编程中的AI助手

Solving bugs is definitely helped by AI, as identifying the root cause is essentially a search problem. AI excels in searching critical code snippets, making the fix process much simpler and quicker.

Prototyping, planning, and conceptual design. When ideas are vague, generating a clickable UI prototype with AI can be more effective than mere brainstorming. By interacting with this prototype, one can more easily spot interaction issues and generate new ideas.

Architectural design. After discussing with AI about the architecture, I often let it draft a preliminary framework for classes and methods, as well as core data structures, which I then review for logical consistency.

Refactoring, but only for small tasks. For instance, when abstracting several methods from one module into a new one, AI usually performs quite well.

Code review. Having AI review my code after completion can lead to valuable suggestions and sometimes even uncover hidden issues.