I'm waiting for the good AI powers software.... Any day now.

Ideally, llm should be able to provide the capability to translate from memory inefficient languages to memory efficient languages, and maybe even optimize underlying algorithms in memory use for this.

But I'm not going to hold my breath