Interview with Kim Ki-beom, CEO of TenX Builders
“AI widens developer productivity gap”
“In order to survive, develop command and verification skills”
The changes in the work brought about by generative artificial intelligence (AI) technology are only at different speeds and are taking place in all jobs. Go players who suffered the AlphaGo shock first started, and software engineers called developers have been directly hit since last year.
Big Tech’s in-house code is used through AI coding tools such as Claude Code, and developers who do not use AI coding tools have become impossible to find.
If so, the question remains, ‘What should developers do?’
“Now all engineers can play the role of existing senior engineers,” said Kim Ki-beom, CEO of TenX Builders, who started as a software engineer at Google in 2013 and has had a decent career. “The ability to set the overall direction of the system and plan, instruct AI to work, and review the results to increase the completeness of all engineers are now the basic roles of all engineers.”
It also diagnosed that AI technology does not level the developer’s skills, but rather serves as an amplifier for the gap. The same thing is happening in software engineering, just as the gap between the top articles and the rest of the articles, which embody AI in the most depth in the world of go, has widened more than ever before in history, he said.
Kim, who has worked at Google for nine years and led the tech lead, started an open-source start-up called 10x.builders this year after serving as co-founder and chief technology officer of financial AI start-up Terminal X.
In the introduction of the company, he wrote, “In an environment where we don’t know how much more time we have left to unleash the capabilities of pure engineers in the development of AI, I want to spend the rest of my career as an engineer most meaningful, rewarding, and fun.” The following is a question and answer with CEO Kim.
All engineers lead and validate AI,
Productivity Differences between engineers will grow further in the future
-How is the role of engineers changing in the AI era.
Rather than having a new role, it is more like the role of an existing senior engineer now played by all engineers. Now, AI has replaced all the implementations of front-line developers. In other words, all engineers have become “one-person senior engineers” who lead their own AI team members.
The competencies required of senior engineers in the past are now important to everyone. The ability to set the overall direction of the system and plan, as well as to efficiently direct AI to work and critically review the results to increase the degree of completion, are now the basic roles of all engineers.
A similar analysis came from Antropic’s 2026 trends report. Detailed implementations such as code writing and debugging are passed on to AI, and the role of engineers is changing to an orchestra that directs and coordinates architectural design and various AI agents.
Software engineering has been the first to face changes, but only at different speeds will happen in every field.
-Is there anything that AI can’t replace in the work of an engineer.
The stage of discussing ‘replaceable or not’ has passed. Although it has been replaced a lot now, the pace of development over the past few years is more evident.
In 2023, prototyping or algorithmic problem solving in a single file unit began at a high level, and in 2024, the entire context of the large code base was grasped and real-world functions were implemented in multi-file units.
In 2025, we have reached the stage of producing performance that is comparable to that of junior engineers. As of 2026, it is already better than 99% of developers. Considering this speed, even if we take a conservative approach, AI will be able to perform all the tasks that developers do that are not far away by 2027.
What AI lacks most at this point in time is the continuous evaluation and direction setting of “what and how to make it.” This is not a lack of intelligence of AI itself, but rather because it is still incomplete to implement AI systems that maintain consistent long-term directions and allow AI to coordinate several other AIs. However, I think this will also be resolved in the near future.
-Will the gap between a small number of smart developers and other developers widen.
Even before the advent of AI, software engineering was already an area where the gap between excellent engineers and average was already extremely large. Steve Jobs also said, “In general, the difference between the best and the average is at most twice as big, but in software, the difference between the best and the average is 50 times and 100 times.”
But now, as coding agents develop, this gap is widening even more.
First, it is because the basic baseline itself has risen to the AI level. Developers who have less coding capacity than AI or lack the ability to verify AI’s results are no longer needed.
Second, the productivity gap widens exponentially depending on how much of the entire development process can be delegated and controlled to AI. There is a fourfold difference in productivity between those who automate 80% of the development process (20% left over) and those who automate 95% (5% left over). It’s 20 times different from someone who automates up to 99%.
After all, AI is not a tool to level skills, but it is acting as an “amplifier of the gap” by leveraging engineers who are good at it. This is also the first change in the Go world. After AlphaGo, some predicted that everyone’s skills would be upgraded as AI researched, but in reality, the gap between the top players and the rest, who embodied AI in the most depth, such as Shin Jin-seo, has widened more than ever before in history.
-As an engineer, don’t you feel more anxious or unpredictable about the future.
If you were an employee of a company, you would be anxious. However, it seems that there are still many things that humans can do in terms of insights and capabilities to handle AI well. Of course, this will be getting less and less, and if so, you may get anxious thinking that you can’t contribute anything, but it doesn’t seem to be at that point yet.
And since I’m a entrepreneur, I’m in the opposite position. Being in a position to leverage AI to create value, (technology development) is more welcome. Although it is not the most important factor, AI also influenced the choice to start a startup. There is an expectation that I will be able to do everything I want to do until the next few years, and for now, the expectation seems to be much higher. Moreover, there have been a number of rumors and official comments that frontier AI companies have recently begun to enter recursive self-improvement (RSI), so further development of AI is expected to accelerate in the future.
세상에 변화 일으킬 의미 있는 기술 개발 목표 “韓 엔지니어 실력 뒤지지 않아…”Dream big”
-What is a meaningful skill as an engineer.
Meaning is very subjective and varies from person to person. For me, the most meaningful thing is to create a technology that does not exist, to be widely used, and to be different before and after it exists.
In fact, most developments in the software industry do not create anything completely new. Most of the work involves reassembling existing technologies and libraries to meet the requirements of each product. On the other hand, technologies such as Git, Docker, Kubernetes, and Large Language Model (LLM) have changed the fundamental way engineers around the world make software. I also aim to make a difference in the world, including the open-source database startups that I just started.
-What is TENX Builders, which is a startup that develops.
It is an open-source database startup. Among the databases, there is a field called a cache database. A cache is a database that temporarily stores data that is often used for traffic. I usually used RAM a lot because I need fast speed, but now the price of RAM is very expensive. So we are designing algorithms that use solid state drives (SSD) as a cache database and perform like RAM.
-It is unique that the initial recruitment was centered on new employees.
What I felt while meeting and hiring numerous engineers at Google and start-ups is that the speed and judgment of learning are much more important than annual leave. In software engineering, where new technologies come out every day, it makes a difference not by what and how much you know in the past, but by how quickly you can absorb new things and make good decisions. So I think an excellent newcomer is much better than a normal 10 year engineer.
-How competitive are Korean engineers in the global market.
I don’t think Korean engineers are inferior at all. Nevertheless, the reason for the wide difference in results is mindset.
In Silicon Valley, dreams are wildly, ridiculously loud. It was not natural at the time that Amazon, an online bookstore, made a cloud that changed IT infrastructure around the world, re-used rockets in the days when no one believed, or went all in to create AGI.
Google founders Larry Page and Sergey Brin emphasized “10x Thinking” at every company-wide meeting. If you announce that “the goal is 100 million users,” you are constantly asking back, “Why not 1 billion won?” which is 10 times that. It’s not just Google’s story, it’s the whole Silicon Valley story. It is paradoxically easier to hold a dream big. In order to make a big impact, it is necessary to gather talent, and the bigger the dream, the better the person can be recruited. There is also no competition because fewer people think so.
In the end, it is not the limit of ability, but the difference between the limit and the size of imagination we drew ourselves.
-From Google to start-up CTO and start-up. What is your career goal.
When I look back on my career, the most regrettable period was when no matter how good the team played a role that others envied, I did not change much before and after spending that time. In the end, it wasn’t the team or the role that was left, it was myself.
So, rather than repeating what I’m good at comfortably, I try to do many things I don’t know and do new things. The reason why I co-founded Terminal X with Big Tech was because I wanted to do challenging things that were difficult to do at Big Tech, such as team building, product direction setting, and fundraising.
And the second criterion that has become more and more important as my career has built is “Can I work meaningfully and joyfully every day?” It is worthwhile in itself to have something meaningful and enjoyable to do every day, but in the end, I think it is the commonality and driving force of people who make a big impact on the world.
