Introduction
Over the past year, I've been creating AI-related content and receiving numerous questions from people asking, "How do I start learning AI?" In this guide, I'll share my learning journey, practical tips, and recommended resources to help you start your AI learning journey with confidence.
1. Start with Solving a Specific Problem
The biggest mistake beginners make is starting with "how to learn AI?" This question is too broad and will lead to information overload and confusion. Instead, start with "how can AI help me solve a specific problem?"
My Example: Knowledge Management
I started with a specific pain point: managing my reading notes. I read 20-30 minutes every day, but my notes were scattered and hard to review. I wanted AI to help me organize my Obsidian notes.
Practical Steps:
- Import highlights from Apple Books
- Use AI to summarize and organize notes
- Create a workflow for continuous note management
2. Learn AI Terminology Like a Baby
When you first enter the AI world, you'll encounter many new terms. Don't get overwhelmed. Learn them like a baby learning a new language.
My Experience
When I first took a course on AI agents, a 30-minute lesson took me an hour to complete because I had to stop and look up every new term.
How to Learn New Terms:
- Use conversational AI tools to explain terms in simple language
- Create a personal glossary
- Review and reinforce your understanding
Example: Understanding CLI
When I first encountered "CLI," I was confused about why there were so many different names like ChatGPT, Codex CLI, and GPT CLI. I used AI to explain the differences in simple terms.
3. Learn by Doing: Solve Problems Step by Step
Once you have a basic understanding of AI terms, start solving problems. Don't wait until you know everything — learn by doing.
My Second Project: Auto-Edit Video Agent
I found editing talking head videos very time-consuming, so I wanted to create an AI agent to help me auto-edit videos.
Practical Steps:
- Define the problem clearly
- Ask AI for solutions
- Implement the solution step by step
- Iterate and improve
4. Recommended AI Learning Resources
Short Videos
- @一枚卓子老师: Explains AI terms in very easy-to-understand language
- @张咋啦老师: Provides practical, down-to-earth content that is updated frequently
Long Videos and Courses
- @秋芝 2046's Codex Course: Great for learning specific tool usage and solving specific problems
- Folder as an App Course: Excellent for learning how to build AI agents
Official Documentation
Always check the official documentation of AI tools. No one knows a tool better than its creators.
Example: ElevenLabs API
When I was learning the ElevenLabs API, the official documentation helped me understand:
- What the tool can do
- What possibilities it offers
- How to call it specifically
Conclusion
AI is developing rapidly, and you can't learn everything. Don't get anxious. Focus on how AI can help you solve specific problems. That's where its real value lies.
Remember:
- Start with a specific problem
- Learn terminology gradually
- Learn by doing
- Use a variety of resources
I hope this guide helps you start your AI learning journey with confidence. I'll continue to share more practical ways to use AI to solve real-world problems.
常见问题
What's the very first thing I should do to start learning AI?
Pick one specific problem you already face and ask how AI can help with it. This is the opposite of starting with "how do I learn AI" — a broad question that leads to information overload. A concrete starting point could be organizing your notes, summarizing long articles, writing better emails, or automating a repetitive task. Once you have a target, you'll naturally discover which tools you need, and every new term you learn will stick because you immediately see how it applies to your problem.
I get overwhelmed by AI terminology. How should I learn it?
Treat it like a baby learning a language: don't try to memorize everything at once. When you hit an unfamiliar term, ask a conversational AI tool to explain it in plain language, add it to your own personal glossary, and review the terms that come up in your actual work. The CLI example in this article is a perfect case — "ChatGPT, Codex CLI, and GPT CLI" sound like three different things but are really one concept with different contexts. Terms you encounter while solving a real problem are the ones you'll remember, so let your projects drive your vocabulary growth.
Do I need to understand the math or coding behind AI before using it?
No. Every project in this guide — organizing notes, auto-editing videos — was built by describing a problem and iterating with AI, not by writing machine learning code from scratch. The value of AI for most people is in applying it to tasks, and the fastest way to get there is by doing, not by studying theory first. You can always go deeper into concepts later if a specific project needs it; starting with hands-on use gets you results immediately and builds the motivation to keep going.
Which learning resources should I trust most?
Combine three layers. First, the official documentation of whatever tool you're learning — no one knows a tool better than its creators, and the ElevenLabs API example shows how official docs clarify what a tool can do, what's possible, and how to call it. Second, short explainer videos for quick wins on terminology. Third, longer courses only when you're committed to a specific tool or skill. The key is matching the resource to your current question: docs for specifics, short videos for concepts, and courses for deep dives — not consuming everything at once.