The World That No Longer Exists
I stopped blogging. Not because I lost interest. Because it stopped making sense.
What happened after that, I never would have expected. But let’s start at the beginning. In a world that no longer exists.
Writing as Thinking
For years, I wrote how-to articles. “How to install CentOS with VirtualBox.” “Your First Docker Container.” “Event-Driven Architectures with Kafka and Spring-Boot.” That kind of thing.
I told myself I was doing it for others. And I was, partly. But the real reason was selfish: I wrote to understand. Taking a vague idea and forcing it into a coherent tutorial was the best way I knew to actually learn something. If I could explain it clearly enough for a stranger to follow, I probably understood it myself.
The blog was my thinking tool. I just didn’t call it that.
And it worked. You write, people read, Google delivers traffic. Everyone’s happy. The world made sense.
Learning in Public
There was something else, too. Something I only recognized in hindsight.
Every article I published was a snapshot of me figuring things out. Not polished expertise. A process. Here’s what I tried. Here’s where I got stuck. Here’s what finally worked. Sometimes: here’s what I still don’t understand.
That’s learning in public. You share your mistakes alongside your insights. You help others, and the act of explaining makes you understand things better than you did before. The Feynman Technique, basically, except with a blog and a comment section.
It built a small community. People emailed me. Other developers linked to my posts. It felt like I was contributing to something.
The Offshoring Parallel
Before I get to the part where everything breaks, a quick detour. Because this isn’t the first time I’ve watched a technology wave roll in and wondered whether it was coming for me.
- I was working at a company where the word “offshoring” started showing up in meetings. The message was clear: your job can be done cheaper somewhere else. Developers in India, developers in Eastern Europe. Same code, lower salaries.
Except it didn’t work out. Not the way they’d hoped. You can transfer code, but you can’t transfer context. You can’t offshore the understanding of why this particular system was built this particular way. Knowledge and context don’t fit in a handover document.
So we survived. And I filed it away as a lesson: human expertise is irreplaceable. The people who understand the “why” will always have jobs. The rest is just typing.
I felt safe.
I would remember that feeling later.
The ChatGPT Disruption
November 2022. ChatGPT launches. Within weeks, the world is different.
At first, I wasn’t worried. Sure, it could answer questions. But its knowledge was frozen in time. It didn’t know about the latest Kotlin release, or the newest Jetpack Compose APIs. If you were learning something fresh, something new, you were still ahead of the machine. Your blog still had value.
I told myself: “I can do things the machine can’t.”
Sound familiar? It should. It’s exactly what I’d told myself in 2009. “They can’t offshore context.” Same logic, same comfort. Same false sense of security.
But this time, something felt different. The machine wasn’t just cheaper labor in a different timezone. It was right there, on the same screen, answering the same questions my blog answered. And answering them faster. No cookie banners. No ads. No scrolling past a life story to get to the command you need.
I just didn’t want to see it yet.
The Traffic Collapse
The numbers didn’t lie, though.
Blog traffic dropped. Not dramatically at first. More like a slow leak. Then faster. The pattern was clear. People weren’t searching Google for “How to set up X” and landing on my blog anymore. They were asking ChatGPT directly. And ChatGPT was giving them exactly what they needed, tailored to their specific situation, in seconds.
Tutorial blogs, how-to websites, even Stack Overflow. Everyone felt it. The content types that could be replicated by a machine were being replicated by a machine.
Replaceable content got replaced. Simple as that.
When the Machine Writes Better
Here’s the part that really stung.
It wasn’t just that AI could answer the same questions. It could write better answers. Consistent structure. Complete coverage. Clear language. No typos. Adjustable to any skill level. Need a beginner explanation? Done. Need the advanced version? Also done. In three seconds.
My how-to articles. The ones I’d spent hours crafting, proofreading, rewriting. A machine could produce something equivalent in the time it took me to open my editor.
For standard content, AI is objectively better. Technical documentation, installation guides, concept explanations, tutorials. The machine wins.
Where it can’t compete: personal experiences. Nuanced opinions. The stuff that only exists because a specific human lived through a specific situation. But that’s not what my blog was. My blog was tutorials. And tutorials were now a commodity.
Automation of Mental Labor
This is where the offshoring parallel breaks down. And where it gets scarier.
In 2009, the threat was human. Other developers, in other countries, doing the same work for less money. But they had the same limitations I did. They needed time to learn. They needed context. They made mistakes. The playing field was uneven, but it was the same field.
AI is different. What industrialization did for physical labor, this is doing for mental labor. Code boilerplate, standard documents, summaries, routine analysis. All automatable. Not by cheaper humans. By machines that don’t sleep, don’t get tired, and get better every month.
And this time, the machine is learning the context. That was supposed to be our moat. The thing that couldn’t be transferred. The thing that kept us safe.
The Question of Purpose
So here I was. The blog was dying. Not because I stopped writing. Because the world stopped needing what I wrote.
Why still blog?
That question sat there, unanswered, for months. I didn’t have a good response. The old reason, writing tutorials to share knowledge, was gone. Machines did that better now. The traffic that used to validate the effort had evaporated. What was left?
I think the answer has something to do with a shift. From “How do you do X?” to “What does it mean that X is changing?” From writing to inform to writing to understand. From chasing traffic to chasing meaning.
But I wasn’t there yet. Not in 2023. That realization would come later, after a lot more happened. After AI didn’t just replace my blog, but started replacing parts of how I work.
That’s what the next post is about.
The false sense of security doesn’t last long. In the next part: AI starts researching the web on its own, Google itself starts serving AI answers, and the last moat, “the AI doesn’t know what’s current,” falls.