My personal thoughts on blog during AI period

The Old Era

2022 was right on the eve of the AI explosion, and it was also my first year working. I remember being so eager to set up my own blog that I spent the early days researching different cloud servers and frontend frameworks. I eventually settled on HEXO + Aliyun + Typora. Then, stealing moments here and there over a month, I tweaked the frontend layout and uploaded a few rough articles. Because the framework updated frequently and I knew nothing about frontend development, I hit plenty of pitfalls during setup, scouring outdated tutorials for solutions. It wasn’t smooth sailing, but the joy of finally debugging something and getting it to work is something I still remember vividly. Building it step by step from scratch, picking plugins, optimizing load times, adding modules—that whole process was incredibly satisfying.

Back then, personal blogs were the best way to learn new tech and solve technical problems. Unlike the incorrect, ambiguous, or outdated content search engines often threw at you, blog posts were original, summarized, and verified by the authors. Most of what you found could be directly learned and applied. So every time I stumbled upon a new blog that matched my aesthetic or had valuable content, the sense of discovery was indescribable. Personal blogs became another calling card for programmers, besides GitHub. The blog’s style could really show the author’s personality and taste—whether elegant or minimalist. From my observation, frontend developers’ blogs were usually visually appealing, while backend developers were often less willing to spend time tweaking the frontend UI, seeing it as pointless. This resulted in the opposite aesthetic: minimal colors and images, mostly just text and code—a stark minimalism. docs.kernel.org is a classic example. Either way, they all contributed to the rich and diverse blog culture.

At first, because my skill level wasn’t enough for high-quality original content, most of the articles on my blog were operational guides and summaries distilled from various sources about Linux and Android tech. But I was still obsessed with it, because the amount I learned throughout the entire process far exceeded what I gained in classrooms or at work. It pushed me to actively explore different fields at the start of my career. Having a blog that’s maintained regularly and systematically in daily life is a really happy thing. I thought I’d keep going like this forever, that one day my blog would be browsed and bookmarked by peers. Then, suddenly, the AI era arrived.

The Impact of AI

This blog was built using Claude + DeepSeekV4, and it took just one hour to set up both frontend and backend and have it running perfectly. A lot of the tech choices were even made autonomously by AI. Compared to the first version of my blog, which took me a month to build by hand, this one is more polished and more stable. On one hand, I really enjoy being isolated from the frontend world—I absolutely hate debugging CSS and HTML pixel by pixel. On the other hand, it left me feeling empty. I didn’t experience any of the surprise or satisfaction I used to feel.

Ever since ChatGPT was born, my frequency of visiting great blogs has dropped dramatically. Nowadays, I’ve forgotten how long it’s been since I solved a technical problem by reading someone else’s blog. Instead, I ask AI everything, big or small. For non-critical code, sometimes I don’t even bother reading the specific error message—I just let the Agent fix it. And by May 2026, the latest Agent + LLM combos can already build and run a complete personal project in the code domain, handling all kinds of errors very accurately. I can’t help but wonder: in the face of an “omniscient and omnipotent” AI in the coding world, what value does a personal blog still hold? Has it just become a programmer’s toy for self-indulgence? I couldn’t find an answer, so I went and looked through many great programmers’ blogs. Many had already stopped updating, and others had expressed their confusion and fear about AI in their columns.

As a newcomer who entered the field around the same time as AI, I’m not like some veterans who cling to “old-school programming.” On the contrary, I embrace AI actively but pessimistically. A significant part of my previous work was improving unit test coverage for an older project. It took about half a year to complete the coverage for one repository. Later, I brought in AI to help. I have to admit, the test cases it wrote were more comprehensive and standard than what I could come up with after much deliberation. As long as the boundaries are well-defined, at least in the C/C++ domain, it’s already stronger than 80% of traditional programmers. Now, I’ll learn the underlying principles and applications of various AIs because of anxiety, but on the other hand, I unconsciously avoid any AI-related content I come across—I just hurry and scroll past it. One day I realized this and reflected on the reason. The conclusion: it stems from deep fear. I’m not afraid it will make me unemployed. I’m more afraid it will make me lose my sense of value and meaning. I’m afraid it will strip away the thrill I get from battling wits with code and bugs. Of course, I could keep seeking that thrill the old way, but work and school are fundamentally different. School is for learning and selection, so not copying answers is reasonable. But work is about output and efficiency. Even if your skills are weak, as long as you deliver on time and in quantity, you’re a good employee. So I don’t think any boss can tolerate someone stubbornly doing “old-school programming” for a day when AI could get the same result in an hour.

The New Era

Coming back to the discussion about blogs, maybe in this era, blogs have gradually become something like personal notebooks. When people encounter problems, their first instinct is to ask AI for help, not to read blogs. Similarly, when learning new tech, the Q&A format with AI is more scientific and effective.

I still believe writing a blog is valuable. At night, quietly listening to music and writing a blog post lets me shut out the fatigue and arguments of the day. Any programmer who’s worked a real job should understand: the workday is filled with meaningless meetings and colleagues passing the buck. Being able to fully immerse yourself in coding becomes a form of rest. I think blogging can extend that “rest.” While writing, you can reflect, systematically organizing your knowledge and insights. It’s also a great opportunity to improve your writing skills. I also have a hunch that in the future, blog content will have less technical substance and more personal reflections—just like this article. Because most mid-to-low-end software tech has become cheap and readily available, no longer worth spending hours carefully crafting a post to share.

I won’t think about whether this is good or bad. When the wheels of progress roll over, individuals can only adapt positively. Just as the spinning jenny didn’t fade away because textile workers opposed it.