Will our functions transform into AI text prompts from interpreter instructions?

I've been implemented a JavaScript AI emoji generator recently. And as OpenAI API responses are not always consistent, or maybe my prompts are far from ideal, I needed to create a backup function that generates random emojis if the response is not parsed properly.

Surprisingly, I found that functions might more look like an instruction for a person than for a machine in future 😅. Here is what I mean:

// regular JS function
function getNEmojis(n) { …

ai api backup consistent emoji found function functions generator javascript node openai openai api programming prompts random responses text