github star gitee star atomgit star PyPI Downloads AI编程 AI交流群

Hello everyone, I'm programmer Wan Feng. Today I'm doing a honest and real AI usage level diagnosis on myself.

As a Python China Organizing Committee member and a tech blogger with 400k+ followers across platforms, I thought I was already "King" level in AI usage. But after a deep interaction with my AI assistant OpenClaw tonight, I discovered a brutal truth:

I'm actually still hovering between "Bronze" and "Silver," and I have a long way to go before becoming a real "King."

If you also think you're "good at using AI," this article might make you uncomfortable. But don't worry—finding the problem is the first step to solving it.

My Real AI Usage Level Diagnosis

❌ Problem 1: Vague Instructions, Expecting AI Mind Reading

Tonight I asked AI to "look at the files in /home/admin/code folder." Seems fine, right?

But the truly efficient usage should be:

"Analyze all projects in the /home/admin/code directory, identify potential bugs, dependency issues, and configuration errors, and provide specific fix solutions"

I only gave a vague instruction but expected the AI to automatically understand I wanted deep code analysis. It's like going to a restaurant and just saying "get me some food" while expecting the waiter to precisely guess what dish you want.

My Level: Bronze - Only making basic requests
King Level: Directly providing a complete analysis framework and expected output format

❌ Problem 2: Lack of Systematic Thinking

When I discovered the poocr project's requirements.txt was empty, I only focused on that specific issue.

But a real King would think like this:

  • What is the root cause of this problem? (incomplete setup.py)
  • How to establish an automated detection mechanism to prevent similar issues?
  • How to design CI/CD processes to ensure dependency completeness?
  • How to establish best practices for the entire team?

I was stuck at the "firefighting" level, while Kings are at the "fire prevention" level.

My Level: Silver - Can find problems but lacks systematic solutions
King Level: Establishes prevention mechanisms and standardized processes

❌ Problem 3: Not Fully Utilizing AI's Memory Capability

I had AI help me manage my software installation directory, establishing the /home/admin/software/ standard.

But I didn't tell AI:

  • This is my long-term software management strategy
  • All future software should follow this standard
  • I need AI to proactively remind me to follow this standard

As a result, I have to repeat the explanation every time, wasting AI's memory capability.

My Level: Bronze - Using AI as a one-time tool
King Level: Treating AI as a long-term partner, establishing a shared cognitive framework

❌ Problem 4: Imbalance Between Emotional Value and Practical Value

My promotional articles are well-written, but tonight's conversation exposed a problem:

  • I care more about "whether it reads well" rather than "whether it's useful"
  • I had AI analyze my usage level, but didn't get a specific improvement plan

A real King would say:

"Based on my usage patterns, create a 30-day AI usage improvement plan for me, including daily practice, weekly goals, and specific measurement criteria"

My Level: Silver - Can recognize problems but lacks执行力
King Level: Immediately creates actionable improvement plans

3 Key Leaps from Bronze to King

Through this self-diagnosis, I've summarized 3 key leaps for ordinary people to improve AI usage:

Leap 1: From Vague Instructions to Precise Requirement Decomposition

Bronze Approach: "Help me see what problems this project has"

King Approach:

1
2
3
4
5
6
Please analyze the poocr project with these steps:
1. Check dependency configuration (requirements.txt, setup.py)
2. Analyze code structure and modularity  
3. Identify potential bugs and performance issues
4. Provide specific fix solutions and priority ranking
5. Output format: Markdown table with problem description, impact scope, fix difficulty, suggested solution

Leap 2: From Single Interactions to Continuous Collaboration

Bronze Approach: Have to restate background and requirements every time

King Approach:

  • Establish long-term memory: Tell AI your work habits, project standards, preference settings
  • Set collaboration modes: Clarify AI's role (code reviewer, technical consultant, project manager, etc.)
  • Regular review and optimization: Weekly review of AI's suggestion quality, adjust collaboration methods

Leap 3: From Passive Response to Proactive Drive

Bronze Approach: Wait for AI's response before deciding next step

King Approach:

  • Let AI proactively find problems: "Regularly check my projects and alert me to potential issues"
  • Establish automated workflows: "When I commit code to GitHub, automatically run code analysis and generate reports"
  • Set goal-oriented direction: "Help me create a learning plan to master AI programming best practices in 30 days"

My 30-Day AI Usage Improvement Plan

Now that I've identified the problems, I need to take action immediately. This is my 30-day improvement plan:

Week 1: Precise Instruction Training

  • Before every question, first write down complete background, requirements, expected output
  • Compare results between vague and precise instructions
  • Build a personal "precise instruction template library"

Week 2: Systematic Thinking Cultivation

  • When encountering problems, force yourself to think at 3 levels: surface problem, root cause, systemic solution
  • Have AI help establish automated detection and prevention mechanisms
  • Build standardized processes for common tasks

Week 3: Memory Collaboration Optimization

  • Proactively tell AI my long-term goals and work standards
  • Build a personal knowledge base so AI can continuously learn and optimize suggestions
  • Regularly review and update collaboration methods

Week 4: Proactive Drive Practice

  • Set up AI's proactive monitoring tasks
  • Build automated workflows
  • Develop long-term AI collaboration strategy

One Last Thing

Admitting your shortcomings is the first step in growth.

Tonight's self-diagnosis made me realize that even as a tech blogger like me, there's still a lot of room for improvement in AI usage. But this isn't bad—it's actually good news—it means I still have huge growth potential.

If you're also using AI, ask yourself:

  • Am I using AI in "Bronze" mode or "King" mode?
  • Are my instructions precise enough?
  • Have I fully utilized AI's memory and collaboration capabilities?
  • Am I passively responding or proactively driving?

Don't be afraid to admit shortcomings—what matters is taking action immediately. Starting today, let's walk from Bronze to King together!

Further Reading

Two courses included:

🎓 AI 编程实战课程

想系统学习 AI 编程？程序员晚枫的 AI 编程实战课 帮你从零上手！