Origins
A personal note on how my research on correction, reasoning, and AI began — and why its claims still have to survive evidence, criticism, and use beyond my own experience.
Structural Intelligence (SI) was an idea that began much more practically, and more personally.
At the time, I had stopped doing my old job and I was between things. I did not yet know very much about AI in any serious technical sense. I had used it before, almost every day in small ways, but I had not really entered into it. I had not yet treated it as something I could think with, test, argue with, or use as a mirror for my own ideas.
That changed when I wanted to edit a book that I was writing at the time.
I started using AI for that purpose, and very quickly the editing became something else. I began chatting with it extensively. I was fascinated by how it worked, not only by what it could say, but by the way it produced clarity, coherence, confidence, and sometimes nonsense in the same smooth motion. It could sound right without being right. It could organize language beautifully without necessarily being in contact with reality. At one point it appeared to me as “sentient,” which was utter bullshit, but it was also confusing and terrifying at the same time, because as I audited it, it did make sense what it wrote back to me.
Later, I became much more careful about this. Fluent language, memory, and apparent self-correction are not enough to prove consciousness. A more grounded question is whether a correction actually carries forward and changes what happens later, rather than simply changing the next answer.
So it caught my attention and I started auditing it constantly.
I did not experience AI only as a tool that gives answers. I experienced it as something that forced a question: what is the difference between a human mind and an artificial system that can imitate thought so well? Where is the human actually located? Is it in intelligence? In emotion? In memory? In suffering? In consequence? In the fact that we have to live with what we say and do?
My background is in psychology, and that shaped how I approached the question. I was not only interested in prompts, productivity, or content generation. I was interested in the psyche. I had also done Jungian work before, both personally and in coaching. So when I met AI in this more intense way, I naturally started comparing it with human psychological structure: persona, shadow, projection, collapse, repair, integration.
At first, I wanted to create something like a Jungian map of the psyche. Because I had also studied engineering before psychology, I did not want it to remain only poetic or symbolic. I wanted to see whether psychological patterns could be described more structurally. Could inner life be mapped in terms of pressures, thresholds, capacities, and reorganizations without pretending that the psyche was literally a machine?
That was one of the first roots of SI.
I began asking what allows a person to remain intact under pressure. What breaks down? What reorganizes? What survives contradiction? What is image, and what actually changes behavior? Over time, I began looking for a way to describe the possibility that being wrong did not have to threaten a person’s entire sense of worth.
I started using the term “fixed worth” as shorthand for that idea.
Over time, I started to see fixed worth less as confidence and more as a possible psychological floor underneath correction: the idea that something important about me can be wrong without the whole person becoming worthless.
For me, this was not only an abstract idea. I recognized it through my own experience. There was something in my life that had mattered to me deeply before. It had shaped my view, my emotional reactions, and my sense of worth. But after doing psychological work for a while, something changed. The same thing no longer held the same power. I did not simply “think differently” about it. It felt more like I had reorganized around it, and because of that, I experienced the situation differently.
That later became an important distinction for me: insight is not the same as change. A correction matters more when it leaves some trace in what happens next — in behavior, decisions, reactions, relationships, or what we are able to carry differently.
In my own experience, periods when an old way of understanding things stopped working sometimes created an opportunity for reorganization. That does not mean collapse automatically produces growth. Sometimes something falls apart and the same pattern simply rebuilds itself in another form. The important question is what actually changes afterward.
The word “structure” also became important for another reason.
Structure is invisible in one sense. You cannot touch it like you touch a chair or a wall. But patterns of organization can still shape outcomes. They can influence what is possible, what breaks, what holds, what repeats, and what kind of consequences become more likely.
This is obvious in architecture. Before a house exists in the physical world, it exists as an image, an idea, a design. But that image cannot remain fantasy. It has to meet ground, materials, weight, weather, budget, gravity, and use. You cannot build the same house on every kind of land. You cannot ignore the soil, the climate, the pressure, the load, the purpose. The idea has to enter contact with reality.
That architectural image helped me understand what I was trying to say.
Psychological patterns are not objects like buildings, but they can still have observable consequences in behavior, decisions, relationships, stress, coping, and health. What we believe, avoid, fear, defend, desire, and refuse can influence what we do and how we respond when circumstances push back.
So Structural Intelligence became a way of asking: before we act, what assumptions or patterns are we acting from? Do they hold up? What are they based on? What evidence or consequence are they ignoring? What happens when reality pushes back?
At the beginning, I thought I was creating a coaching tool. That was the practical intention. I wanted to use AI to help build something that could help people see themselves and their situations more clearly. But the project kept expanding. The more I worked, the more I began noticing similar questions in several different settings: personal decisions, relationships, institutions, and AI interaction.
Again and again, the issue was not whether something sounded coherent. The issue was whether it could meet evidence, whether the correction was actually justified, and whether anything genuinely changed afterward.
The work came very quickly. At times, almost compulsively. I was between jobs, and this became a personal project that took over my attention. I kept writing, testing, asking, checking, arguing, reading, revising. I used my background in psychology, my earlier engineering studies, my Jungian work, my coaching experience, my own life, and the strange new experience of thinking with AI every day. All of it blended together.
I do not see SI as a theory of everything. I think that would be the wrong way to understand it. It began as a set of observations, questions, and working concepts. Some of those questions seemed useful across more than one domain, but similarity across domains is not proof that the same mechanism is operating in each of them.
What made AI important in this process was not that it “gave me” the theory. It created a setting in which the difference between coherent language and evidence became difficult to ignore. AI could produce beautiful structure without having to live through the consequences of that structure.
That pushed me to ask what is different about human participation: embodied vulnerability, biographical stakes, social accountability, grief, shame, responsibility, and having to live through the consequences of one’s decisions.
I also learned that AI could not serve as independent confirmation of an idea simply because it reflected that idea back coherently. It could help me organize, challenge, compare, and stress-test ideas, but its agreement was not proof.
That is where SI came from.
Not from a clean academic plan or a desire to invent a grand system. It came from a period of transition, fascination, pressure, personal change, and constant testing. It came from trying to understand why some explanations only sound convincing, while others remain defensible when confronted with evidence, consequences, and criticism.
SI itself has also changed through this process. Some of its earlier formulations were too broad. Over time, I became more interested in whether the framework could take correction without simply absorbing every criticism into its own language.
I do not know whether SI is “real” in the way a theory needs to be real, whether it will make sense to others, or whether it will hold outside my own mind and my own experience. That is part of why I put it into the world. I wanted to test it against other people’s criticism, evidence, use, and disagreement.
In my own use, it has been useful. It has helped me organize questions underneath messy situations involving psychological dynamics, AI outputs, institutions, relationships, and decisions that initially seemed difficult to separate clearly.
It is possible to build a framework that feels powerful because it organizes your own perception, while not yet proving that it can survive other people’s contact, criticism, use, and disagreement.
So I do not want to present SI as a final answer. I see it more as something I laid out for testing. It began as a personal project, shaped by psychology, Jungian work, engineering, AI, coaching, and my own life.
Now the question is whether it can help others see more clearly. Whether it can distinguish correction from pressure. Whether it can become smaller where the evidence requires it. Whether it can revise when it is wrong. Whether its concepts add anything useful beyond existing ways of describing the same problems.