Very few posts on this blog explain how to do something in software. On the blogs I ran before this one, and killed later, I wrote a lot more of that kind of content, because I needed it more back then. But I noticed that with AI, this need slowly went down for me too. At some point I asked myself why I don't go back to it, and how much sense that would even make.
I used to open Google, type my problem, and somehow find a solution close enough to what I needed. Now the numbers tell a different story.
In 2026, less than a third of Google searches end with a click to any website. For the rest, AI Overviews already answer the question, so the user never leaves the search page. On queries where an AI Overview shows up, click rates sometimes drop as low as 17-20%.
So which posts still get clicked? Not the "how do I do X" ones. Comparison posts and posts that share real experience get more clicks. The other type gets pulled out by AI and handed straight to the reader, without the blog in between.
That's tiring, honestly. My writing gets picked up by AI before it even reaches my site, gets used, and my site gets no traffic from it. But does that actually matter? Is it worth quitting over?
I sat down and went through my old posts, and a pattern showed up. I wasn't writing "how to use X." I was writing what happens when you actually use X in production, and what I learned from it. Turns out I was already doing the right thing for this era, without planning it.
The first type is generic reference material. It's already in the docs, repeated in ten other blogs, and AI can summarize it faster than I ever could. The second type is something that happened to me. Which mistake I made, why I made it, how I noticed it. There's no documentation that can summarize that, because that experience only exists in me.
Research backs this up too. AI isn't killing traffic, it's redistributing it. Clicks are dropping on generic, unbranded information queries. But for the sources AI actually cites, the ones that say something real, the traffic that comes through is smaller but worth more. Posts with original research and real opinions are doing better than plain SEO posts right now.
So the question isn't "should I write how-tos." It's "which how-to is actually worth writing."
Writing "here's the syntax for X" doesn't mean much anymore. That information is everywhere, and Google hands it over faster and cleaner than I can. But writing "I used X on a real project, got three things wrong, here's how I fixed them" is still worth it. Because nobody else can tell that story but me.
I'll keep writing how-tos. I'm just going to write less "how it's done" and more "how I did it." Google can somehow give the other answer. It can't give the story.
Sources:
Originally published on Listen. Code. Reflect.