How AI Coding Tools Help Ordinary People Unleash Their Creativity: Reflections Based on Real-World Experience

For the past six months, I have been deeply interested in AI programming. I have experimented with various tools and models. Whether in my professional work or my independent side projects, I have continuously used these projects to learn AI programming, exploring its best practices and functional boundaries.

For a long time, both my fellow programmers and I viewed AI programming tools merely as a means to improve development efficiency. The slogan "Use AI to increase your development efficiency by 10x" is everywhere. AI is a capability amplifier; it can multiply the output of an experienced programmer. As for ordinary people who don't know how to code, they have no baseline capability to amplify.

However, a recent event changed my perspective. My view of AI programming has begun to shift from "efficiency enhancement" to "creativity enhancement."

I Improved Communication Efficiency with My Product Manager by 300%

As a programmer, I naturally deal with product managers a lot in my work. The protagonist of this story is a product manager I collaborate with—let's call her Emily.

Like many typical product managers, Emily doesn't know how to code, but she is responsible for defining product requirements and features, telling the programmers what the product should look like.

In the usual development process, Emily often has to perform a tedious task: once she conceives a product feature in her mind, she needs to express it through wireframes (product screenshots + Photoshop + text descriptions) before handing them over to the UI designer for visual drafts. This communication process—whether through drawing, writing, or verbal exchange—is extremely time-consuming and laborious.

One day, I had an idea for a new product feature and wanted to discuss it with Emily. I faced the same hurdle: how to convey my thoughts to her clearly and accurately.

On a whim, I chose an AI programming tool to help me draw a prototype.

The tool is called v0. Those familiar with it know it is a tool for writing frontend code. I knew about it because I had used it in my independent projects. But this time, my goal wasn't to write frontend code for a project, but to use frontend code to render a product prototype.

I entered a prompt describing the requirements into v0, and it quickly generated the first version of the visual. After making a few minor adjustments, I had a prototype ready and sent to Emily in just 20 minutes. When Emily saw the image, she instantly understood my idea.

I knew that without this visual, a verbal discussion between us would have taken at least an hour. Just like that, I boosted my work efficiency by 300%!

And all of this required nothing more than entering a few simple prompts.

The following video demonstrates the process of using v0 (actual company business information has been hidden; only the example workflow is shown):

A Bridge Between Technical and Non-Technical

At first, I was simply smug about my "broad knowledge," and Emily praised me for introducing her to a useful tool. It wasn't until a certain day this month that my perspective truly shifted.

Since April, I have been intermittently sharing my insights and experiences regarding AI programming, independent development, and AI Agents on various social platforms. I write these things partly to practice the Feynman Technique and partly to leverage my strength: explaining knowledge clearly and understandably.

As AI programming tools continue to grow in power, even ordinary people who cannot code can build simple small projects. I began to wonder if I could teach non-programmers to master AI programming through my explanations. At that moment, I realized Emily was the first "ordinary person" I had successfully taught.

This is a fragmented world. On one hand, AI programming tools are flourishing, helping countless technical people free their hands and save time. On the other hand, there are perhaps millions of "Emilys" waiting for an AI tool to boost their efficiency. It’s not that they are stuck in their ways; they simply don't know how to find the tools that suit them. Most AI products have not yet developed a truly great ecosystem or user-friendly interfaces, leaving a massive number of users locked outside.

This is where someone needs to act as a bridge between the technical and non-technical worlds, utilizing the power of AI programming tools to help non-technical personnel express and realize their ideas more efficiently. My advantage in "explaining things clearly" might play a larger role here. Moving forward, I need to uncover more use cases and think about ways to improve efficiency for non-technical people.

Unlocking the Infinite Potential of AI Programming

2025 is being called the inaugural year of AI Agents. In this wave of AI Agents, AI programming is the largest crest.

This is no coincidence; it is determined by the characteristics of Large Language Models (LLMs). The current training capabilities of LLMs rely on Reinforcement Learning (RL) for advancement. Because RL requires an immense number of iteration cycles, it heavily depends on clear, quantifiable reward feedback mechanisms. Tasks with objective standards, such as mathematical reasoning and programming, are naturally easier to train. Since LLMs excel at programming, it naturally follows that AI programming products have become popular.

Consequently, AI Agents have made massive strides in the relatively technical field of programming, leading to an interesting phenomenon: the potential of AI programming has actually barely been scratched.

First, as a technical field, the promotion and popularization of AI programming are very limited.

Currently, AI programming is mostly popular within developer circles. Ordinary people find it difficult to take an interest in something that "looks professional" at first glance. Thus, even though tools like v0 can be used by anyone without a barrier to entry (requiring zero coding ability), many people have never heard of them and don't know they can be integrated into their workflows.

This is also why AI programming is far less famous than AI drawing: it is easy for people to imagine what AI drawing can do for them, but they are completely unaware of what AI programming can achieve.

Second, programming itself is a highly creative act. It is not just a job for programmers; it is a means for many people to realize various ideas.

For example, in Emily's case, she used v0 to generate frontend code, but her goal wasn't to write code—it was to use the code to draw the prototype she wanted. In this process, writing code was the means, and drawing the prototype was the end.

We can easily imagine more similar scenarios: an entrepreneur using AI to generate a product prototype, a marketer using AI to generate a campaign proposal, a teacher using AI to build an interactive learning tool for students... Writing code is just the method; completing work efficiently and producing creative output quickly is the goal.

Combining these two points leads to a startling conclusion: the creativity of AI programming can help people in countless fields, but it is currently unknown to the majority. This means there is infinite potential in AI programming waiting to be tapped.

Most AI programming products today are busy launching new features and chasing ARR, neglecting how to make these capabilities accessible to the masses. AI programming capabilities are still advancing, but who knows when its true potential will be fully released.

Shifting from Technical Ability to Problem-Solving Ability

Lately, as AI programming becomes more powerful, the narrative that "programmers will be replaced by AI" has become rampant. However, I believe the real crisis for programmers is not being replaced by AI, but rather a lack of problem-solving ability.

Like many technological waves before it, Generative AI will eliminate some technical roles while creating a batch of new ones. The current difficulty in finding programming jobs is mainly due to decreased demand caused by the economic downturn, while many new demands of the AI era have not yet emerged. As a profession with technical barriers, programmers are not that easy to replace.

However, the real career crisis for the programmer community actually has nothing to do with AI; it lies in a stubborn and hidden mindset: over-pursuing technical skills rather than problem-solving skills.

"I know Java," "I have low-level CUDA experience," "I work on AI Infra"—too many people approach job hunting with this mindset. Granted, this is the norm for hiring in internet companies today; companies need a series of "screws," so engineers are forced to specialize in a specific technology to prove their worth. Even when facing the threat of AI, many people's first reaction is to find a niche tech stack that AI isn't good at. But when the environment changes and a certain technology is no longer useful, programmers who rely solely on that tech to find work struggle to articulate their value.

"I can decompose vague business requirements into executable features," "I can package complex tools into easy-to-use products," "I can get a system running as fast as possible with limited resources"—these are problem-solving abilities, and they are the true skills that help programmers survive cycles. Unfortunately, internet companies often overlook these, so programmers have become increasingly lazy about cultivating them.

This is the greatest inspiration Emily's story gave me: when "tech-only" thinking is at its peak, how do we find a path to cultivate our problem-solving abilities?

Being a "cross-disciplinary empowerer" might be one approach. The potential of AI programming tools has not yet been fully explored, and this is entirely the domain of technical people. By using your understanding of AI and programming to help more people build new workflows and create new productivity, you are not only performing a noble act of helping others but also gaining an excellent opportunity to exercise your problem-solving skills.

When my thinking is no longer confined to the technology itself, but instead treats programming as a means and a tool—leveraging imagination to move the technological lever—I will be able to create much greater value.