14 Sep 2025

Morsel #26: vibe recursion

Filed under: AI | tech | vibes

What if you wrote a prompt (Prompt n) that asked the LLM to write a better prompt (Prompt n+1) than this prompt (Prompt n) and to ask for a better prompt to ask for later (which would be Prompt n+3). And then you repeat the process over and over to see the results.

Would it become asymptotic and reach a point where it couldn’t produce a “better” prompt or would it go wild and lose quality and meaning?