Honestly, I used to think coding was way out of my league. I’m just an ordinary person—coding? That's what programmers do. But last month, I found myself forced to use AI to write code due to a particularly silly requirement.
The request was simple: I watch dozens of English videos every day and wanted to quickly generate Chinese summaries with timestamps. The tools out there were either too expensive or too complicated. I thought, can’t I just create a tool myself? That night, I opened up a free AI coding tool.
And guess what? In just three hours, I built my first working tool from scratch. Sure, the interface looked like a joke, but it actually worked! In that moment, I felt like I had suddenly grown a pair of programmer's hands. This wasn’t about “learning to code”; it was about “cheating the system.”
Later, I did the math: if I had outsourced the same requirement, it would have cost at least 5,000 yuan. A subscription for a ready-made tool would have set me back a few hundred each month. But I only spent three hours and a few cents on electricity. But that’s not the main point.
*The real revelation I had was a bit unsettling: our generation is experiencing a silent “ability divide.” *
On one side are those still asking “which tool is the best.” They scroll through reviews, watch comparisons, and get bogged down in parameters without taking action. On the other side are people like me, who create a simple solution regardless of the tool's quality. The gap isn’t about the tools; it’s about the distance between “thinking” and “doing.”
I later tried out all the mainstream free tools: Cursor, Claude Code, and various open-source projects. To be honest, the differences between these tools are nowhere near as significant as the marketing hype suggests. They can all help you write code.
The difference lies in their approach: some act like patient teachers, guiding you step by step, while others are like efficient assistants, churning out results based on your commands. But the core remains the same—helping you, a novice, turn your ideas into code. Here's a surprising fact: the best way to learn isn’t “learn first, then do,” but rather “learn while doing.” You don’t need to understand data structures or algorithms; you just need a specific problem. For example, “How can I automatically generate QR codes from a column of data in Excel?”
Then, you tell the AI your request in plain language. It’ll provide you the code and guide you through each step, telling you where to click and what commands to type. If you run into errors, just copy and paste them to the AI, and it’ll help you fix them. It’s that simple.
But let me throw some cold water on this. Many people get stuck in these three traps:
Searching for the “perfect tool” before starting. Wake up! Tools are always evolving. The “best” tool you find today might be obsolete next month. Your core asset isn’t familiarity with a tool; it’s the muscle memory of “using AI to solve problems.”
Setting overly ambitious goals. Don’t jump straight into building a “product” for others. Your first project should be embarrassingly small. For instance, I created a tool that automatically organizes my saved articles on WeChat into a weekly reading report. It served only me, but it saved me two hours each week.
Thinking you need to “study systematically.” This is the most harmful mindset. AI programming is a practical skill, not a theoretical subject. You don’t need to watch tutorials for three months first. What you need is: an idea -> take action -> encounter a problem -> solve it -> keep going.
Leveraging your skills in this way is the greatest benefit of our time. My current workflow looks like this: I have a chat AI (like DeepSeek) on one side to break down complex problems and design architecture, and a coding AI (like Cursor) on the other side to handle the actual coding and debugging. With both windows open, I feel like a conductor leading two tireless super brains.
But the most satisfying moment isn’t just how useful the tools are. It’s when a colleague is still complaining about a repetitive task, and I’m quietly thinking, “I’ll just write a tool to automate that tonight.” That confidence—that “I can solve this”—is the strongest foundation AI programming gives you.
Tools can help you write code, but your perspective helps you see the problems. In the past, I looked at the world and saw “what tools do I need?” Now, I see “what tasks can be automated?” This shift in perspective doesn’t just enhance efficiency; it explodes your options.
Finally, here’s the plain truth. Free tools are out there, and tutorials are abundant. What you’re missing isn’t resources; it’s that fire of “I’m going to figure this out today.” Don’t wait any longer—open a tool right now, find a repetitive task that annoys you, and try to use AI to eliminate it.
Even if you fail the first time, you’ll still be ahead of 99% of people who only stay in the “thinking” phase.
Top comments (0)