The episode is largely a deep dive into OpenAI's Sora app, with the hosts describing it as more than a video model and instead a new social-media modality built around short generated clips, personal cameos, remixing, and highly shareable strange or funny scenes. They discuss its rapid rise in the App Store, invite-only rollout, the technical jump in Sora 2 Pro, voice and character consistency, and the ways the app is already changing how they think about video, deepfakes, and even the simulation hypothesis. A major thread is the business and cultural impact of Sora: the hosts argue that likeness controls, meme culture, and fan-made IP uses could create new monetization models, including ad-supported video generation and possible revenue-sharing with rights holders. They also discuss how Sora may become a creator-friendly tool rather than a threat, how its clips are spreading to other platforms as memes, and then close with recommendations for Weapons and The Studio, plus a brief look at OpenAI's newer ChatGPT app and image-generation products. Key topics Sora as a social video app: The hosts repeatedly frame Sora as a social feed, not just a model, describing generated clips as shared daydreams or thoughts and emphasizing its strange, personal, and culturally sticky feel. Cameos and likeness permissions: They explain the cameo feature, where users can create avatars and set guardrails for likeness use, including restrictions on political content or other categories. Mark Cub