Using AI to Write Prompts for AI
User says "draw a cat at sunset," but the image model needs an entirely different description — the fix is using an LLM as translator
THE QUESTION THIS PAGE ANSWERS
ANSWER FIRSTWhat is the key idea behind “Using AI to Write Prompts for AI”?
User says "draw a cat at sunset," but the image model needs an entirely different description — the fix is using an LLM as translator
Make the claim earn its place. Use this page as a decision aid, not a definition to memorize. Connect the idea to one real task, one observable result, and one failure that would change your mind.
Write one question you could answer with evidence after trying this idea.
A conclusion that sounds complete but leaves the key assumption untested.
① Users can't write image gen Prompts: they don't know whether to specify "golden hour lighting" or "illustration style"
② Image models can't understand vague intent: "daydreaming" is not a visual description to the model
③ Each image model has its own dialect: Midjourney, DALL-E, and Stable Diffusion each prefer different Prompt styles
How “Example 1: Daydreaming on the Balcony” changes an answer
“User says "draw a cat at sunset," but the image model needs an entirely different description — the fix is using an LLM as translator” shows that a model does not process the “word count” we see. It processes Token pieces. Tokenization affects input length, how much context fits, and how much computation a request consumes.
Length, information, and context are different
As “User says "draw a cat at sunset," but the image model needs an entirely different description — the fix is using an LLM as translator” grows, separate three questions: how many Tokens the text becomes, which pieces can change the current decision, and whether older material has fallen outside the context window. Removing repetition is often more useful than simply making the window larger.
Keep what can change the decision
Use “User says "draw a cat at sunset," but the image model needs an entirely different description — the fix is using an LLM as translator” as an A/B test: keep the same question while removing repeated background, compressing format, and trimming irrelevant history. Compare answer quality, latency, and Token count.
From “Example 1: Daydreaming on the Balcony” to “Example 2: Morning Cooking”
“Example 1: Daydreaming on the Balcony” grounds the problem in “User says "the character is daydreaming on the balcony" LLM translates into image generation Prompt A young woman named the character standing on a sunlit balcony, leaning against the railing, gazing into the d…”. “Example 2: Morning Cooking” then moves it toward “User says "the character is cooking" LLM translates into image generation Prompt A young woman named the character in a bright modern kitchen during morning time, cooking breakfast. She has shoulder-length dark…”. Together, they show that the lesson is not just a conclusion to remember, but a claim with conditions.
Carry the judgment into the next situation
For long text, keep what can change the conclusion before compressing format and history. A larger context is worth its cost only when the added information is useful.
- “Example 1: Daydreaming on the Balcony”: User says "the character is daydreaming on the balcony" LLM translates into image generation Prompt A young woman named the character standing on a sunlit balcony, leaning against the railing, gazing into the d…
- “Example 2: Morning Cooking”: User says "the character is cooking" LLM translates into image generation Prompt A young woman named the character in a bright modern kitchen during morning time, cooking breakfast. She has shoulder-length dark…
The final “Finish by testing the claim” brings the discussion to “User says "draw a cat at sunset," but the image model needs an entirely different description — the fix is using an LLM as translator”. The useful thing to carry forward is knowing which judgments must be revisited when input, scale, or risk changes.
I turned one judgment from this article into a small experiment I could run today. Knowing what to observe next is more useful than simply remembering the conclusion.
After reading this, I first looked for the conditions behind the idea instead of copying the method into a project. That order made the later trade-offs much clearer.
When this judgment reaches real work, which constraint should be added first? I am curious which step matters most between reading and the first practical attempt.
No discussion on this article yet.