Old problems, the flamboyant beauty of blossoming flowers. (This captures the essence and poetic nature of the original.) Alternatively, a more literal translation could be: “Old ailments, beautiful eyes like blooming flowers.” However, the first option is generally preferred for its aesthetic quality.

For many years, I’ve focused on backend development, and recently started to explore AI programming while dipping my toes into some frontend-related content. However, during this period of tinkering, I gradually realized I was falling back into an old habit – being dazzled by shiny new things. I constantly try to use AI to create a frontend interface, but in reality, these attempts haven’t provided much practical benefit for my current work and are actually wasting my energy.

AI Use Cases

In small projects, AI tools can truly shine, particularly when writing independent functions with low coupling to the system and simple business logic. These tasks typically have clear inputs and outputs, and rely less on context – making them well-suited for the current capabilities of AI-assisted programming.

However, when facing complex system architectures or deep business logic, AI’s limitations begin to emerge. It may generate code that appears reasonable but is actually detached from the project’s real needs, or even introduce potential issues that are difficult to debug. In these scenarios, AI is best suited as an assistive tool rather than a fully autonomous code generator. We need to conduct rigorous review and testing of the generated code to ensure it meets actual requirements.

Errors and the Cost of Learning

While attempting to generate frontend code using AI, I encountered numerous challenges. As frontend development isn’t a domain I’m familiar with, troubleshooting often proved time-consuming and frustrating. Even after adjusting prompts to have the AI rewrite the code, it was difficult to avoid the appearance of some low-level errors. This iterative process not only wasted time but also highlighted that my current focus should be on backend business logic rather than groping around in unfamiliar territory.

Looking back at the project completed over the weekend, I’m even more convinced that focusing on backend development and user interaction logic, implementing functionality through a console, is the most efficient approach currently. Perhaps systematically learning frontend knowledge would be a better strategy when I have more time and energy.

Frontend Learning Plan

The frontend technology stack is complex and diverse, so it’s unrealistic to quickly master it. I plan to first choose a framework, such as Vue.js or React.js, and deeply learn its core concepts and usage methods. Only after becoming familiar with the fundamentals will I attempt to use AI to generate frontend code, which can effectively avoid errors and wasted time caused by unfamiliarity.

In short, the focus for this stage should be on backend development, steadily building my core skills. When the timing is right, I’ll then explore the combination of frontend and AI – potentially yielding greater rewards.