I wonder if we will ever get AI-optimized language sets, ie code instructions given specifically to AI agents to develop code for a JIT interpreter. This gives the same results each and every time the instruction to AI is given no matter the version of LLM used.