I first encountered AI-assisted programming—often referred to as “Vibe Coding”—last year. I had never built a Chrome extension before, but with the help of AI, I managed to complete one in just ten minutes. The extension converts all of a user’s browser bookmarks into a web-based navigation page; I used AI to generate the web interface, and it worked perfectly once deployed. It struck me that AI has significantly lowered the barrier to entry for programming—to the point where even a complete novice could become a developer.
Later, while maintaining a legacy VB6 application, I ran into an issue where the MSCOMCTL.OCX control failed to load. Upon inspecting the code, I discovered it relied on a ProgressBar control. I used AI to rewrite the component, simulating a progress bar using a PictureBox instead. The task was completed with almost no modifications to the original source code, demonstrating an incredibly high level of development efficiency.
Through further exploration, I’ve found that “Vibe Coding” is a fascinating approach that enables one to accomplish unexpected tasks—offering immense benefits, especially to programmers. While there has long been a concern about programmers becoming “obsolete” after the age of 35, mastering AI-assisted programming allows even older developers to build a vast number of applications quickly and efficiently. In this new landscape, younger programmers may no longer hold a distinct advantage. Programming is no longer just “manual labor”—since AI can handle the bulk of the coding—so regardless of whether you are a professional programmer, it is well worth learning about AI-assisted programming and trying your hand at building something.
Drawing from my own experience, I will walk you through the entire process—from setting up your “Vibe Coding” environment to development, debugging, and final deployment.
- Environment Setup
For your editor, choose VS Code or one of its AI-enhanced variants, such as Trae, Antigravity, or Cursor. Once you are familiar with the interface, configuration, and features of VS Code, you will find it easy to navigate almost any other AI-powered editor. You can download Visual Studio Code from the official website: https://code.visualstudio.com. By default, the application is in English; however, you can switch the interface to Simplified Chinese by searching for “Chinese” in the “Extensions” sidebar and installing the first result. For code management, use Git—an open-source distributed version control system. Download it here: https://git-scm.com
Post-installation configuration:
git config –global user.name “Your Name”
git config –global user.email “Your Email”
For code hosting, use GitHub. To sync all your GitHub projects to your local machine at once:
Install GitHub CLI (run in PowerShell): winget install GitHub.cli
Log in to GitHub (run in CMD): gh auth login
Batch-sync all projects to your current local directory: for /f “tokens=1” %a in (‘gh repo list –limit 1000’) do gh repo clone %a
Accessing GitHub from within China can be unreliable (with roughly a 50% failure rate), so setting up a proxy is recommended.
The local proxy server is assumed here to be: socks5://127.0.0.1:10808 (you may adjust the port number to match your specific setup).
To configure Git to use a proxy when accessing GitHub, follow these steps:
Open your terminal and run the following commands:
git config –global http.https://github.com.proxy socks5://127.0.0.1:10808
git config –global https.https://github.com.proxy socks5://127.0.0.1:10808
Alternatively, you can modify the configuration file, which is typically located in your user home directory:
C:\Users\YourUsername.gitconfig
Open the file using any text editor (such as Notepad or VS Code) and append the following content to the end of the file. [http “https://github.com”]
proxy = socks5://127.0.0.1:10808
[https “https://github.com”]
proxy = socks5://127.0.0.1:10808
To verify that the configuration was successful, run:
git config –global –list
For programming, choose Python—it is currently the mainstream language for AI-assisted development. Python is remarkably similar to VB6; their syntactic structures are nearly identical, and the underlying logic is almost exactly the same, with only minor differences in syntax. Much like the old DOS-based GW-BASIC, the code is intuitive and reads almost like natural language. In essence, Python can be thought of as a cross-platform, feature-rich, and library-heavy “enhanced version” of VB6. I have even successfully converted legacy VB6 programs into Python.
There are many AI models to choose from. International options include Claude Code, OpenAI Codex, Microsoft Copilot, and Google Gemini, while domestic (Chinese) options include Doubao, GLM, Deepseek, Kimi, and Qwen. Note that accessing international AI models requires the use of a proxy.
To configure VS Code to use a proxy server:
First, install proxy server software locally. Then, open VS Code, go to “Settings,” and search for “proxy.” Enter your proxy server address in the “Http: Proxy” field—for example: "http.proxy": "socks5://127.0.0.1:10808". Once configured, you will be able to access international AI models within VS Code.
GitHub Copilot is the default AI model for VS Code; its free usage quota resets monthly, but I found that 20% of the monthly allowance was consumed in just one day—hardly enough for software development. OpenAI Codex can be installed by searching for “Codex” in the extensions marketplace; its free quota resets weekly, and while I used 27% in a single day, the allowance replenishes daily, offering greater flexibility. Claude Code can also be installed via the extensions marketplace (search for “Claude”); it provides an initial, one-time quota, after which you must pay for further usage. Be aware that Claude Code is not particularly user-friendly for those in China, as accounts are frequently banned.