Let the LLM's magically write code for the optimal estimators, using existing theory; but that code will implement interpretable math with provable characterist

Let the LLM's magically write code for the optimal estimators, using existing theory; but that code will implement interpretable math with provable characteristics, as opposed to magic.