Why My Developer Colleague Moved to Sales

A senior developer at our company with 18 years of experience landed a great opportunity and is moving on. It might sound strange to call someone leaving a "great opportunity." But I know they pulled more all-nighters here than anyone and cared endlessly about the team, so wherever they went, I'm sure it would be somewhere better.

Anyway, the decision to leave was already made. What caught my attention when I first heard about it was that they weren't going into development. They were going into sales, specifically technical sales. They have plenty of experience, they speak logically and clearly, and they'd always said they were interested in management and business. Still, I was curious about the reasons, so we had a coffee chat before they left.

Below are the meeting notes Claude summarized for me.

1. A senior who wants to leave development behind

They had already figured that whatever their next career move was, it wouldn't be as a developer. They made up their mind at the end of last year. Not that long ago.

With a career like theirs, the usual paths are VP of Engineering, CTO, or running a research lab like they did before. Becoming an executive means letting go of development in a different way. Apart from all that, they said they'd grown tired of development itself.

The irony is that the trend right now is going the other way. The legends they look up to, who left because they were burned out, are coming back to hands-on work in the AI era. Vibe coding has made it possible for them to build things directly again.

2. Anyone can code, but not just anyone can build

This was the point they pressed hardest. In the AI era, developers aren't running out of work. It's heading toward a world where you have to be really good to survive.

With vibe coding, anyone can code now, but unless you're good at it, nobody can build a real product.

I asked if that meant competition had gotten fiercer. They said no. The core issue is that everyone has the illusion that programming is easy now. In the old days, development itself was the barrier, so you could even try selling a clumsy product. You can still sell something simple with just an idea. But building a proper product takes a huge amount of knowledge, and people no longer build up the knowledge they need. So they can't build anything solid. And yet they believe they can.

It's not only developers. Non-developers also assume that these days anyone can build things and AI does it all, so they keep turning out products that are a total mess. It's a kind of race to the bottom. They were tired of it. They said they've been saying the same thing since last year: in the AI era you have to study more and be smarter, and you can't just start building things blindly. That trend is accelerating, and they see it heading toward a world where only the real experts survive.

3. Not wanting to be a developer in Korea

As they see it, overseas it's actually an era where the better and smarter people survive. People overflowing with ideas or people who learn fast make it. But in Korea, people look at that and seem to think they can do it too. So if they're in Korea, they don't want to be a developer anymore. Overseas, they'd rather keep developing, but they don't want to compete in Korea while hearing that kind of talk.

What they do have, they said, is a lot of knowledge. With about 18 years of experience, they figured guiding others would play to their strengths. Becoming a manager or a consultant seemed better, so last year they were thinking of tech consulting or tech education as their next career. They wanted to do sales too, but sales doesn't give anyone a chance. Then a role opened up where they could do something worthwhile, so they went into sales.

We also talked about the mood in Korean society. It heats up fast, like a pot coming to a boil, and everyone gets swept up in it. There's a lot of content that's just posturing. They said that in the past, they would have gone their own way regardless. But they're tired now. They don't know how much longer they'll keep working, and they don't want to prove themselves all over again while people argue about who's right and who's wrong.

I've been proving myself this whole time, and I don't want to start proving myself all over again.

So they want to stay on the sidelines until it's settled. They might be wrong, so they'll watch until it becomes clear whether they were right or not. Just like the legends coming back, they can always go back too. And since they're joining a foreign company, if the Korean market doesn't shift toward the mood they want, they can give up on the Korean market and go abroad.

4. Not having a CS degree, and domain

They didn't major in computer science. I asked if not having a CS degree gave them a kind of freedom. They didn't think so. Why does it matter whether you studied CS? You don't use a single thing you learned in school at work anyway.

I also heard how they got into development. Originally they planned to do something else, making military supplies. But they liked development, and engineers absolutely had to know basic coding. They're an image processing developer, and image processing developers need to be able to turn research papers into code. For someone who loved coding, that was great practice. Along the way they got hooked on the mechanics of being a builder. Weaving together knowledge they'd studied and implementing it with their own hands was fun, like playing SimCity.

They've never focused on a particular domain. Each time, they followed whatever domain their company was focused on. They started in defense and went through fields like finance. People say it's an advantage to become an expert in one domain, but they never tied themselves to one.

This is where our views split. I think the job function matters, and the domain doesn't matter that much. You can't suddenly switch from marketing to design, but in my experience, if your role stays the same, you can move from securities to entertainment and it's just a learning curve. They said it's the same for developers. Going from a finance developer to a commerce developer is doable if you want it, but going from a finance developer to a finance product planner isn't easy. If your role stays the same, the domain carries over. They went a step further and said developers and non-developers are the same in this respect. Non-developers seem to think domain differences are huge, and they said they're actually curious why non-developers see it that way.

5. The legends are coming back to hands-on work

They recently shared what they were wrestling with over tea with an acquaintance, a US citizen who works at a foreign company. They were told that for the very same reason, people overseas who had stepped away are coming back. The legends they look up to, who must be in their fifties or sixties by now, are returning to hands-on work.

I asked why. It used to be that you had to go up into management to work on bigger things. Rather than doing the work yourself and having your knowledge recognized, the way to bigger projects was to coordinate and design things between people. But now you can get that done by directing AI. You can play the manager role and still develop directly, which isn't bad at all. You can even do it without managing humans, they said. And people respect that now. They think Korea would respect it too at that level. The problem is just that the mood here gets swept up too easily.

6. What they read and what they're into

I asked about their go-to sources. They don't really follow people, since they're not that into social media. Mostly they read Google articles, Google's recommended content and GeekNews (a Korean tech news aggregator, similar to Hacker News). They read GeekNews and the comments underneath, Reddit posts and comments, and if they try something and it works, they share it.

They used to watch a lot of YouTube, but not much anymore. Especially Korean YouTube. They said when they see content like "I used 10 agents," the person doesn't even come across as capable. Content keeps pouring out about using a bunch of agents, how much money someone made, who hit some monthly revenue the fastest, how they did this or that with multimodal. Every time the trend changes, the same kind of content floods in. They used to watch it because they felt they had to keep up, but they realized it sits way at the front of the Dunning Kruger curve in terms of depth. It's content for beginners. It might help beginners, but there's no reason to watch more than one, so they rarely watch YouTube now. They still watch some US content, but the English barrier keeps them from watching much. As it happens, that kind of content has cooled off too, and we talked about how a lot of people who were running a bunch of agents have now deleted them all.

I asked what they're into these days. In June 2026, they said, they're actually reading about psychology and mental models. Since they're switching to sales, they watch some sales videos too, but surprisingly few. They see the current era as a data fight, so they read technical material on managing and preprocessing data. And they've had one routine for a long time. They check in on what's happening in specific regions. They watch vlogs or videos from people who've actually been to places like India or the US to see how the economy and job market are doing. If they go somewhere, they want it to be where the opportunities are. They said even with year-end fireworks, if a place has fewer than last year, they take it as a sign the economy has gotten worse.

I asked which regions they're interested in. They still think about South America. But that's where you go if you're running full speed with passion. These days it's more and more Southeast Asia. As a place to start winding down the run, Southeast Asia seems right. They see it as a place where they can still keep their career going, with safety, cost of living and a lifestyle that fit. They're not abandoning Korea. They're thinking of going back and forth, and maybe moving for good when they're older.

That led to a conversation about perspective. Back when they lived in Korea, they thought there was plenty to do in Korea and that people live the same way everywhere. But after spending some time in the UK by chance, they found out that wasn't true. They realized how small their view had been. Places that are too intense are hard, so their conclusion was that Southeast Asia lets you break out of your bubble a bit while still having a safety net. Even so, they still want to keep breaking out of their bubble. In Korea, you only ever meet Koreans, and they want to meet people from around the world and see what opportunities they have. Korea is the most stable and everyone speaks the same language, so they stop making an effort, and since there's no need to make new friends, their view gets narrower.

7. Good developers and good communities

I asked whether everyone is as disillusioned as they are with that mood among Korean developers. They said the people who are really good, truly capable and using AI exactly where it fits, don't seem to hang around the noisy communities much. They're probably just quietly doing their work well. There are some attention seekers among good developers, but they said, carefully, that of the 10 people making noise, maybe one is the real deal, and the other nine have no real skill and are pretending to be good. Those are the people going down the influencer path.

So I asked how you judge whether someone is good at development. Being able to build something isn't what makes you good at development, they said. You have to build something the other person is satisfied with, something they can trust and use. It has to be stable and well finished. There are surprisingly many products that break one time in five but, hey, at least they got built. They agree with the startup mindset that taking action matters more than being perfect from the start. But the important thing there is that problems can happen, but the same problem must not happen twice. Plenty of people let it happen twice, three times, a hundred times, and brush it off with some fantasy like "this doesn't matter, we need to focus on what's more important," with no intention of fixing it. It's not that they can't fix it for lack of ability. They think of themselves as fast movers.

They were also clear on the difference between a good community and one with nothing to offer. They've been part of a lot of developer communities and still meet with two or so. The communities with nothing to offer are full of people who only talk about "here's what I did." "I've been learning this AI stuff and doing this with it" has no substance. The really valuable communities are where everyone is working on their own stuff, asking about what they're curious about, and having a lot of conversations helping each other with what they're struggling with. "I can't do this one" is actually the valuable thing to say. They said communities split that way from the start. Some gather people who want to show off and promote themselves, and some gather people who come to learn what real developers are wrestling with and what help they can get. If a community was born the wrong way, you go once or twice and stop going.

What they want from a community is one of two things. Either they can help someone with a problem, or someone can tell them something they were curious about. If they come away with one of those, it's a really good day. It doesn't have to be just development. It can be career stuff and a lot of other things. They said the community they've kept going for several years works that way. Everyone meets up, does their own work, and when someone gets stuck, they ask and help each other. It worked up to ten-something people, but once it hits a few dozen, it's not a community anymore.

8. Why sales

The role they're taking is solutions engineer. In technical sales there are solutions engineers and solutions architects, and they're going the engineer route. What both roles share is POCs and hands-on sessions, and that's what they really wanted to do. Presenting to people on how to use a good product to solve the things they're struggling with, listening to people's pain points and showing them how to solve them. Since they'll be at a company, they'll be using a specific product, but being able to keep doing that kind of work was the draw.

The foreign company they're joining has also built a lot of AI products. They said they want to learn how a company that actually makes money approaches those products from the seller's side. If anything, it's a good opportunity. Last year, they were focused on how to combine development and business. They wanted to do development that was stronger as a business or more useful to people. They still want that. But now they're kind of tired of hearing people talk about AI. Most companies doing that aren't producing anything from it.

9. AI is just a tool

When I asked why they got interested in psychology, it led into the core skills of a developer. They said they once drew up a chart themselves. Of a developer's abilities, development skill is 10 percent and communication is 50 percent. They even gave talks on it, long before the AI era. The most important skill for a developer isn't coding or specialized knowledge. It's understanding what the other person is doing through communication. Especially with other teams, and above all, really understanding the requirements of the people asking for things and making an effort to understand the domain knowledge. Mental models are about how people build up their problems in their heads, so they figured understanding that would help you put context into prompts better in the AI era. That's how it connected to their interest in psychology.

When I asked what changed before and after the AI era, the answer surprised me. They said they're not sure how big a deal the phrase "AI era" should be. They see it as being on the level of before and after smartphones, or the shift from phone calls to texting. It's big. The world changed when smartphones arrived, and it changed when texting took over. But their read is that expectations right now are heading into science fiction territory. The more they study AI models, the more they see them as just tools. In the past, to recognize speech you needed a speech recognition tool, to recognize text you needed a text recognition tool, or else you had to feed it in through a fixed language called a programming language. Now you can bundle voice, images and so on into one input and easily get back answers in the form you want. That's all.

I said it's similar to when Excel came out. Before Excel, you did it by hand, and Excel made it easier. Excel didn't suddenly let you do things you couldn't do before. They were on the same page. A tool like that doesn't suddenly make the car go faster. Input got easier, and now anyone can easily get the output they want. The concept evolved one step, and people who understand that concept benefit, no matter what field they're in.

I shared something I've been struggling with. AIs don't just work with each other. For me to understand, communicate and make decisions, I end up having to know everything myself, and the volume has gotten huge. With all the multitasking, my head feels like it's going to explode. Their answer was clear. When they had ten people under them, they didn't know everything those ten people were doing either. They just needed to understand the overall context, delegate well and check the work well. The AI era is the same. You don't have to know everything. You need to understand the core, explain clearly why something is needed, and be able to judge the results.

It just brings back results faster than a human. Nothing about the essence has changed.

10. To junior developers

Finally, I asked what they'd tell complete juniors to focus on. There's a lot of talk that AI is making things hard for developers, so what would they recommend to people who've just fallen in love with coding? They named two things.

The first is learning how to talk with the people you work with, especially external communication. Most people are only interested in what they're doing, and they said if you only care about your own stuff, you'd better be a genius. Otherwise you won't survive. The second is your learning curve, and curiosity. You need to have curiosity, they said.

They also stressed two things that communication absolutely needs: humility and sharing. They said sharing is a developer's core value. Sharing the know-how you have with others. The very idea of open source came from sharing what you know, so others give freely of what they have too, and everyone gets better. But these days, starting with prompts, a lot of people don't share, they said. Some people sell their prompts, and getting paid to share is still sharing. The problem is people who don't share at all, saying, "This is my know-how. Why should I share it with the team?"

They believe sharing is what makes you the best in the end. If you're a super genius, you don't have to share, but the more of a super genius you are, the more sharing actually pays off. The hardest asset to get is influence, and you only earn a reputation by giving things away. Sharing is how you prove you're really smart. You share because you're confident you can keep growing while others are still learning what you gave them. When they see someone who doesn't share, all they can think is, "So that's all you've got." Who would want to share their own stuff with someone like that?

They summed it up in a few lines. If you're 20 or 30 years into your career, you can hold out without sharing, work another 10 years or so and be done. But if you're five years in and don't share, how are you going to get through the next 25 or 30 years? In the end, it comes down to mindset and attitude. Of everything in communication, they said, that's what's needed most right now.

Originally written in Korean on Brunch, June 22, 2026.