The episode opens with a long discussion of the OpenAI / Scarlett Johansson controversy. Andrew says he had a direct view of GPT-4o voice development, that OpenAI hired actors with disclosures and fair pay, and that there was never an intent to copy Johansson’s voice. The hosts argue that the resemblance was driven by audience expectations shaped by Her and by a familiar voice archetype, not by a plan to mimic her. From there, the conversation moves through AI’s strengths and weaknesses: emotional companionship, chatbot use for counseling and rewriting messages, multimodal video analysis for inventorying property, and hallucinations or odd outputs from models. The latter half shifts into space news and speculation, covering Ed Dwight’s Blue Origin flight, NASA’s Artemis and Starliner problems, commercial launch competition, the X-37 military spaceplane, whale communication and alien contact analogies, Dyson sphere detection, and finally a recommendation for Severance. Key topics OpenAI voice casting and the Scarlett Johansson controversy: Andrew says OpenAI hired real actors, gave disclosures, and had no intent to copy Johansson’s voice. They discuss how people linked the voice to Her and to a voice archetype, and why the public narrative became simplified. AI companionship and emotional dependence: The hosts discuss AI as a personal companion, including pretend girlfriends and emotionally supportive chatbots. They treat it as potentially helpful for some users but also poten