This conversation presents a practical operating model for innovation in an AI-driven world. Josh argues that companies need independent frontier teams, informal sources of ideas, fast experimentation, and a willingness to pivot or kill weak products. He describes how to identify product-market fit through durable user behavior and visible user excitement, then explains Google Labs’ lifecycle from zero-to-one through scaling. The discussion also covers AI-era skills such as unlearning, sustained intensity, trust, and cross-functional collaboration; planning through rolling windows and short milestones; and the organizational conditions that help labs succeed. It closes with competing visions for future products and the rituals Google Labs uses to reinforce product quality and team values.