AI becomes more useful when treated as a collection of specialized working modes rather than a single-purpose chatbot. By using specific commands, users can direct AI to adopt different thinking styles for tasks like debugging, algorithm design, or research planning. This shifts the interaction from a simple question-and-answer format to a structured workflow involving goals, context, specialized lenses, analysis, and iteration. The text details numerous "lenses" or commands categorized by function, including transforming text, controlling tone, structuring writing, compressing information, and managing meetings. Project management lenses help decompose large goals into executable systems with measurable outcomes through OKRs and KPIs. Risk analysis and root-cause analysis lenses provide frameworks for identifying uncertainties and understanding failures. Productivity lenses focus on prioritization and feedback, while learning lenses encourage active retrieval and practice. For developers, coding-specific commands cover generation, explanation, debugging, refactoring, and optimization, each with distinct purposes. Algorithmic and data structure lenses guide problem-solving for technical interviews. The text also highlights lenses for working with data formats like SQL, designing APIs, and conducting research. Crucially, it emphasizes the difference between opinions and testable hypotheses, and between peer review and simple criticism. The /audit lens offers a broader inspection capability, and the meta-lens /framework helps select appropriate problem-solving methodologies. Ultimately, leveraging these specialized modes allows for more precise and effective AI utilization.