📖 This is Part 5 of 5 in the “AI Path L0→L1 Upgrade Guide” series — Series Navigation + Graduation Checklist.
Series Navigation
| Part | Topic | Core Content |
|---|---|---|
| Part 1 | Understanding Your Tools | LLM fundamentals (not a search engine), working memory vs. long-term memory, mainstream platforms and specialized tools |
| Part 2 | From Vague Questions to Precise Instructions | The RBGO prompt framework, Chain-of-Thought reasoning, format constraints |
| Part 3 | Turning AI Into Your Collaboration Partner | Iterative follow-up questions, context management (new conversations / progress summaries / chunked processing), role-playing |
| Part 4 | Building Your Personal System | Prompt library, scenario-to-tool mapping (international and China options), layered knowledge management |
| Part 5 | Graduation & Next Steps | L1 graduation checklist, L1→L2 dual-path preview |
Graduation Checklist: Have You Reached L1?
Four weeks of practice — and here we are. Before I declare you “graduated,” go through this checklist honestly:
- Can you explain the core working principle of an LLM (probabilistic text generation, finite context window)?
- Do you understand the difference between working memory and long-term memory, and know which your current platform supports and how to manage it?
- Have you internalized the RBGO prompt framework, naturally including role and background in everyday prompts?
- Can you tell when to use Chain-of-Thought (analytical reasoning) and when it’s unnecessary (simple lookups)?
- Have you built a habit of iterative follow-up — rarely accepting AI’s first answer as the final one?
- Do you maintain a prompt library with at least 10 battle-tested prompts?
- Do you know which tool works best for each of your core use cases?
- Can you gauge the reliability of AI output, and proactively verify critical information?
If all eight boxes are checked — congratulations, you’re a certified L1 user. You’ve gone from “hoping for a good answer” to “designing for one,” and that leap will keep paying dividends every single day.
If a few items are still unchecked — that’s perfectly fine. Go back and drill those weak spots. This guide isn’t meant to be read once and forgotten; it’s your reference manual.
What Comes Next: The L1→L2 Threshold
If L1 already feels natural and you’re starting to wonder “what else is out there,” the direction to L2 is already coming into view:
The key leap from L1 to L2: from using AI to building AI-powered tools.
L2 means you’re no longer just chatting with AI — you’re starting to make AI do things for you. There are two paths:
Path 1: Build automation tools via APIs. Sign up for an API account, write a few lines of Python, and print your first AI-generated reply. International options include Anthropic API, OpenAI API, and OpenRouter. China-based options include DeepSeek API, Kimi API, 智谱 (Zhipu) API, MiniMax API, 火山引擎 (Volcano Engine) API, and 小米 (Xiaomi) API — note that DeepSeek, Kimi, Zhipu, and MiniMax also offer overseas subscription services, so choose based on your region and cost preferences.
Path 2: Use autonomous execution AI directly — tools like Claude Code, OpenAI Codex, and OpenCode. You don’t need to write code, but you do need to be comfortable describing tasks from the command line. If the “autonomous execution AI” concept from Part 1 caught your eye — having AI batch-process files and automate repetitive operations — this path may deliver results faster than learning APIs.
And if neither path appeals to you right now — that’s completely fine. L1 skills already cover the vast majority of everyday scenarios. Knowing how to drive doesn’t mean you need to build a car; knowing how to use AI doesn’t mean you must call APIs. When you eventually hit a genuine “I really wish I could automate this workflow” moment, L2 will come naturally.
The planned L1→L2 upgrade guide will involve command-line operations or basic programming. We’ll meet again when that time comes.
Getting from L0 to L1 took you a month. But from this day forward, every conversation you have with AI is no longer a roll of the dice — you know what you’re doing, you know how to get the results you want, and you have a system that’s truly yours. This is the real starting point of treating AI as a tool.