“What is this?” “Why does this happen?” “How do you even cook fresh chestnuts?” “Is this restaurant actually good value?”
I do not know the answer. Searching manually feels annoying. So I ask ChatGPT.
It answers.
Then I push back.
“But doesn’t that mean this?” “What about the opposite case?” “So what is the actual cost?” “Is that really a good deal?”
The conversation grows.
Behind the scenes, those conversations are cleaned up and transformed into articles.
When I first wrote about this, the rough count was 1,800 articles and around 100,000 lines.
One correction matters:
Not 100,000 articles. About 100,000 lines. The article count is roughly 1,800.
That is still absurd, because I do not feel like someone who sat down and wrote 1,800 articles.
I lived, got curious, and asked because it was easier than researching everything from scratch.
That was basically it.
A normal blog starts with “What should I write?” This starts with life first
Traditional content production usually starts with the article.
Choose a topic. Research keywords. Build an outline. Research sources. Write. Edit. Publish.
This workflow runs in the opposite direction.
Life happens first.
A question appears.
The question goes into an AI conversation.
The discussion branches into more questions.
Only afterward does someone notice, “This is already an article.”
So the flow is not:
content production → life
It is:
life → question → talking things through with AI → article
That distinction matters.
You are not turning your entire life into material on purpose.
You are simply collecting the “I don’t know,” “this is annoying,” and “wait, how does that work?” moments that were already happening.
It feels less like content creation and more like attaching a drainpipe to everyday curiosity—then discovering that an article factory was connected to the other end.
A 3-hour walk can be “just a walk” to the person and several articles to the system
From the human perspective, one evening can be summarized like this:
“I walked around for about 3 hours, went into a restaurant, had a drink, and went home.”
That could be 4 lines in a diary.
But during those 3 hours, small questions keep appearing.
Why does a city feel different late at night? Is this menu actually cheap once all fees are counted? What is solo drinking like? How do the payment methods a restaurant accepts affect where you choose to eat? Why does alcohol feel stronger on some days than others? How useful is it to live somewhere you can simply walk home?
The person is not walking around thinking, “I need ten article ideas tonight.”
They are just walking.
But if each question becomes a conversation, one ordinary evening later splits into several independent topics.
Human: “I went for a walk and had a drink.”
System: “Excellent. We harvested quite a lot today.”
One ordinary evening, two very different reactions.
Months later, “How do I cook fresh chestnuts?” suddenly comes back
The even stranger part happens after publication.
One day, an access-monitoring system reports:
“This page has started getting views.”
The title is something like “How to prepare fresh chestnuts.”
And immediately the reaction is:
“Oh yeah. I remember that.”
Not perfectly.
Not “What is this and who wrote it?”
Just enough memory returns.
There were chestnuts. I had no idea what to do with them. I asked because looking it up manually was annoying. We talked about it. It became an article.
At the time, I was not trying to preserve a memory.
I just wanted to eat the chestnuts.
Then months later, search traffic brings that small moment back.
SEO suddenly starts acting like a photo album.
That is a very strange feature for SEO to acquire.
Photos preserve what you saw. These articles preserve what bothered your brain
A normal album stores visible moments.
A trip. A meal. A view. A person. A place.
An archive built from AI conversations stores something slightly different:
what you were curious about at that exact moment.
How do I cook this? Why is this price structured this way? Why does this system behave like that? Why is this fictional character acting like this? How much work is it really to keep a pet?
Those thoughts rarely deserve a photograph.
They may not feel important enough for a diary.
But asking an AI is effortless, so the question gets recorded naturally.
Over time, the archive becomes less like a photo album and more like a map of your mental footprints.
Not merely:
“What did I see that day?”
But:
“What was catching my attention during that period of my life?”
That can feel more ordinary—and more intimate—than a curated album.
The best part is not remembering perfectly. It is remembering a little
With around 1,800 articles, nobody remembers everything.
Of course things fade.
But they do not become completely foreign either.
A title is often enough to trigger:
“Oh, that was from that day.” “Right, I was wondering about that.” “I remember where my head was at.” “I was looking into that around then.”
That partial recognition is exactly what makes it fun.
If you remembered everything perfectly, you would not need an archive.
If you remembered nothing, the archive would feel like someone else’s work.
The sweet spot is having forgotten just enough for the rediscovery to feel surprising.
And you do not choose when the rediscovery happens.
You are not intentionally browsing old posts for nostalgia.
Someone searches for something. They land on an old article. Traffic moves. The system tells you.
A forgotten question is suddenly back in the present.
It is memory on shuffle.
The strangest part is that the extra effort per article can be close to zero
A number like 1,800 articles sounds like a huge writing workload.
You imagine daily writing sessions, topic planning, long editing hours, and endless publishing chores.
But the source behavior was already happening.
Ask ChatGPT instead of searching manually. Use it to organize a thought. Push back with “but…” Ask for another angle. Keep going until the thing makes sense.
Those conversations would have happened even if there were no article archive.
So the article is both a finished product and a by-product.
That is why making these articles feels different.
The person does not feel, “I completed another writing task.”
It feels more like:
“I lived a normal day, and apparently a few more articles appeared behind me.”
Useful.
Slightly creepy.
Very funny.
This does not mean “publish every detail of your life”
There is an important boundary.
Personal names, addresses, workplaces, exact routes, and other people’s private lives do not need to be published.
What matters to other readers is the question they might have too.
Instead of “I was at this exact restaurant,” the useful version is:
“How should you evaluate value when a cheap menu item sits on top of fixed charges?”
Instead of “I bought these exact chestnuts,” the useful version is:
“What does someone holding fresh chestnuts actually need to know first?”
My experience starts the conversation; I keep what helps someone else answer the same question.
That allows the same piece to be two things at once:
a memory for the person who lived it, and useful information for a stranger.
That dual role is one of the most interesting parts of the whole system.
You thought you were building a blog.
At some point, it quietly became an album.
And the person turning the pages is a search engine plus a stranger you will probably never meet.
