I’m getting ready to shut down my server.
To be precise, I’m not shutting down the blog — I’m shutting down the backend server that runs it. The site will still be here. It just becomes a quieter, lighter, cheaper static site. No database, no admin panel, no comments, no dynamic endpoints, none of the things I once thought “a website should at least have some features.”
It sounds a bit like arranging a dignified funeral for my own site.
My blog was born in 2012.
The internet wasn’t what it is today. Grassroots webmasters hadn’t fully left the stage yet. Personal sites, resource sites, forums, blogs, link directories, download sites were everywhere. Back then a lot of people genuinely believed that if you could write a little, knew a bit of technology and were willing to wrestle with a server, you could claim a patch of ground online.
Personal tech blogs got a slice of that too.
You wrote a tutorial, solved an error, recorded a mistake you’d made, and search engines brought the people who needed it to you. Someone read it, someone bookmarked it, someone reposted it, someone left a comment asking a question. Even with almost no income, it made you feel: this thing has value.
Blogs weren’t as polished back then. Less SEO gaming, less AI-generated filler, less of the constant “most complete guide on the internet,” “hand-holding tutorial,” “everything you need in one post.” A lot of posts were just an ordinary person sitting at a computer writing down the problem they had just solved.
The layout might have been ugly, the screenshots blurry, the code inelegant — but it was real, it was useful, and it had a human temperature.
Then life got harder for independent webmasters.
Filing and registration, review, comment moderation, forum restrictions, content compliance, one policy requirement after another, layer after layer pressing down. A lot of grassroots sites didn’t die of technology, or of lost enthusiasm. They died of “too much hassle.”
Personal blogs were never built to make money. Most webmasters got a bit of traffic from search engines and occasionally a little ad revenue, barely covering servers, domains, certificates and object storage.
Even so, many people kept going.
Because for us, a blog wasn’t just a website. It was more like a public notebook, a personal archive, a record of growing as an engineer.
Read something you wrote ten years ago and you might find it naive, might even want to delete it — but it recorded how you understood the world back then, how you solved problems, how you slowly turned yourself from a beginner into someone who knows their way around.
Then came the app wave.
Traffic moved from PC to phone; users moved out of browsers into one closed app after another. Information stopped flowing freely across web pages and started being distributed, recommended, filtered by algorithms.
Official accounts, short video, Xiaohongshu, paid communities, every kind of content platform — they slowly siphoned away the attention that used to belong to the open web.
Personal sites kept losing traffic.
Tech blogs were luckier, relatively speaking. Technical problems still need searching, errors still need solving, config files still need someone to paste them. As long as programmers keep copying error logs into a search box, personal blogs held on to a last bit of room to live.
But now AI is here.
This time it’s different.
In the past, search engines at least still brought users to the site. A user searched a question, saw a title, clicked into the blog, read the post. Even if they stayed for thirty seconds, that was a visit.
To a webmaster, that visit meant the site was still alive, that the content was still being seen.
Now users don’t even need to open the site.
They ask AI directly:
“How do I fix this error?”
AI answers directly.
They ask:
“How do I write this config?”
AI generates the config.
They ask:
“How do I integrate framework X?”
AI gives the steps, gives the code, and throws in a warning about where you’re likely to trip up.
From the user’s side, that’s obviously more convenient. Honestly, I use it too, and I enjoy it.
The problem is that for a personal blog, this is close to the final cut.
A lot of what AI answers with was learned from those very articles, docs, blog posts and forum threads out on the internet.
We small webmasters wrote up the problems, organized the experience, filled in the details — and the content got absorbed into a model and turned into a more efficient answer machine. Users got answers, AI got the applause, search engines kept the traffic, and the people who wrote the original content slowly lost their visits.
This isn’t a complaint about AI.
I’m not against AI. Quite the opposite — I use it every day now, writing code, looking things up, organizing docs, generating scripts. I even run the structure of this post past AI sometimes.
It’s just that as someone who has been writing a blog since 2012, I feel this very clearly:
The era of the personal blog is over.
It didn’t die suddenly.
It got filtered by policy first, then had a share taken by apps, then got diluted by short video and content platforms, and finally AI closed its hand around its throat in a way that was elegant, efficient and even reasonable.
So I’ve decided to stop forcing it.
I’ll shut down the current backend server and publish everything — existing posts and whatever I write later — as static pages.
The site will still be reachable and the posts will still exist. It just won’t depend on a dynamic backend, won’t need a database, won’t require maintaining a pile of online features that few people use.
The benefits are obvious:
- Low cost
- Little maintenance
- Smaller security surface
I used to think a personal site should be “complete”:
It needs an admin panel, comments, tools, endpoints, dynamic pages.
Thinking about it now, none of that is necessary.
The core of a personal blog was always the posts themselves. As long as the content is there, the links are there, search engines can still index it, and someone can still land here through some keyword now and then — that’s enough.
As for the developer tools on my site, most will probably stop working.
If you’ve used any of them, or still need one later, just contact me. I’ll send you the code and you can deploy your own.
None of those tools were complicated anyway. I wrote them one afternoon and put them on the site for myself and anyone else who wanted them.
The money I save on servers is going into renewing Codex and Claude Code.
It’s darkly funny:
I shut down a blog server whose traffic AI took away, then handed the money to AI coding tools.
But maybe that’s just the direction things are going.
Personal blogs won’t disappear completely.
Some people will keep writing. Some will keep missing the era of the open web. Some will keep needing a corner of their own that belongs to no platform, isn’t fed by an algorithm, and doesn’t depend on a recommendation feed.
It’s just no longer the kind of thing to measure in traffic and revenue.
From now on my blog is more like a static archive.
It won’t be busy. It won’t update often. It won’t have commercial value.
But it will try to last longer, weigh less, and stay quiet.
In a sense this isn’t an ending either.
It’s just that I’ve finally admitted:
A personal blog doesn’t need a server to prove it’s still alive.