How I Grew a New Blog to 302K Impressions in 3.5 Months by Writing Every Post by Hand
In late February, I launched a new project called Darko Unity, named after myself and the Unity game engine. My goal was to share insights from my 11 years of Unity development experience while also attracting tutoring students to help fund this new adventure. At first, I didn't expect much. I thought blogging was dead and that nobody would care about what I had to say. But I was wrong. What happened next completely changed what I thought I knew about marketing.
Your competition is AI slop, and I mean it
I come from a software engineering background, and I am always eager to understand how things work under the hood. That led me to learn SEO, and not just the typical kind you get from a WordPress plugin. I decided to build my own CMS and my own backend for the Darko Unity website. Along the way I learned about technical SEO: how Google renders pages, and how that affects the way you show up in Google and other search engines.
As an engineer, I have at least a surface-level understanding of how AI training works. These big companies don't want to train on data that already exists, or, even worse, on data that was generated by AI itself. There is a phenomenon called data dilution. Here is one explanation:
In the context of artificial intelligence, data dilution refers to the degradation of AI output quality that happens when models are trained on AI-generated content.
What happens when you train a model on data that was generated by a model, which was itself trained on data generated by another model? You get the point. The data gets diluted until it is close to worthless. This is well understood in the LLM and AI industry.
So these big tech companies tune their algorithms to avoid training on that kind of data. And if your blog is written by AI, then search engines, especially Google (let's be honest, the one that matters most), are far less likely to recommend your content. Meanwhile, as I was building the website and writing my posts, I realized how many people had been sold on the dream of AI automation: you sleep, and AI writes your blog for you. I tried it too. I even started building an LLM-based CMS that could run on its own.
But then it hit me: the content was barely ranking. And when I shared those posts on social media, readers rejected them, because people can tell when an article was written by AI, and they don't want to read it. I looked at other blogs, and at the influencers I followed, and most of them were using AI to write their content. I noticed they were losing impressions, and the writing itself lacked any personality. It was just slop. Content with no value. I was convinced this was data dilution at work. And sure enough, around April, Google rolled out an update aimed at fighting AI slop, because the internet had turned into an AI slop fest.
So I had an idea. What if I wrote the content myself? I had no idea whether it would change anything, but I tried.
How writing every article by hand got me to the top of Hacker News
I started this experiment and quickly noticed a shift. People were more receptive to my writing: more likes, more comments, more genuine interest. By the third hand-written article, I had another idea: what if I posted it on Hacker News? I had an old account I had never used, and I made my first-ever submission. I also shared it in a few other places, like Reddit, X, and some Facebook groups. Then I went on with my day and forgot about it, since I rarely visit Hacker News.
I was about to go to sleep when I noticed the live analytics in my CMS starting to explode. My first thought was that I had a bug in my code, since the feature was only two or three days old. But the reads kept coming. My read tracking works a bit like YouTube's: I don't count refreshes or accidental clicks, so the numbers are legit. At first I assumed Elon Musk had shared my post on X or something, because I was getting around 100 reads every five minutes. The day before, it had been 100 reads in a whole day. I checked all my social media. Nothing. Nobody had shared it. Then I opened Google Analytics, saw the traffic coming from Y Combinator, and it clicked. My first-ever Hacker News post had blown up.
There was nothing especially remarkable about that post. It was just me writing about the current open-to-work epidemic on LinkedIn and what I thought was causing it. But I believe people responded to it because it was real. These days almost every post is written by AI, and people are fed up with it. My guess is that some of the Hacker News moderators read the article, liked it, and pushed it up.
I figured I had just gotten lucky. But over a short span, two more of my posts were also featured on Hacker News. People simply liked them.
Here are the two other articles, if you want to read them:
How I Learned Unity the Wrong Way
C# in Unity 2026: Features Most Developers Still Don't Use
I always assumed Hacker News was an elitist tech clique, but it was not. It is the one place I have found where people are genuinely open-minded. They don't take things out of context, they actually read the full article, and they try to understand what you meant. It was the first place in my life where I noticed that commenters weren't trying to one-up you or put you down. I had assumed the only content that did well there was filler like "APPLE GETS A NEW CEO," but one of my articles was a C# coding tutorial for Unity developers, and people still read it. They enjoyed discussing it.
What did I learn from this?
I learned that people are still reading blogs, maybe even more than before, since the end of the coronavirus pandemic. My theory is that this is largely because of AI. Chatbots are text: you have to read to use them. So I suspect people are reading more now and watching a little less YouTube, which really dominated before modern LLMs.
Human-written content beats AI-generated slop by a wide margin, because AI systems are hungry for genuinely new ideas to train on. I know there are ethical questions tangled up in this, but what can I do?
My personal theory is that almost any kind of content worth consuming will, for the foreseeable future, have to be made by humans. So I would not worry too much about AI replacing artists and writers. I don't think it will happen. The bigger risk is to programmers and similar fields. Our work isn't seen, and it isn't art. I know some of you will disagree, but programming was never about art. It was always about solving problems and creating value. A lot of people think writing code by hand is something sacred. It isn't, and I say that as an engineer. I just accept things as they are and keep going.
Post mortem: Google Search Console showed massive organic growth
Anyway, if you made it this far and you care about SEO, try what I did: write your content yourself, and see if it works for you.