Why I still write
Table of contents
A few years ago, I had a problem to solve at work — one of those annoying, specific-enough-to-have-no-obvious-answer problems that are still common enough that someone, somewhere, must have dealt with them before. I did what everyone does: I went to Google, found a forum post with the perfect solution — working code, clear explanation, just the right amount of context — copied it and solved the problem without changing a single line. When I went back to the page to thank the author, I discovered the author was me: I had written that tutorial months earlier and forgotten about it completely.
It’s a slightly silly story, and it was funny at the time, but what interests me about it is something else: nobody asked me to write that post, nobody paid me, and almost nobody read it. I wrote it because I had just solved a problem and wanted to put the solution somewhere — partly to help others, partly because putting things into words helps me understand them better. Years later, the person it helped turned out to be myself. That cycle — write, publish, forget, rediscover — is what keeps me blogging in 2026, when maintaining a personal blog already feels like a gesture from another era.
Writing to think
I blog because I have a kind of need to share what I know, think, and do. I wouldn’t call it altruism — it’s closer to the impulse of telling someone a story after you’ve lived it. The act of writing forces me to organize what I’ve learned, to find the logical sequence between steps, to notice the gaps in my own understanding. I often start writing a post thinking I’ve grasped a subject and discover, somewhere around the third paragraph, that I haven’t understood it quite as well as I thought. The post forces me to go back, test again, confirm. By the time I’m done, I know more than I did before I started.
There’s also something else, harder to explain. I admire people who write — people who maintain technical blogs, who publish notes about what they’re learning, who document their projects in public without expecting anything in return. Being part of that scattered community — no headquarters, no charter, no algorithm organizing it — is something I value. It’s not a community in the way social media uses the word: nobody likes, shares, or drops a fire emoji in the comments. It’s a community in the older sense: people doing the same thing, each in their own corner, who every now and then cross paths along the way.
I don’t blog to make money. I have made money from it before (and even taught others how to), but becoming a “professional blogger” was never the point. I don’t blog to build an audience, a personal brand, or a reputation. I do blog because I like to write, and the blog is where that writing gets an address and stays available for whoever might need it — including, as it turns out, myself.
The internet that shrank
Anyone who has used the internet long enough remembers when it was something else. Not better in every way — it was slower, uglier, harder to navigate — but decentralized in a manner that now seems improbable. Everyone could have their own corner: a GeoCities page, a Blogger site, a forum about any subject imaginable. The content was amateurish and often rough, but it was made by real people who wrote because they wanted to, not because an algorithm rewarded the right format at the right time. Traffic came from referrals — one blog cited another, a forum linked to a tutorial, people followed links the way you walk through a city without a map. There was something organic about that ecosystem, even if we hardly used this word.
In 2026, the landscape is unrecognizable. Recent studies estimate that over half of all internet traffic is generated by bots. Cloudflare’s CEO has predicted that automated traffic will surpass human traffic by 2027. Reddit co-founder Alexis Ohanian said openly that much of the internet is dead — overrun by bots, semi-automated content, what he called “LinkedIn slop.” The dead internet theory, dismissed as a fringe conspiracy just a few years ago, is now discussed in academic papers and mainstream press — not because the theory is entirely correct, but because its observable part — the progressive displacement of human content by synthetic content — is increasingly hard to ignore.
The slop
The word “slop” was named Merriam-Webster’s word of the year in 2025, and it describes with precision what floods the internet today: AI-generated content that exists only to exist; blog posts written by nobody, about everything and nothing, optimized to rank on Google and convert clicks into ad revenue; YouTube channels with synthetic voices reading Reddit posts over lo-fi beats; Instagram accounts publishing “art” that looks like a melted corporate color palette. The volume is immense and the quality uniformly mediocre — technically functional, but with no human intention behind it.
The problem isn’t that AI produces bad content; it’s that it produces indifferent content at industrial scale. A blogger who writes a bad post has at least tried to say something. An AI-powered content farm isn’t trying to say anything — it’s trying to occupy space in search results. And the more space it occupies, the harder it becomes to find the text that someone actually wrote, thought through, revised, and published because they had something to say.
The walled gardens
Social media completes the siege from the other side. In the early years, platforms like Facebook and Twitter worked as amplifiers: you published something on your blog, shared the link, and people clicked through to read it. That model worked for a while, until the platforms realized that every click on an external link took the user out of their ecosystem — and away from their advertising. One by one, they all rewrote their algorithms to boost native content and bury outbound links. Twitter was the last to fall, and when Elon Musk took over in 2023, referral traffic all but vanished for anyone who depended on it.
The result is a self-devouring system. The content that survives is what is born and dies inside the platform — short, fast, optimized for immediate engagement. Anyone who writes long texts with links, references, and depth loses by definition: not because the text is worse, but because the platforms’ business model cannot accommodate the attention it demands.
Small fires
Maintaining a personal blog in 2026 is an anachronistic gesture, but it’s not the only one: there is a whole constellation of practices out there that, by the dominant logic, should have disappeared long ago — and that are not only still alive but gaining new people.
In January 2026, the Associated Press published a story about the revival of letter writing. Pen pal clubs, typewriter clubs, communities dedicated to calligraphy and wax seals — all on the rise, while the rest of the world optimizes every second of communication to fit into a three-word message and an emoji. In Dallas, a DJ created a monthly “analog gathering” — vinyl listening sessions, letter writing, coloring — because, as he put it, real connection happens when we can touch or see something, not when we swipe through an infinite feed. Postcrossing, a project that connects people around the world to exchange postcards, remains active and growing. SXSW 2026 dedicated an entire panel to what it called the “Analog Revival” — how creators use authenticity and scarcity to build communities in the age of AI.
And there’s more. Zines — independent, self-printed publications distributed hand to hand or by mail — never died, with dedicated fairs in dozens of cities worldwide. Ham radio, which predates the internet and should have been made irrelevant by it, maintains active communities, annual events, and its own culture of experimentation and knowledge sharing. Book clubs, which existed before Gutenberg’s press democratized the book, still meet in cafés and living rooms — and in 2026 are, I’d venture, more popular than they were a decade ago.
What all these practices share is not, it seems to me, nostalgia. Nobody writes letters by hand because they think the postal service is more efficient than email, and nobody listens to vinyl because the sound is objectively better than streaming. What they share is a conviction — often unspoken, often not fully conscious — that the process matters, that slowness is not a flaw but the point; that sitting down, thinking about what you want to say, and choosing your words carefully has a value not measured in reach, likes, or time on page. They are small fires, each in its own corner: they don’t light up the world, but whoever passes nearby finds them — and warms up.
The blog as permanent draft
I maintain my blog on a stack that, to anyone outside the field, sounds like alphabet soup: Hugo, GitHub, Cloudflare Workers, Pages CMS. In practice, what this means is that there is no intermediary between what I write and what gets published, no platform deciding who sees my text, no algorithm measuring engagement and adjusting distribution accordingly, no AI model training on my posts without my consent (well, that last one probably happens anyway). The text is mine, the address is mine, and the content stays where I put it for as long as I want it there. This isn’t a political statement — at least it wasn’t one when I made the choice. It’s a practical decision that, over time, acquired a meaning it didn’t originally have.
Jeremy Williams, who has maintained a sustainability blog for over twenty years, wrote in early 2026 that his generosity in publishing openly had been “abused” — he hadn’t just given his articles to readers, but also to Google, Facebook, and OpenAI. Richard MacManus, of Cybercultural, compared browsing the indie web to browsing a record store: you have to look, but the treasures you find are genuine. Jeff Geerling, who maintains over three hundred open source projects, described how the flood of AI-generated contributions is degrading the infrastructure that sustains the real internet. Different voices, different perspectives, all describing the same thing: the space for content made by humans, with intention and care, is shrinking — and those who keep producing it are increasingly aware that they do so for reasons no metric can capture.
I don’t know if blogging in 2026 is an act of resistance. Probably not, at least not in the strong sense of the word. I use AI in my work every day — to automate tasks, explore ideas, and speed up what would be tedious to do by hand. I’m not against the technology, far from it. I’m against the indifference it enables when used without intention. Slop is not an AI problem; it’s a problem of people who have nothing to say and now have the tools to say their nothing at scale.
I have things to say — not always important, not always useful to anyone else, but thought through, written, and published by a real person at a real address. If someone finds them, great. Otherwise, the text stays there all the same — like that post I once found on Google that, without knowing it, I had written for myself.