Creating Software with No Coding Experience: A Week with Vibe Coding

Discover how a business student created three software applications in a week using Vibe Coding, a method that simplifies programming with AI.

Introduction

Last week, I was just a business student who couldn’t even write “Hello World”. Faced with a screen full of code, I felt overwhelmed, like it was a foreign language. However, within a week, I successfully created three functional software applications: a resume generator, an image filter tool, and a mini-program for automatic resume screening.

In this article, I want to share my experiences from that week and the methods that made it possible for an ordinary person to code. If you’ve ever hesitated to dive into programming or thought, “If only I could create a tool myself,” then keep reading.

What is Vibe Coding?

“Vibe Coding” translates to “coding by feel.” It may sound abstract, but its core concept is simple and almost humorous.

Traditionally, learning to code involves mastering syntax, data structures, algorithms, frameworks, and databases. By the time you write your first useful program, months may have passed, and many give up on this long and tedious journey.

Vibe Coding is different. Its central idea is: You don’t need to know how to write code; you just need to know what you want to achieve. Imagine yourself as a director, with AI as your versatile crew of cameramen, lighting technicians, and prop masters. You don’t need to know how to operate the camera or set up the lights; you just tell your crew, “I want a rainy night scene with the protagonist standing on a street corner, and the mood should be somewhat melancholic.” Then, the crew brings your vision to life.

Vibe Coding operates on this principle. You provide the “vibe”—your ideas, needs, and scene descriptions—while AI transforms these vague concepts into concrete code. You are no longer just a student memorizing syntax; you are a product manager, a designer, a decision-maker.

In simple terms: Previously, humans adapted to machines; now, machines adapt to humans.

How Can Vibe Coding Enhance Our Lives and Work?

You might still be skeptical: can this really work? Is it practical? Let me share some real-life scenarios.

Scenario 1: Boosting Workplace Efficiency

During my internship at a consulting firm, I had to process dozens of resumes weekly. HR provided them in PDF or Word formats, and I had to open each one, extract names, schools, majors, and internship experiences, then fill them into an Excel spreadsheet. This tedious task consumed an entire afternoon.

Later, I used Vibe Coding to create a “resume screener.” I simply needed to drop a bunch of resume files into a folder and tell AiPy, “Extract the key information from these resumes and organize it into a table, filtering out those that don’t meet our requirements.” A few minutes later, a neatly organized Excel file was generated, highlighting candidate match percentages.

What used to take an entire afternoon now took just five minutes. My boss looked at me differently, thinking I had taken an efficiency course.

Scenario 2: Side Projects

A friend of mine runs a small photography studio. He was looking for a tool to batch apply filters to client photos, but existing software was either too expensive or overly complicated. Eventually, he created a small program using Vibe Coding that tailored five filter styles to his portrait photography.

Now, he not only uses it himself but also sells this tool to other small studios. He remarked, “I used to think software development was for programmers, but now I realize that we know our needs best.”

Scenario 3: Bringing Personal Projects to Life

Have you ever thought, “If only there were a tool that could…”? It could be a tool to automatically organize desktop files, track spending habits, or create an interactive webpage for an anniversary.

Previously, these ideas remained just that—ideas—because the cost of realization was too high. But with Vibe Coding, the distance from concept to product could be just a cup of coffee away.

This is the essence of Vibe Coding: it lowers the barrier to programming from “professional skill” to “expressive capability.” As long as you can clearly describe what you want, you can create.

Vibe Coding is Simple: A Vague Idea + The Right AI

I know some of you might still be doubtful: “This sounds great, but can I really do it?” I can responsibly tell you: yes, you can. And it’s even simpler than you think.

Step 1: Have a Theme, Even if Vague

You don’t need to have every detail of functionality figured out. For example, if you want to create a “budgeting tool,” you only need to know: “I want to record how much I spend daily and see a summary at the end of the month.” That’s enough.

The rest is up to AI.

Step 2: Find the Right AI Tool

There are many AI tools that claim to write code, but most are merely “code completion tools” that require very precise instructions. This is still too high a barrier for beginners.

In my week of experimentation, AiPy stood out as a tool that truly delivers on “you describe the need, it provides the product.”

Why AiPy?

  1. Strong comprehension: You can describe your needs in plain language, and it understands. There’s no need to learn “prompt engineering”; just communicate like you would with a colleague.
  2. Complete code generation: It doesn’t just write a function; it structures the entire project, including interface and logic.
  3. Improvement and execution: This is crucial. Many tools can only generate code, but if you can’t debug or deploy it, the code is useless. AiPy can run it directly, and if you’re not satisfied with the results, just say, “Make this change,” and it will adjust.

Step 3: Let AI Generate Code in Bulk and Quickly

This step is the most enjoyable. You simply sit at your computer and tell AiPy your ideas as if you were chatting. For instance, when I was creating the “image filter” tool, I said: “I want a simple tool that allows users to upload an image and choose from several filter effects, like black and white, retro, cool tones, and warm tones. The left side should show the original image, and the right side should display the filtered effect. Users should be able to download the final image.”

Then I went to grab a drink, and when I returned, the interface was already created. Although the first version had some issues—like small buttons and insufficient filter effects—I simply told AiPy, “Make the buttons larger and enhance the filter preview effects,” and it made the adjustments immediately.

Throughout this process, I didn’t write a single line of code, didn’t look up any syntax, and didn’t even open any programming software. I was merely “specifying requirements” and “accepting results.”

This is the magic of Vibe Coding: you focus on “what to do” rather than “how to do it.”

My Experience: A Business Student’s Week Challenge

Now that I’ve covered the theory, let’s get into the practical details. Here’s a breakdown of how I accomplished this in a week, including the pitfalls I encountered and the three outcomes I achieved.

I am a junior at a prestigious business school, minoring in marketing. My programming experience is zero. I once took a Python introductory course and barely passed the final exam, forgetting everything afterward.

The reason I wanted to try Vibe Coding was that I noticed my peers who could code had a clear advantage in internships and projects. They could build data analysis models, create automation scripts, and even take on freelance work. Meanwhile, I felt I had nothing to offer beyond basic Excel skills. I was determined to change that. However, I couldn’t muster the motivation to pick up programming textbooks and start from variables and loops again—the time investment seemed too high.

Then I heard about Vibe Coding. I decided to give myself a week to see if this method was truly effective.

Day 1: Exploration and First Outcome—Resume Generator

The first day was quite confusing. I registered for AiPy and stared at the input box, unsure of what to ask it to help me create.

Later, I remembered that the fall recruitment season was approaching, and many of my classmates were revising their resumes. However, many people (including myself) struggle with formatting, often producing resumes that are either too flashy or too simplistic.

So, I tried inputting:

“Help me create a resume generator where users can fill in their basic information, education background, internship experiences, and project experiences, then choose a template to export as a PDF. The template should be simple and professional, suitable for fresh graduates.”

About ten minutes later, a web-based resume generator appeared. The left side had a form, and the right side displayed a real-time preview. I filled in my details, selected a template, and clicked “Export PDF”—and it actually generated a beautifully formatted resume file!

Of course, the first version had issues: sometimes the exported PDF would have garbled text, and there were only two templates available. But I wasn’t worried; I continued to communicate with AiPy, saying, “Fix the PDF export issue and add more template options, ideally allowing for custom font colors.” By that afternoon, those problems were resolved. My first software was born.

Day 3: Advanced Attempt—Image Filter Tool

With the success of the first day, I felt bolder. My girlfriend loves taking photos but often complains that the built-in filters on her phone aren’t appealing enough. I thought, why not try creating an image processing tool?

This time, the requirements were a bit more complex:

“Create an image editing tool where users can upload an image and choose from various filter effects, including black and white, retro, cool, warm, and high contrast. After uploading, the left side should display the original image, and the right side should show the processed effect. Users should be able to adjust the filter intensity and download the final image once satisfied.”

This time, AiPy generated a significantly larger amount of code since it needed to handle image data. During the first run, I encountered an error indicating “missing dependencies.” I felt a bit anxious, but recalling my previous experience, I copied the error message to AiPy. It informed me that I needed to install an image processing library, which I did with a single command.

This process taught me: don’t fear errors; they can be your best teacher.

By 8 PM, the filter tool was officially launched. I tested it with my girlfriend’s photos, and the results were surprisingly better than some paid apps. Her delighted expression was one of the most fulfilling moments of my week.

Day 5: Practical Application—Resume Screener

The final project was based on real needs when I returned to my internship company.

As mentioned, we had to process a large number of resumes weekly. Manual screening was time-consuming and prone to errors. I casually mentioned this to my supervisor, who said, “If you could automate it, that would be amazing.”

So, I decided to tackle it.

This time, the requirements were the most complex:

“Help me create a resume screening tool that can batch read all PDF and Word resumes in a folder, extract key information such as names, schools, majors, degrees, internship company names, and internship durations, and organize it into an Excel table. Then, based on the criteria I set, automatically filter out candidates who meet the requirements and indicate their match percentage.”

This project took me two days. Not because it was difficult, but because it required repeated adjustments. For instance, some resumes had inconsistent formats, leading to inaccurate information extraction; the matching algorithm needed weight adjustments.

But every time I encountered an issue, I simply described the situation to AiPy, and it helped me modify the code and adjust the logic. By the evening of the fifth day, the screener was officially in use.

The next day, I demonstrated the tool to my supervisor, and she was astonished, asking if I had hired an external tech team to develop it. I told her no; I created it using an AI tool myself.

One Week Summary

In just seven days, I produced three outcomes. It sounds unbelievable, but it truly happened.

Reflecting on the week, my biggest takeaway is: The essence of programming is not writing code but solving problems. I used to think I had to master syntax to solve problems, but now I realize that tools are just means; the goal is what matters.

Of course, I didn’t learn nothing. Throughout my use of AiPy, I subconsciously grasped some basic programming concepts, such as front-end, back-end, and APIs. But these concepts were learned through practice, which is far more effective than rote memorization.

More importantly, I built a sense of confidence: I am no longer afraid of technology. Where I once felt overwhelmed by code, I now think, “Can this function be implemented with code? If so, how should I describe it to AI?”

This shift in mindset is more valuable than creating three pieces of software.

Tips for Beginners

If you’ve read this article and want to try Vibe Coding, here are a few tips:

  1. Start simple; don’t aim for a big platform right away. My first project was just a resume generator with a single function and clear requirements. Don’t start by trying to create the next WeChat or Taobao; that will only lead to frustration.

  2. Be as specific as possible when describing your needs. While it’s called “vague ideas,” you should still strive for clarity when communicating with AI. For instance, instead of saying, “Create a good-looking interface,” specify, “The interface should be simple, with blue as the main color, buttons on the right side, and using the Microsoft YaHei font.”

  3. Don’t fear mistakes; they are the best teachers. I encountered five errors while creating the image tool for the first time. But each error helped me understand how the tool works better. Share the error messages with AI, and it will help you resolve them.

  4. Embrace iterative thinking. Don’t expect to create a perfect product in one go. Start with a usable version and continuously improve it based on user feedback. Each of my three software projects went through four to five iterations before finalizing.

  5. Choosing the right tool is crucial. I primarily used AiPy during my week; its understanding of Chinese and code execution capabilities are quite strong. Of course, you can also explore other tools to find the one that suits you best. If you’re worried about token limits, use the invite code c8W3 for an additional two million tokens.

Conclusion: Everyone Can Be a Creator

As I wrap up, I’m reminded of a saying: “When the wind blows, even pigs can fly.”

Is Vibe Coding the next big trend? I can’t say for sure. But it certainly presents an opportunity for ordinary people to engage in technological creation.

In the past, technology was a privilege for a select few. You had to spend years learning, go through layers of screening to enter major companies, and toil away in the background as a cog in a large machine. But now, things are changing. AI has lowered the barriers to technology, giving everyone a “magic pen.” You don’t need to understand the principles of magic; you just need to know what you want to draw.

I’m not advocating for the “programming is useless” argument, nor am I saying professional programmers will be replaced. On the contrary, I believe the future will be a world where professional developers and amateur creators coexist. Professionals build the infrastructure, while we “amateurs” fill in the niche needs that large companies overlook with our creativity and understanding of life.

This week’s experience taught me that our limitations are not our abilities but our imagination.

If you feel even a little inspired by this article, why not open your computer now, find a suitable AI tool, and try to turn one of your small ideas into reality? Don’t strive for perfection, don’t worry about others’ opinions; just seek that feeling of accomplishment that comes with “I did it.”

You might discover that you, too, can be a creator. And this is just the beginning.

Was this helpful?

Likes and saves are stored in your browser on this device only (local storage) and are not uploaded to our servers.

Comments

Discussion is powered by Giscus (GitHub Discussions). Add repo, repoID, category, and categoryID under [params.comments.giscus] in hugo.toml using the values from the Giscus setup tool.