[D] Is OOD generalization still a future in the LLM era?

I think OOD generalization is an important issue because it pulls in the distance from reality. But I am concerned that recent conferences like ICLR, ICML, NeurIPS etc. don't have many people working on this problem. So, I wonder if it's because of some difficulties in this field that people stopped studying it. Or is it because of some other reason.

conferences etc future iclr icml important issue llm machinelearning neurips people problem reality think