Visual Generation Unlocks Human-Like Reasoning through Multimodal World Models | AI Papers Podcast Daily | Podwise