[The Collapsed Coding Wall]⑤Will Vibe Coding Take Away Jobs?... "High-Level Developers Will Be in the Spotlight"

Vibe Coding as an Assistant for Developers
Superior to Humans in Repetitive Tasks like Code Generation
"Developers Who Can Grasp Structure and Revenue Models Are Needed"

Editor's NoteWe are now in the era of 'vibe coding,' where simply explaining your idea verbally allows artificial intelligence (AI) to write code for you. This marks a dramatic shift from the last decade, when coding fever swept the nation as developer jobs soared in popularity. Now, anyone can code, but landing a job in the field has become more difficult and security concerns have escalated. How is this rapid shift in the development paradigm changing our everyday lives? The Asia Business Daily takes a closer look at the frontlines of industry transformation brought about by vibe coding.

"Just over a year since its debut, vibe coding has moved beyond individual developers and is now being actively adopted at the organizational level."

Hyunseo Jung (pseudonym, 31, male), a developer with three years of experience, works at a company that subscribes to the enterprise plan for the vibe coding tool Cursor and uses it for development projects. Jung's way of working has changed dramatically compared to a year ago. He explained, "Developers handle the overall design and planning, while Cursor generates the necessary general code."

When vibe coding first emerged in February of last year, many developers were astonished by its performance. However, reactions within the industry were sharply divided. Some welcomed the arrival of an era where coding skills were no longer a necessity, while others expressed concern that bugs and security issues made it too risky for use in professional enterprise projects.

AI as an Assistant for Developers

A developer at O’Park is using an AI agent to conduct both user experience responsiveness testing and functionality testing simultaneously. Hye-rin Kim, CEO of O’Park, said that the AI agent demonstrates the work capacity of about four developers, enabling simultaneous work on other projects. Photo by Dongju Yoon

Vibe coding has already established itself as an assistant to corporate software developers, boosting development productivity. Typically, developers go through a seven-stage software development life cycle: planning, requirements analysis, design, development, testing, deployment, and operation/maintenance. AI mainly handles repetitive and standardized tasks during the development and testing phases, such as drafting code, organizing repetitive logic, and supporting test code, while developers focus on requirements analysis, exception handling, and operational management, which are critical to service development. Jung commented, "AI takes care of the 'miscellaneous tasks' that are additionally required for objectives like code revisions and bug fixes. In the past, making a single commit per day (the process of applying changes to the development management system) was the norm, but now, dozens of commits a day are possible."

Jehoon Lim (pseudonym, 36, male), a web developer with five years of experience, said that after introducing vibe coding, he was able to do the work of three people on his own, speeding up development by two to three times. Lim explained, "Vibe coding writes about 30 to 40 percent of the code, and I focus on the detailed parts or reviewing internal regulations. I review the code generated by AI, select the version that best fits our internal systems, and check for any missing components." He added, "Tasks that require a responsible decision-maker, such as server infrastructure or database design, and log history management according to company rules, are still difficult to leave to AI."

As vibe coding has become a supplementary tool for developers, companies are hiring fewer new developers. Dongseung Baek (29, male), a systems developer in the financial sector with three years of experience, said, "I use AI when developing financial app screens or maintaining old code logic." He continued, "For small, repetitive tasks like screen design and functional testing, AI already outperforms humans, so instead of hiring new developers, we subscribe to the paid version of Claude using company funds."

No Complete Development Without Human Involvement... Developers Who Grasp Context in Design Are Preferred

Although the amount of code produced by AI is increasing, developers do not fully trust vibe coding. This is because AI-generated code may contain security vulnerabilities, and AI still has limitations in understanding the full structure of large-scale projects. Baek explained, "AI still lacks the ability to fully grasp and design large-scale projects. Typically, developers take charge of the overall design, while partial tasks are delegated to AI."

Developers are no longer just coders; they are now planners and designers of IT services. As a result, since the beginning of this year, Silicon Valley companies have increasingly sought out brilliant and unconventional developers. The ability to identify and fix flaws in AI-generated code and to understand the overall context and architecture of a project has become more important.

Experts predict that only developers with high-level design skills will survive in the future. Choi Byungho, professor at Korea University Human Inspired AI Research Institute, said, "Coding AI will evolve to perform multi-step reasoning, identify and fix code defects, and even grasp the architecture of a design. Planning skills based on domain knowledge such as industry know-how and regulation will be crucial for setting the direction of IT systems."

Jaesung Lee, professor of AI at Chung-Ang University, commented, "Direct programming—writing code by hand—will gradually disappear. Rather than developers who rely on apps to find logic and problems or look up technical documentation, companies will need senior developers with diverse experiences and expertise." He added, "Experience and a spirit of exploration are now essential, enabling developers to handle programming even in unfamiliar situations."

Jinho Park, professor of Computer Science and AI at Dongguk University, analyzed, "AI has already internalized source code and coding experiences that enable high performance, and coding automation has become part of everyday life. There will be more demand for developers who use AI effectively rather than just technicians or beginners." He added, "Only the coding stage of the software development life cycle has been automated; requirements analysis, design, testing, and maintenance still require human developers. To use AI effectively, one must have logical thinking and background knowledge to understand the overall structure, direction, and business models of development projects."

Related articles: [The Collapse of the Coding Barrier]

①"Can You Get Some Work Done While I'm at Lunch?"... 'Vibe Coding' Changes the Way We Work

②Two Hours of Vibe Coding... Even a Reporter Can Build an Investment Profile Analysis Site

③"Job Seekers Out, Office Workers In"... How Vibe Coding Is Changing Education

④"Just Say 'Find the Weakness and Infiltrate'"... Hacking Has Become Child's Play

This content was produced with the assistance of AI translation services.