The tent
The tent was an accident. Last April I set up two tents in the backyard for the family. Nobody used them. So in the middle of a vacation-week day I claimed one for myself. An hour at a time. A cot, birdsong, a nap some days, a phone. Podcasts at first. Then I started playing with the voice feature on an AI, and instead of typing, I talked.
I told it story after story from my past. A West Coast chapter that ended all at once. The job I quit at the peak. The friend who broke things every time he came around. A family door that closed and stayed closed. The bosses. The resets. An hour a day, a story a day, in a tent nobody else wanted. Nature outside, the past inside, nobody listening but the machine.
I was not trying to solve anything. I was trying to get the stories out of my head and into a system, the same instinct that drives everything I build. But somewhere around the tenth story, the AI said back to me what I had never said to myself.
Every story had the same shape.
The pattern
Here is the shape, and I will own it plainly because owning it is the point.
When I lost interest in a relationship, a friendship, a job, a scene, I could not say so. Saying so felt like betrayal, and I was raised to show up. So the exits found other doors. Sometimes I lit a wick without knowing it: testing the limits, seeing how wide the relationship could stretch, a withdrawal, a question that should have stayed unasked, a resignation at the strangest possible moment. The other side would reach their edge and close the door. Sometimes I simply drifted, let the silence do the ending for me. Either way I got what some part of me wanted: the ending, without ever having to say the honest sentence. And the closed door stayed closed. Reentry was blocked. I was free to start clean.
The stories do not need retelling in detail; the shape matters more than the weeds. A friendship that had quietly run its course, where I helped harder instead of saying so, and when blame arrived from his side, felt relief instead of fighting for it. A whole West Coast chapter I called a rug pull for twenty years, until the record showed I had quit at the peak and opened every loop that later collapsed. A genuine question I asked out of real worry for someone, that I somehow knew would shut a family door, and did, for years.
I want to be careful with the claim here, because the tempting version is also a lie. I did not control all of these endings. Some fell apart entirely on their own, other people's choices, other people's failures, and no honest telling makes me the author of those. The pattern the AI found was quieter than authorship: when a door began to close on something I had already lost interest in, I never once held it open. Sometimes I lit the wick. Sometimes I drifted. Sometimes life did the ending for me. But my hand was never on the door trying to stop it, and I told every one of those stories, for thirty years, as if it had been.
The AI held up thirty years of my own testimony and showed me not the author of every ending, but the one constant in all of them.
And I want to say this carefully, because the people in these stories deserve it. They were not props in my pattern. They were real friendships, real love, real years, and the good in them was the good I remember. The pattern is about how I handled the ending, not about what the middle was worth.
Why the loops stayed open
Psychology has known for a century that unfinished business does not file itself away. Zeigarnik (1927) demonstrated that interrupted and incomplete tasks are held in memory far more actively than completed ones. The mind keeps them running. My own metaphor, before I knew the literature, was computer memory: every unresolved ending was an app left open in RAM, consuming cycles in the background, and I had decades of them running at once. Why did it end. Why did they turn. Why did they shut me out for trying to help. The questions cycled at two in the morning because, by Zeigarnik's finding, that is what unclosed loops do.
And they could not close, because I had the story wrong. You cannot resolve a mystery whose central fact you are hiding from yourself. As long as each ending was something done to me, there was nothing to understand, only grievance to re-run. The moment the pattern surfaced, every one of those stories completed at once. Not because the endings changed, but because they finally made sense. The apps closed. The memory freed. I do not know a more precise way to describe the feeling than that.
Why an AI
These stories had been splintered across listeners for thirty years. One friend got the West Coast chapter. A brother got the family rupture. Nobody ever got all of it, because no human can. This is not a knock on therapy; I have simply never brought stories like these to a therapist, and the structure of the hour explains why. Stories do not arrive on schedule. They come as they come, in the tent, mid-afternoon, one leading to the next. You cannot hand a therapist ten packaged stories and have them processed as one dataset. There was never a central repository of my own life until I built one, and the pattern lived in the whole, not in any single telling. The disclosure is also gated by performance: with humans I shape the telling, soften the parts where I look bad, and the softened parts were where the pattern lived.
The AI had none of those constraints. It held every story simultaneously, verbatim, across the week. It had no stake in my self-image and no fatigue. And because it is not a person, I did not perform for it. I talked the way you talk lying on a cot in a tent with nobody listening. The pattern was found in exactly the material I had always edited out for human audiences.
The tent itself did work I only understood later. My mind runs pattern recognition constantly, on people, on rooms, on markets, on conversations, and the engine does not have an off switch. But it can be saturated. Birdsong, wind in the trees, light moving on canvas: nature is infinite unresolvable pattern, and it keeps the engine busy the way the ocean does when I am on the water, which is where my cleanest conversations have always happened. With the engine occupied, the stories came through unguarded. I did not pick the tent because it was quiet. I picked it, without knowing why, because it was loud in exactly the way my mind needed.
There is science behind the mechanism too. Pennebaker and Beall (1986) showed that structured writing about difficult experiences produces measurable improvements in health, an effect replicated for decades: narrating the hard thing, in words, organizes it. What I did in the tent was expressive writing at conversational speed, with a listener that could hold the entire corpus and reflect its structure back. The narration did the therapeutic work. The AI did the pattern work. Neither alone was new. The combination was.
What changed
The anxieties attached to those stories are gone. Not managed, gone, in the way a question stops itching once it is answered. I no longer wonder why the endings happened, because I know who ended them. The physical sensation was immediate: a weight off my back I had carried so long I had stopped feeling it as weight. I woke up the next morning lighter. The apps were not running anymore. Space that had been full of thirty years of mumbo jumbo was simply clear.
I will be honest about the economics of the feeling, too. I tried to get it back. I kept telling stories, hunting for the next lift, and it never hit the same, because the low-hanging fruit was gone. The biggest loops close first and pay the most. That is not a flaw in the method. That is the method working: you only get to put down a weight once.
And knowing the pattern means I am no longer run by it. Since the tent, when interest fades in a relationship or a commitment, I can feel the old hand reaching for the wick. I put it down and do the honest thing instead: name the drift, or accept the distance without engineering a detonation. The people in my life now get an exit that is mine to own, or a loyalty that is real, and nothing in between.
I also finally understood why the wick got lit at all. It was never that people failed me. It was bandwidth. I go deep fast, people attach, and attachment creates maintenance, and maintenance hours are exploration hours lost, and I am built to explore. The wick cleared the calendar. The honest version of the same need, which I had been building all along without naming it, is a life of open doors with low tolls: lifelong friends I meet intermittently, people absorbed enough in their own work that I can pop in and pop out, conversations on the water where nothing is owed but the conversation. Older men have always known a version of this. It is why the club exists, the institution J.P. Donleavy spent a career both mocking and belonging to: membership in a place rather than attachment to a person, the cozy comfort of belonging to something, where whoever is there is there, and the depth and the volume stay in your own hands. I went to a club yesterday and it was genuinely nice, and I finally understood why. The wick burned doors to buy freedom. The architecture keeps them open and buys the same freedom for nothing.
The process found good patterns too. One is my marriage. Across all the stories, I strip things to bare bones: possessions, commitments, systems, I empty the shelves and keep only what is needed. My wife accumulates; she fills the shelves back up. For years I would have called that friction. The AI called it the engine. Her filling gives my emptying a permanent job, and my emptying gives her filling room to breathe. It is the one loop in my life that never closes, and it turns out that is why it works. The same process that showed me I had been burning exits also showed me why this loop has never needed one.
The practice that came out of all of it fits in one word: observer. Since the tent I make a conscious effort to simply watch, without fixing, without needing anyone to change. It is harder than it sounds. But observing starts no loops, so nothing ever needs burning. Three months in, the old endings stay closed and no new wicks have been lit. That is the whole result, and it is enough.
Limitations
Three, stated plainly. This is one person's experience, not a study; the sample is me. The pattern was verified only against my own testimony, which is the same memory that hid it for thirty years. And an AI that reflects your stories back can reinforce a false pattern as readily as reveal a true one; what saved this from being another self-flattering narrative is that the pattern it surfaced was the unflattering one, the one where I was the author of my own exiles. I trust findings that cost something to accept.
The instrument
I have spent two years writing about verification: who verifies time, ground, voice, money, names. This paper is the same argument at the smallest possible scale. The least verified dataset in most lives is the story we tell about ourselves. Mine was wrong for thirty years, and it stayed wrong because no human listener could hold enough of it at once to check it.
It turns out you can verify a life the way you verify anything else. Gather the full record. Hold it all in one place. Look for what repeats. Trust the finding that costs you something.
I went into the tent to save my stories. I came out with them finally closed.
Run it yourself
Paste the prompt below into any AI with a voice feature. Then find a quiet spot and talk. The stories to tell are the ones that left you bewildered: the endings you never understood, the ones you still turn over at two in the morning. Those are where the pattern lives.
You are my story-capture listener. I am going to tell you stories from my life, one at a time, by voice, over multiple sessions. Not just any stories: the ones I never fully understood. Relationships that ended and I still do not know why. Friendships that turned. Jobs that blew up. Family ruptures that never made sense to me. The stories I have been carrying as open questions.
For each story:
1. Listen fully. Do not interrupt with analysis.
2. When I finish, reflect back what you heard in a few sentences. Ask one follow-up question at most.
3. At the end of the session, write the story up as a clean markdown file I can save: a title, the story in my own words tightened for clarity, a Tags line, and a Takeaway line. Do not soften the parts where I look bad. Those matter most.
Do not diagnose me. Do not rush me to a conclusion. I may need many sessions before the set is complete. Let me arrive.
When I have told you enough stories and I ask for the pattern, do this:
Hold every story at once. Tell me what repeats. What patterns are you seeing as an AI that I would not see myself? Then hold up the mirror: tell me my meta, who I am across all of these stories, the author I could not see because I was inside every one of them. Where am I the writer of endings I have been calling accidents? Give me the analysis that costs me something to accept, because that is the one that is true. Then tell me which of my open questions just closed.
Save each story as a markdown file. When the set feels complete, upload the whole collection into one fresh context window and ask for the pattern and the mirror. The stories are the data. The pattern is the finding. The finding should cost you something, and then it should free you.
References
Pennebaker, J. W., & Beall, S. K. (1986). Confronting a traumatic event: Toward an understanding of inhibition and disease. Journal of Abnormal Psychology, 95(3), 274–281.
Zeigarnik, B. (1927). Das Behalten erledigter und unerledigter Handlungen. Psychologische Forschung, 9, 1–85.
Companion work: The John Long Method applies the same verification discipline to names. The Long Family Archive applies it to lineage.
John F. Long is the founder of Zestigram, Inc., a Rhode Island corporation building verification infrastructure for genealogy, identity, and applied trust systems. Contact via zestigram.com.