
What is the most painful thing about the age of AI?
It's watching the extreme voices on both sides speak up every time, and seeing their views continuously reinforce each other.
I hold both affection and hatred for AI.
The affection comes from the part that shattered my worldview, and the hatred from the part that eliminated my work.
The affection is greater.
The truth is, the worldview of a peripheral Korean programmer is very narrow. I can't speak for people programming inside the United States, but the world of Korean programming is built on a kind of dialect.
Korean AWS documentation has less content than its American counterpart, contains mistranslations, and carries some outright errors. On top of that, the books that actually get imported and become popular tend to be around 5 to 10 years old, so the knowledge stays stuck in the past. The Korean government framework, too, is essentially built on what are now considered modern anti-patterns.
Before I encountered AI, I was someone whose knowledge was composed of decidedly outdated ideas.
In that sense, AI is what shattered my worldview. The realization that the code and design patterns I had been using were outdated, and that things I considered concise could actually be anti-patterns, is exactly what I mean.
Learning to program in Korea used to cost a lot of money.
For starters, if you didn't live in Seoul, you had to cover Seoul's housing costs, and on top of that you had to spend over one million won per month on something called a bootcamp. There were bootcamps that offered financial support, of course, but as everyone knows, those large-company-led bootcamps prioritized hiring based on academic background and prior experience, making it very difficult for people starting from the bottom to actually study.
And all of it was in Seoul.
In practice, people outside Seoul simply couldn't afford the cost of learning. They would go there to learn how to use things like Microservices Architecture and Kubernetes, but looking back now, Korean corporations had essentially outsourced education for their own benefit and systematized it into the bootcamp model.
While overseas tech giants were producing documentation and spreading high-quality free study materials, Korean conglomerates were effectively presenting a system that isn't sustainably accessible. And honestly, that's understandable. In Korea, a company's business card functions as a measure of human worth, and placing a financial barrier in front of the hurdle to obtain that business card makes it very convenient for privilege to be passed down.
Looking back now, all of those skills were ultimately about getting into the companies known for good compensation.
After much deliberation, I accepted that I couldn't reach that level of treatment, joined a mid-tier Systems Integration firm, got scammed there, and have been building my career since. Luckily, in the provinces, being a step behind on technology trends was still enough to get by. And luckily, I landed a contract with a certain company, then used that track record to land another, and that's how the work kept going.
Of course, after AI came along, that kind of work decreased significantly. That said, behind the "hatred" that AI eliminated my work lies an admission I had to make: a good portion of what I was doing in the past had amounted to translator-level work, memorizing and assembling the "how-to" of outdated Korean-style frameworks.
Looking back now, I'd even call it shoddy. Taking people who hadn't yet formed a programming mental model, teaching them how to use Kubernetes, having them memorize design patterns without knowing whether any of it was right, and now teaching them how to use Langraph instead, I find myself wondering whether just teaching the usage of whatever framework happens to be trending is really the right approach.
I have a certain aversion to those instructors who say AI can't teach you anything. Honestly, they're right that it can't, in a sense, but reading the framework docs and actually building things with AI while learning a framework is a far more effective way to study. For the underlying principles, you can look up papers or seek out well-regarded foreign books on the subject. Even that used to be inaccessible to Korean learners, but these days AI translation makes it much easier.
On top of that, most of the communities I sought out were filled with people who had no interest in teaching, only in lecturing, telling you to read a book or check the manual first. That's understandable in retrospect, because indie app developers are essentially self-employed, and asking for technical knowledge in those spaces is essentially asking someone to give away their trade secrets.
Looked at that way, when I reflect on how expensive learning used to be, I genuinely feel a sense of relief about AI. Learning back then came bundled with unwanted social friction and enormous costs, but now you can just open a paper, open the docs, and study alongside AI.
On the other hand, people who were part of programming communities and could communicate with others may grieve watching those communities fall apart, but the world I experienced wasn't really like that, so I actually felt a sense of relief.
And East Asian media is fundamentally on the side of corporations, putting out only AI-friendly coverage that benefits their domestic semiconductor companies, which makes this all the more true.
Actually, it's not just me; the older programmers around me say the same thing, that AI is better than they are. But their skills, and this is hard to say, are stuck 20 years in the past. Those who keep their skills current are rare, so there's also a tendency to trust AI-generated code blindly.
I probably would have ended up the same way. It's common for the skills of provincial developers who have no outside connection to grow stale.
So I feel both affection and hatred toward AI: hatred because it's shrinking my work, and affection because it makes keeping my skills current so much easier.
The American friends I occasionally keep in touch with, by contrast, mostly tend to strongly resent AI. It destroys their communities and steals their knowledge. And what they say is true as well.
Seeing all of this, I'm always reminded that the same thing can look completely different depending on where you stand and where you belong. For my American friends, AI is a monster that destroys communities, but for me, a peripheral developer outside Seoul, AI is the only window that lets me connect with the world without an expensive tuition fee.
These days I'm actually transitioning my occupation to "someone who talks with Claude and GPT." I don't know how much longer my programming skills will hold up. But at the very least, the world has become one where I no longer have to pay a premium to learn an outdated dialect, so what does it matter.