[bilekas]

> Now consider a different arrangement. The LLM generates code that actually runs inside your application – at call time, every time the function is invoked.

I'm sure there's a lot of effort put into this, god knows why, but I pray I never have to have this in a production environment im on.