

Example output

For a smart 12-year-old, in under 100 words:

A database index is like the index at the back of a textbook. Without it, finding 'photosynthesis' means reading every page. With it, you flip to the index, see 'photosynthesis - page 47,' and jump there. The cost is that every time someone adds a new page, you have to update the index too. So indexes make finding things fast, but make adding things a tiny bit slower. You add them for the questions you ask a lot, not for every question you might ever ask.

Common mistakes

Don't pick an audience you don't actually have to explain to - 'a 5-year-old' is overdone and produces baby-talk that no real 5-year-old needs. Pick the actual person: 'my non-technical co-founder,' 'a new hire on day one,' 'a skeptical CFO.' The more specific the named audience, the better the output. Also: 100 words is a hard cap, not a target - if the model returns 95 words, it's calibrated; if it returns 150, repeat the constraint.