We describe the persona selection model (PSM): LLMs learn to simulate diverse characters during pre-training, and post-training elicits and refines a particular "Assistant" persona. Interactions with an AI assistant are then well-understood as interactions with the Assistant — something roughly like a character in an LLM-generated story. The work surveys empirical behavioral, generalization, and interpretability evidence for PSM and discusses consequences for AI development, including anthropomorphic reasoning about model psychology and seeding positive archetypes in training data.