This one was really interesting for me personally as my experience with code generation is limited, but the spectre of LLMs seems to haunt any current conversation about AutoML.

It's also a companion to last month's about foundation for optimization. Give that a listen if you haven't already!