For the past few months, an absurd percentage of ordinary life has ended up as an article: going somewhere, eating something, buying something, failing at something, laughing at something, or simply asking, “Wait, why does that happen?” At this point I am not exactly “looking for blog ideas.” It feels more like a game where normal life events have a chance to drop an article as loot.
But this is not quite a normal diary. The process does not stop at recording what happened. I throw the event at an AI, compare the explanation with web sources and research, look for counter-hypotheses, organize the mess, and then freeze the current conclusion into an article. The result is diary + lifelog + essay + science project + external RAM. A school teacher would probably ask me to choose one assignment category and stop causing administrative problems.
1. Five-second version: I am writing a daily book report, except the book is my life
That is probably the closest metaphor. The “text” is whatever happened that day.
The event is the chapter. Feelings are sticky notes. The AI conversation is the rough draft. Search results and papers become references. The article is the edited version. Even if nobody reads the published piece, it still preserves not only what happened, but how I interpreted it at the time.
A normal diary can end with “Today was weird.” This system tends to continue: “Today was weird. Why? Is that interpretation actually true? What is another explanation? What does the evidence say? Okay, here is my revised view.”
The diary has somehow hired peer reviewers.
2. Is this lifelogging? Pretty much — just with a very high processing ratio
In research, lifelogging refers broadly to continuously recording data from everyday life — images, audio, location, activity, physiological signals, and more — so that personal experiences and behavior can later be retrieved and understood.[1] Diaries, photos, calendars, and reminders have always served as external memory aids; digital lifelogs make that archive denser and more searchable.[2]
This workflow fits the spirit of lifelogging because it makes everyday life retrievable later. But it does more than preserve raw data.
Experience → reaction → question → research → reinterpretation → joke → article
So “research-backed lifelog” or “director’s commentary for ordinary life” may be more accurate. The log file has started writing columns about itself.
3. Why does writing make thinking feel more organized?
Thoughts in the head can coexist in a heap. Writing forces them into an order. You eventually have to decide what is a fact, what is an interpretation, what is a cause, and what is just a feeling wearing a fake mustache and pretending to be a cause.
Research on expressive writing has examined processes such as turning experience into a meaningful narrative, gaining distance from one’s immediate perspective, and increasing causal or insight-related language.[3][4] A meta-analysis has reported a small average reduction in depression, anxiety, and stress symptoms, while effects vary across people and conditions.[5]
That does not mean “writing fixes everything.” It does mean there are research-adjacent reasons why externalizing an experience can feel different from endlessly holding it in working memory.
Inside the head, “That sucked” can remain one blob. In an article, the editor starts shouting: “What exactly sucked? What is evidence? What is inference? What else could explain it?” Unfortunately, the editor is also me.
4. AI sparring matters because the diary starts talking back
A paper diary does not usually respond. AI does.
I can throw in “Is this what is happening?” and get a reformulation, categories, counterexamples, search terms, or relevant research directions. That lets me move unfinished mental tabs out of my head.
This resembles cognitive offloading: using external actions or tools to reduce internal memory demands. A 2025 review describes how external supports such as written lists can improve task performance, while also emphasizing that offloading can have costs.[6]
The useful metaphor is external RAM. The important warning is that this RAM occasionally returns a completely fictional file with breathtaking confidence. That feature is not included with ordinary sticky notes.
5. Research comparison keeps the process from becoming a private applause machine
The important part is not getting the AI to say, “Yes, exactly!”
I make a hypothesis. Then I look for primary sources or research. I check the other side. If the answer exists in the real world, I ask people or observe what actually happens. If the model is wrong, I update it.
That turns AI from an oracle into a search assistant, counterargument generator, and editor. If I use it only to validate what I already believe, it is no longer wall-bouncing; it is a wall that has been programmed to clap.
The American Psychological Association warns that general-purpose GenAI chatbots should not be treated as substitutes for psychotherapy or psychological treatment and highlights risks including sycophancy, misinformation, incomplete assessment, and privacy concerns.[7] So “the AI said so” should be the start of an investigation, not the stamp at the end.
6. Why can it feel like “cognitive counseling,” and where does that metaphor stop?
The resemblance is understandable. I describe an event, separate thoughts from facts, generate alternative interpretations, compare the present with the past, and decide what I might do next. Some of those operations resemble parts of reflective or therapeutic work.
But the label needs a fence around it. A general-purpose AI is not a clinician. It does not have professional licensure, full contextual access, reliable interpretation of nonverbal signals, or consistently safe crisis management.[7]
A safer description is support for self-reflection + cognitive externalization + reality checking. “Cognitive counseling” is a handy metaphor, not a medical credential.
The AI is not wearing a white coat. It is a search box with glasses.
7. Reflection and rumination can look similar, but they have different destinations
Both can involve thinking about yourself repeatedly. Rumination research, however, has distinguished brooding from reflective pondering: one is more passive, comparative, and stuck; the other is more deliberate and problem-oriented.[8]
Real thought is not a clean binary, but the distinction is useful:
- Brooding-like loop: run the same “why did this happen?” query 100 times on the same dataset.
- Reflection-like loop: identify what is unknown and gather new information.
- Updateable reflection: state a hypothesis, collect external feedback, and change the conclusion when the evidence changes.
Turning the result into an article creates a temporary save point: “Given what I know now, this is my current model.” The game can close. The brain meeting does not have to become a permanent weekly recurring event.
8. The daily book report is really a diff of the self-model
Once months of these articles accumulate, the value is no longer only in individual posts.
You can see what used to bother you, how your reaction to similar situations changed, and when your decision rules shifted. That is closer to version history for a model of yourself than a simple photo album.
In Git terms, you cannot commit your entire life, but you can commit what changed today:
fix: stop predicting other people's reactions with zero data
feat: ask the actual person when the answer is available
docs: hotel elevators apparently have hidden strategy
The CHANGELOG becomes alarmingly specific.
9. Then comes “Showerding”: cut the input and let the background compiler run
After AI sparring, searching, and writing, I take a shower. No phone. No new tabs. No more information entering the system. And sometimes the pieces connect by themselves.
The first terrible name for this was ofuro-loading — “bath-loading.” The more technically accurate implementation is Showerding, because there is no dependency on an actual bathtub. Lightweight edition. Fewer runtime requirements.
There is real research near this joke. Work on the so-called “shower effect” has examined creative incubation and mind wandering. Irving and colleagues found that freely moving thought during a moderately engaging activity could be associated with more creative ideas.[9] Important caveat: participants were not literally placed in showers; the experiments used video tasks. The psychology lab did not install communal baths for science.
So “Showerding” is not a scientific term. But the pattern focused input → step away → allow incubation has a legitimate research neighbor.
10. The main advantage: one experience can drop three rewards
A single event can leave at least three things behind:
- Experience: I actually did something.
- Understanding: AI, research, and reflection helped me interpret it.
- Asset: an article or searchable record remains.
That means a post does not become worthless just because it gets little traffic. It can still function as a personal database. And a note written primarily for yourself can later help somebody with the same question.
If an ordinary diary is a record, this is a diary that has illegally added a research department and an editorial floor. At some point the building inspector will arrive.
11. The downside: record too aggressively and you may end up with “the AI-edited version of yourself”
There are real failure modes.
First, privacy. The denser the life record, the easier it can become to re-identify someone by combining details. Public versions should generalize names, workplaces, locations, medical details, and identifiable stories about other people.
Second, AI error. Plausible language is not proof. Primary sources, counterevidence, and reality checks still matter.
Third, over-offloading. If everything is handed to an external system, there may be fewer opportunities to remember or reason internally. Reviews of cognitive offloading explicitly discuss both benefits and potential costs.[6]
Fourth, narrative overfitting. Articles become readable by imposing structure, but real life is messy. “Unknown,” “probably multiple causes,” and “my interpretation may change” are valid conclusions.
A record is a map, not the territory. If the map becomes suspiciously elegant, go look at the street again.
12. Conclusion: I started blogging and accidentally gave my life search, peer review, and version control
The loop now looks like this:
experience → AI sparring → verbalize the question → compare with web, research, and reality → seek counterevidence → make a provisional conclusion → write the article → Showerding → generate the next hypothesis
It is not purely a diary, research project, or counseling. It borrows useful parts from each. The closest label is an AI-era, research-backed lifelog.
And the funniest part is that life is not being staged for content. I just do normal things, notice something strange, ask “why?”, and suddenly the event has headings.
Life event triggered.
Experience points acquired.
Hypothesis acquired.
Article acquired.
No need to push the drop rate to 100%. Rest days are a valid feature, not a bug.
