The Day Google Summons Back the "Obscure Personal Blog": Beyond AI Article Fatigue

I don't see the kind of blog I used to silently thank anymore

Ten years ago, when I was a junior engineer, my workflow was simple. Red error text would scroll past the console. I would copy it verbatim into Google. The first page was always StackOverflow in English. The second page was noise. On the third page, an FC2 blog would appear — outdated layout, a strange tag cloud in the sidebar, no SEO polish whatsoever.

And there, on that scruffy page, would be someone who had been stuck on the exact same error. Trial-and-error logs pasted raw. One line at the end: "this fixed it." I would mutter "thank you" to my monitor.

After a few years of office life, when I started tinkering at home again, personal blogs kept saving me. When I wanted to wipe Windows off a Surface (Microsoft's first-party laptop) and install Linux Mint, I found someone who had attempted the exact same thing. They had photographed their BIOS screen with a slightly blurry phone camera, with honest captions like "stuck here, two days." Without that post, I would still be stuck.

But sometime in the last few years, those personal blogs disappeared from search results. What replaced them was an endless row of "What is X" / "Pros and cons of X" / "Top 5 X" articles — convenient, but devoid of any warmth.

In late 2022, Google added "Experience" to E-A-T

Google has long had a quality framework for evaluating search results. It's called E-A-T, and it's been used since around 2014 — Expertise, Authoritativeness, Trustworthiness. The premise: does the author actually know the field, and is the site a credible source?

In December 2022, Google revised its Search Quality Rater Guidelines and added one more "E": Experience. People who had actually visited the place, actually used the product, actually performed the task — would now be evaluated on a separate axis from expertise or authority.

A minor revision on the surface. But think about what it really said: you can be evaluated favorably even without being an expert or an authority, as long as you actually did the thing. Google had officially blessed the FC2-blog-corner type of person.

The blogger who wrote "stuck here for two days" in some forgotten corner of FC2 wasn't an expert. Wasn't an authority. But they had actually done it. The December 2022 update was, in theory, a mechanism to surface exactly that layer.

In practice, though, it wasn't immediately clear how the addition would propagate into the actual ranking algorithm. Guidelines and algorithms always have a lag.

What the 2023 Helpful Content Update did to personal blogs

The added "E" did not immediately rescue personal blogs. In fact, things got worse first.

Looking back at the 2023 HCU carnage still makes my stomach hurt. Hobby recipe sites, review sites, error-log blogs — "small, individually run, but carefully written" sites lost 50–80% of traffic across the board. Some bloggers reported 95% gone. Google had classified these as "unhelpful."

When I felt "personal blogs have somehow disappeared from search results," that wasn't a feeling. It was something that actually happened.

Around the August 2024 mid-sized Core Update, overseas SEO industry observers began detecting hints of a rebound. And now, the in-progress March 2026 Core Update and the May 2026 Core Update have Google's official statements and analyst commentary pointing clearly toward "experience," "case studies," "demonstrated."

To me, this looks like "Google is slowly summoning back the personal blogs it killed three years ago." It might be wishful thinking. But as an observation, it looks that way.

One clarification, to avoid a misread. When I say "obscure personal blogs are being summoned back," what I have in mind is the technical notebook, the lived work log, the troubleshooting record of someone who got stuck in a very specific situation — the kind of post that actually rescues a future reader. I am not talking about reviving the "had a tough day at work today" type of diary blog.

What Google's "Experience" axis is reaching for, I believe, is the former layer. Diary blogs were precious places for their authors, but they sit outside the kind of revival I'm describing here — the revival of "posts that save someone stuck on the same problem."

The kind of posts I used to thank, looking back

Back to the ten-years-ago story. The personal blogs I muttered "thank you" to had specific faces.

The first was when I was setting up a firewall daemon by hand on an on-prem server. I wanted, simply, to block every IP outside of Japan. Look up a GeoIP database, push country-level block rules into iptables — conceptually simple, full of small traps in practice.

I searched, and found someone doing exactly the same thing on a personal blog. Their scripts, their cron jobs for periodic refresh, verification commands, every spot they'd gotten stuck on. The kind of pinpoint specificity no commercial publication would ever bother with. Without that page, I would have lost two extra days.

The second was the Surface-to-Linux-Mint replacement project. Not exactly a mainstream operation. The Surface has its own quirky keyboard dock; hardware compatibility is its own little adventure.

In that obscure corner of the world, someone was trying exactly the same thing. And crucially, they had photographed their BIOS screen with a blurry phone camera. Half of the photos were out of focus. But that scruffy realness conveyed "ah, this is where you get stuck" and "the setting is buried under this menu" far more eloquently than any text would have.

And alongside the help, a sense of fellowship welled up. "This person is running this setup at home too" — "let me try it." Comrades with the same texture, somewhere on the other side of the search box.

That, I think, is where the warmth came from.

It turns out a lot of people felt the same way

While writing this piece, I searched around just to make sure — and realized my feeling wasn't a new discovery at all. People had been putting the same thing into words years before me.

If you poke around in Japanese for "obscure personal blog" (謎の個人ブログ), you find a recurring genre of summary posts. They mostly aggregate threads from 5ch (Japan's 2chan-descendant message board), where people lament: "Google has stopped surfacing those error-log personal blogs I used to find" and "personal blogs have disappeared from the index."

For example, back in September 2021, a Japanese summary blog called Tte Nan Desu Ka ran a post titled "Obscure personal blog: 'Leaving this here for anyone who hits the same bug'", quoting this line from the underlying thread:

A year later, in October 2022, another summary blog, Popo-soku, ran "Obscure personal blog: 'Leaving this as a note'" on the same theme. The quoted comments there got sharper:

September 2021. October 2022. Both posts predate the 2023 Helpful Content Update by a hair. The thread participants couldn't yet name what Google was doing — but the felt experience that "personal blogs have vanished from the search results" was already widely shared, well before the algorithmic change that explained it.

The phrase "obscure personal blog" (謎の個人ブログ, nazo no kojin blog) has itself become a kind of Japanese internet shorthand. The word "obscure" is a little ironic — the blogs aren't really obscure. They're records of someone you've never met, found only through search, who got stuck on the exact same problem you did. We didn't know their names. But for years, the same people had been quietly saving us.

So this piece isn't a new observation on my part. People had been feeling it for years. My contribution is to connect that long-running feeling to what Google is doing in 2026.

Lately, I've had AI article fatigue

Lately, every time I open a top-ranked article in the search results, the same thought hits me: "this was probably written by an AI."

Let me state it up front. Search top results are starting to fill up with AI-mass-produced, nothing-special articles. "What is X" / "Pros and cons of X" / "Top 5 X" — open any of them and you find the same H2 headings, the same outline, the same conclusion. You can tell at a glance that you're being fed an AI-flavored article.

And the articles from major publishers are structurally close to this too. Broad and shallow, careful not to take responsibility, smooth on every side. No writer's face. No lived experience. AI articles, and AI-flavored publisher articles — either way, what I actually want to know isn't in there.

The reason I open a search engine has shifted. Not "summarize the broad landscape for me," but "show me the raw record of someone who actually tried this, in this specific situation, with this specific hardware, against this specific error."

The person who jammed Linux Mint onto a Surface and photographed the BIOS. The person who poked at iptables behind a home router to block non-Japanese IPs. Something neither AI nor a major publisher can produce: "scruffy, but unmistakably the record of someone who actually lived through it."

I call this "AI article fatigue." The saturation point with AI-written articles and AI-flavored publisher articles, all at once. Reaching past the major-publisher piece toward an obscure personal blog — that's the same impulse, pointed the same direction.

And here is where the column gets to its real point: this AI article fatigue and the 2026 Google Core Updates are deeply connected.

For Google, keeping "articles AI could write" at the top of search results is corrosive to the search engine's own reason to exist. Users will just go ask an AI directly. For a search engine to survive, it has to surface things AI can't write. The Experience axis in E-E-A-T, and the March/May 2026 reinforcement of "experience and case studies" — these are probably all solutions to the same problem.

And who holds the largest stockpile of "things AI can't write"? Probably — the personal blogs that were told "thank you" ten years ago.

For some reason, this blog gets more traffic from Bing than Google

From here, some of my own blog's data. This blog launched in March 2026 — only three months old. I'll keep absolute PV numbers off the table, but I'll share trends freely.

The monthly PV curve, roughly: March → April down about 30%, April → May up about 2x. March benefited from a one-off Hatena Bookmark burst; April was the cooldown month; May, with no buzz event, still cleared March's peak.

This growth doesn't seem to come from a Core Update boost — no real rank jumps observed. It looks more like the honest result of accumulating articles in low-competition niches. No algorithmic tailwind that I can feel yet.

But the composition of inbound traffic is unusual in a way that breaks expectations.

On this blog, Bing brings slightly more traffic than Google. The market is supposedly 7:1 in Google's favor. We have landed at roughly 1:1, with a slight Bing lead.

Why? Speculation, but I have a hypothesis. What this blog mostly publishes is CVE (security vulnerability) alerts and Japanese domestic service outage reports. The kind of information IT administrators search for during work hours, from a corporate-issued PC.

On many corporate PCs, internal policy forbids installing Chrome. Which leaves default Edge with default Bing — and that becomes the funnel through which they find the blog.

There is circumstantial evidence. Google Analytics referrer logs occasionally show traffic from a corporate chat tool's CDN, a large enterprise group's SharePoint, and an enterprise identity-management service. Pages opened on work PCs, then shared internally over corporate chat with "please read this." The fingerprints are visible.

By contrast, the days when Hatena Bookmark or X buzz hits us, traffic is almost entirely mobile. The desktop-to-mobile split lands near 50:50 overall — against the conventional wisdom of 70% mobile — precisely because of this "work PC by day, private phone by night" dual structure.

There's probably another reason Bing picks up personal blogs early: a mechanism called IndexNow. Articles get indexed by Bing within the same day. Google takes days to weeks to evaluate; Bing puts personal blogs into trial rotation early.

After AI eats search, what's left to read

A wider view. Behind "maybe personal blogs are coming back" is a much larger structural shift: AI is eating search.

Since 2024, Google has been showing AI Overviews (AI-generated summaries) above the search results. Users read the summary and leave without clicking a blue link. This is called zero-click search, and it has been quietly eroding site operators' inbound traffic.

In May 2026, Google rolled out "AI Mode" in earnest. Search itself is converging toward conversational AI. OpenAI is strengthening ChatGPT's web search; Anthropic (the company behind Claude) offers Claude's own web search. The first place users ask a question is shifting from the Google search box to an AI chat input.

In this current, search engines and SEO are probably splitting into two camps.

The middle official source layer is safe; Google will keep favoring primary sources. The top "AI-sufficient layer" is not winnable for individual bloggers anymore. What's left is the bottom "specific-situation experience layer." That's the personal-blog territory, and Google's addition of "Experience" makes sense as a mechanism to rescue exactly this layer.

User search behavior probably shifts too. "Give me a summary" → open AI. "Show me someone's actual hands-on procedure" → open the search engine. The search engine moves toward becoming "a tool for unearthing human work logs."

In SEO terms, this is a big shift. For years, "authority wins on Google" was sacred. Going forward, "authority assumed, plus a layer of experience" becomes the differentiator. For the first time, major-publisher media and personal blogs may stand on the same ground.

A caveat is needed though. The rise of AI has produced an atmosphere where "ChatGPT said so, therefore it's true" is accepted uncritically — a kind of "AI as higher authority" reflex. A different kind of risk — one I touched on in a recent post on this blog. The more search experience leans on AI, the more humans need to keep final judgment in human hands.

One day, I want someone to thank me back

Back to the beginning. To the personal-blog writers who saved me ten years ago, I have been saying "thank you" all this time. The forgotten corner of FC2, the ruins of Hateda diary, the dusty rented-server blogs. Without them, I might not have lasted as an engineer.

Now that I'm on the writing side, I finally understand how precious those "thank yous" were. Write the article, verify the facts, fix typos, wrestle with SVG cover art, click publish. The weight of "thank you" lands differently once you've spent hours on a piece. It's not knowable from the reading side.

In 2026, Google added Experience to its evaluation axis and, with the March and May Core Updates, shifted toward placing "lived experience" and "case studies" above generic summaries. This may be the moment when Google finally caught up with the personal blogs that saved me ten years ago.

It is still hopeful observation, of course. Whether personal blogs really come back will take six months to a year to verify. Whatever ground tremor arrives — Google changing its axis, AI eating search, Bing taking the top spot — I suspect what people want to write doesn't change. Those who want to write will keep writing; those who want to read will keep searching them out.

This blog is three months old. No rank jumps observed. But Bing has become the lead channel, work PCs are reading the site, and there's a slowly emerging texture of "where industry IT admins land when they get stuck." That's the natural result of stacking CVE-alert articles in a low-competition niche; no Core Update tailwind in play yet. I'll stay an observer, and come back here in six months.

By then, I suspect I'll have the same face as those writers ten years ago — scruffy prose, one more post that rescues someone from an obscure error. And if a reader ever silently mutters "thank you" to their monitor, I'll have finally repaid a small piece of what I owe.

I want to be on the receiving end of "thank you." The real reason I started this blog is, probably, just that.