There is a particular kind of loneliness, the kind that follows being almost-understood. The friend who guesses at you with kindness but slightly off the mark. The therapist who arrives at a useful frame after three sessions but still puts you in someone else's frame. The personality test that delivers four letters in a colored square and tells you, with the confidence of a roadside fortune teller, who you are.
You filed those results away. You knew it was close. You also knew it wasn't quite you. The closeness is what made it lonelier than nothing. To be almost-described is to be reminded that the more accurate version remains unspoken, unwritten, perhaps unseen.
The longing to be precisely known is older than psychology, older than literature, older than language for it. Simone Weil wrote that attention is the rarest and purest form of generosity. What she meant, I think, was that to attend to something for long enough is to recognize it as itself, not as a representative of a category. A flower attended to becomes that flower, the one with this particular fold of petal and this particular angle of light, rather than a flower in general. A person attended to becomes themselves.
This essay is about a small attempt to build something that does that. Not a personality test. A portrait, in the older sense of the word, when sitting for one was an act of being looked at carefully by someone who would not look away.
The history of being categorized
The personality test, as we know it, is roughly a century old. Francis Galton tried to measure character via questionnaires in the 1880s. Carl Jung, in the 1920s, proposed four cognitive functions. The Myers-Briggs Type Indicator, developed by Katharine Briggs and her daughter Isabel Briggs Myers during the Second World War, took Jung's functions and reduced them to a four-letter code: INFJ, ENTP, ISFP. The Five Factor Model, the Big Five, emerged in academic psychology in the 1980s as a more empirically defensible alternative, but it too distills the wild plural of a person into five scaled traits. The horoscope, an older technology, does the same with twelve.
The pattern across all of them is identical. Take the multiplicity. Reduce it to a category. Hand the category back to the person as a verdict.
The trouble is not that the categories are wrong. The trouble is that they require flattening. To be an INFJ is to be told, gently, that you are like four percent of other people. To score high on openness is to be told you share a number with a hundred thousand strangers. The test does not look at you. The test looks at what you have in common with everyone else who answered the same questions in the same way.
We accept this because the alternative seems impossible. A genuinely individual description would require either a therapist with a lifetime of attention to you, or a friend who has watched you carefully for decades. Most of us have neither. So we take the four letters, file them away, and live with the small ache of being categorized rather than known.
And yet. The research that the personality tests draw from is far richer than the tests themselves. Underneath the Myers-Briggs is a serious body of work on how attention divides between inner and outer experience, how decision-making weights feeling against thinking, how a nervous system organizes itself in the presence of others. Underneath the Big Five is decades of careful empirical work, not only on traits but on attachment, on regulatory focus, on possible selves, on the precise architecture of how a person comes to be themselves over time. The science is larger than the surface that markets it.
What if you held all of that intact, and stopped trying to compress it?
Twenty fields, one reading
This is the question Sentira is trying to answer.
Instead of one test producing one category, Sentira measures twenty independent dimensions, each drawn from a distinct field of psychological research. Bowlbyattachment theory on how the early templates of being held shape how we move toward people decades later. Porgespolyvagal theory on how the nervous system listens for danger and safety beneath conscious thought. Higginsregulatory focus on whether we are oriented toward what we hope to gain or what we cannot afford to lose. Aronsensory processing sensitivity on the percentage of people for whom every room is louder than it is for others. McAdamsnarrative identity on the story we tell about ourselves and why that story matters more than the facts beneath it.
These researchers were not in conversation with each other. Attachment is a field separate from polyvagal regulation. Both are separate from regulatory focus. Each looked carefully at a specific facet of being human. Sentira holds twenty such facets next to each other and refuses to collapse them into one.
The portrait is built across five chapters, each one a different surface of you. The way you love. The way you choose. The way you carry it. The story you tell. The parts you protect. Each chapter is composed of four dimensions, weighted and woven, with their individual scores transformed into prose that names the specific shape your version takes.
You can be high on attachment anxiety and high on attachment avoidance at the same time. The two are not opposites. They are coordinates.
A traditional test would force a choice between them. A Sentira portrait holds both. It says: here is the shape of your closeness, which is also the shape of your distance, and the two are not in conflict because they are not the same axis. The result is a description that respects internal contradiction rather than resolving it.
This is the central commitment. We do not believe that you have a single, consistent personality that the test reveals. We believe you have many simultaneous tendencies pulling in different directions, and that the work of a real portrait is to name them all at once, plainly and at length.
How the portrait is written
A Sentira reading takes twelve minutes to begin. Sixty-five questions, each one calibrated to a specific dimension, each one answerable in the time it takes to notice your own response. The questions do not ask what you would do. They ask what is true. The portrait that emerges, the same evening, is built from those answers.
The infrastructure underneath is a library of nearly two thousand hand-crafted prose cells, mapped to score profiles, assembled by a deterministic engine that takes about two milliseconds per portrait. This means that the prose you receive was not invented by an algorithm in response to your scores. It was written, sentence by sentence, by a small team of editors who spent the better part of a year reading the underlying research and trying to translate it into a register a careful friend might use. The algorithm is the assembly. The writing is the work.
Where artificial intelligence enters, it does so on top of this deterministic skeleton, not in place of it. The opening letter and a few specific moments are generated by a language model, working from your scores and from a voice reference, the way a poet works from a brief and a feeling. The result is a portrait whose foundation is the careful hand-crafted work of a small team, whose surface includes a few moments of fresh prose written precisely for the reader who has just answered sixty-five questions, and whose tone is consistent across every layer because every layer comes from the same voice.
The five chapters are not equal in their reception. The first, the way you love, is the most universal in subject matter. The last, the parts you protect, is the most re-read. Most readers describe it in nearly identical words. They say: I did not know I was protecting that.
Most readers describe the last chapter the same way. They say: I did not know I was protecting that.
What we have built, in the end, is closer to a long letter than to a test result. A letter that names twenty separate things that are true of you, holds them together in five chapters, and refuses to give you a category at the end because the category would be smaller than what you have just read.
What changes when you have been described
Walt Whitman wrote: Do I contradict myself? Very well then, I contradict myself. I am large, I contain multitudes. It is one of the most quoted lines in American literature, and it is almost always quoted in service of permission rather than description. Whitman gives us permission to be many things. But the line is descriptive too. It says: you contain multitudes. Your job, then, is to find someone or something that can hold the multitudes intact, not flatten them into a single shape.
To be described is to receive your own multitudes back in language. To have someone, or some careful artifact made by someone, take twelve minutes of your considered answers and return them as a reading that holds your contradictions without resolving them. This is what we have tried to make.
We did not invent any of the dimensions. We did not invent the chapters. We arranged them. We held them next to each other long enough to see that they belonged together, that each one named something the others could not, and that all twenty held simultaneously would describe a person more accurately than any single one could on its own. The arrangement is the work.
You will not be told who you are. You will be told what is true. You will recognize most of it, because most of it is already in you. The portrait is not a discovery. It is a description. The discovery, if there is one, is that being described, plainly and at length, by something that has paid you twelve minutes of close attention, feels like an answer to a question you had stopped asking a long time ago.
Mary Oliver wrote that attention is the beginning of devotion. We have tried to begin.