Ask HN: Does AI research need "world models" more than bigger LLMs?

A Hacker News user asks whether AI research needs 'world models' more than larger large language models, suggesting that generating hypotheses is no longer the main challenge. The user argues the harder problem is continuously collecting real-world data, building accurate world models, and using them to guide experiments, questioning if automated research is heading in that direction. Applications for an unspecified program are open until July 27.