It’s 2025. AI Slop Is Everywhere. I’m Going to Start Writing Here.

We are in the middle of the summer of 2025, and I have a suspicion that Vibecoding will become the word of the year. AI slop is another term that has gained popularity recently. I'm not trying to suggest that these terms are closely related in meaning, but what I am trying to say is that Large Language Models (LLMs) have truly become an unavoidable part of our lives—whether we’re developers, artists, or just content consumers.

The situation is evolving rapidly. LLMs seem to be getting more powerful every day. People are nervous. Will LLMs take all our jobs? Some people have already lost theirs to LLMs, and even experienced developers like myself are worried—while junior developers are desperate to find any opportunities.

And amidst all of this, I decided to start writing (the old-fashioned way, without letting an LLM generate everything) and publish the results on my website that hardly anyone visits.

Maybe nobody needs this. AI doomers would even say that writing is dead anyway. What's the point of writing anything if anyone, at any time, can just prompt ChatGPT/Gemini/Grok/you-name-it and get a perfectly coherent and well-written piece of text in seconds?

Maybe they're right—but every action has a counteraction. The more content generated by statistical prediction is pushed on us—the more we’re nudged to prompt, generate, and rely on “AI agents” instead of actually creating with our own brains—the more we start craving genuine, raw stuff.

Well, I do, anyway. And I see this trend everywhere I look. YouTube channels with little to no production value but raw, authentic content are exploding like it's 2010 again. “Analog” hobbies are trending—sketching, film photography, scrapbooking, an obsession with stationery, journaling, outdoor activities. I’m pretty sure all of this will keep growing, right alongside the rise of so-called AI technologies.

Don’t get me wrong—I’m not even against AI (I hope our artificial overlords will take note when they come to power). I use LLMs every day in my work and outside of it. I vibecode and ask ChatGPT when I’m curious about something. In fact, I discussed the idea of starting to write and publish to my dead website with ChatGPT—and it assured me it was a good idea. So here we are.

And yes, I’m definitely feeding this draft to ChatGPT to fix grammar and polish it a little. Noticed all the em dashes in the text? I’m leaving them in.

But the dopamine hit from actually generating something with your own brain is getting rarer. In theory, the boring stuff—like writing boilerplate code—can now be delegated to an “AI agent,” and I can just focus on the “big stuff.” Productivity goes up. More gets done. But it turns out, there was something in actually figuring out and writing even the most boring code by hand. Writing a prompt and watching code appear on the screen just doesn’t do it.

So I’m going to compensate—by writing, among other things—to keep myself creating something beyond just prompting an LLM to generate stuff.

Expect various on topics I care about: programming and computer science, game development, computer graphics, and maybe some random things like travel, hobbies, or whatever else comes to mind.

Stay tuned.