Founder of Baixing.com: My Fourteen Tips on Using Claude Code

A brief record of my experience with Claude Code so far. This is purely my personal exploration and may not be applicable to everyone.1. Focus on one tool and use it intensively. I use Claude Code. I don't think it's better than Codex, but the ROI of the effort spent comparing tools isn't necessarily higher, although being able to articulate the differences convincingly gives a false sense of accomplishment.2. Remember the most important keyboard shortcuts. Control+G opens the editor, which helps with writing longer text; Control+A, Control+E, and Control+U are very useful shortcuts for quickly moving the cursor in the command line. Although they're not new to the AI era, they're just as important as Control+C and Control+V when using them.3. Use voice input. HoldSpeak is very helpful. 4. For each project, start by writing a PROJECT.md file, using a structured approach to write down all your thoughts at once. 5. Claude agents are the default. 6. Claude Code, github.com, and cloudflare.com are a perfect match; delegate the build, deployment, and all domain-related operations to the infrastructure. 7. Separate human-written and machine-written code. Manually maintain the core CLAUDE.md file; do not read Claude Code's .md files or code. Machines are machines, and humans are humans. 8. Drag and drop files into the Claude Code window—audio, video, documents, screenshots—if you can't explain it clearly, use Command+Shift+5 to take a screenshot and then drag it over; this is the fastest method. 9. Reconstruct the memory system. Centered on ~/.claude/CLAUDE.md, categorize and reference multiple memory files. The requirement is to avoid using the project's memory and place all memory files in Git, syncing them to GitHub (private). This ensures your memory is permanent and accumulative, rather than scattered across various projects. 10. Write skills, and after each task, require Claude to "contain what he has learned into the skill"—this can be automated. 11. When possible, use Ultracode to trigger dynamic workflows for complex tasks. Although expensive and slow, the results are guaranteed. 12. Accumulate and refactor skills along the way. Skills need to be stored in Git. 13. Use Git documentation as the output of the previous task and the input of the next. Ensure agents have clear handover documentation, rather than relying on context for connection. 14. Treat Claude Code like a horse (or a person), not a chariot. A chariot turns under its own command, a horse has its own ideas; we only need to set the goal and boundaries. Its autonomous pathfinding feature is a characteristic, not a bug.