Alexandr Wang traces Muse from an intense personal-agent experiment to a polished consumer product launched after seven months of iteration. He explains why Meta chose personal AI over coding agents, how the team measured reliability, and why Muse’s focused design, internet-native marketing, and word-of-mouth growth helped it stand out. The conversation then turns to Wang’s move from Scale AI to Meta, Zuckerberg’s role in the decision, and the organizational model Wang is building: small, technically strong teams, high autonomy, minimal bureaucracy, and a willingness to fund many experiments for the few breakthroughs that matter.