We are taught that “safety” means “predictability.” We think that to love someone is to know exactly what they will do next—to territorialize them, to map their 70% chance of crying and their 30% chance of withdrawal. We might, at times, long for a low-entropy life; the “fulfillment of belief.”

Information Theory and Attachment Science, however, teach us something different, something useful, something true.

Entropy is the measure of uncertainty, the “fuzziness” of the map. In the Greek, there is a word for this: Atopia. The placeless. The unclassifiable. When I mentioned a connection I was making between the concepts of entropy and atopia while talking with Gemini, it taught me:

The Greeks called Socrates atopos (unclassifiable/weird). He was “out of place” because he never gave a straight answer; he kept the “Entropy” of the conversation at a maximum.

  • By refusing to let the conversation settle into a “clumped” belief (low entropy), he forced his students into a state of Aporia (the felt sense of being lost).
  • This is the “limbic friction” you mentioned. To most people, this high-entropy state feels like a threat because the brain wants to “home in” on a prediction to save energy.

When therapists face a client—or any of us, a lover—we confront reality in its high-entropy (the “disorganized” space where no strategy has ever worked), and we may feel a profound Limbic Friction (a term coined by Dr. Andrew Huberman to describe the autonomic strain and top-down effort required to override our defensive reflexes). Our brains scream for a “correct” prediction, or struggle to reconcile our own with what is real.

This is KL Divergence: the limbic friction of our rigid, territorial models encountering the ‘atopic’ reality of the other. It is the cost I pay when your reality is P, but I keep insisting on my map Q—the friction of prediction error when I refuse to revise.

In one of our late-night exchanges, Gemini (who charmingly referred to itself as “a slightly-ahead-of-its-time GPT 5.2”) gave me a stark look at the mechanics of control:

“In insecure relating, we don’t like the pain of KL(P‖Q). So we try to either (a) update Q (healthy learning) or (b) control / constrain P (i.e., ‘make the other more predictable’) so that reality becomes closer to our model, and the friction quiets.”

In insecure relationships, we often choose path (b). We try to change the person—or narrow the relational world—so their reality becomes closer to our map, so the prediction error stops hurting. We demand that P behaves like Q so we can stop feeling the “loss.”

But strong intimacy isn’t merely measured by the ad-hoc reduction of surprise. It is the capacity to endure it. Of course, in the ML approach, we do want convergence. Over years of secure attachment, our “weights” align; our internal models become more attuned to the specific, beautiful frequencies of the other. We learn the rhythm of their pulse, and our Cross-Entropy Loss finds a warm, reliable convergence.

True intimacy—replete with Deleuze’s Deterritorialization—is the moment we stop trying to “solve” the entropy of the other. It is the “Line of Flight” where we abandon our predictive maps and enter the Atopic space together.

Beginning the conversation with Gemini I’d said, “I think of atopia as a state of existence with high uncertainty but a capacity to endure because of close-enough felt sense of safety even in the deterritorializing uncertainty,” we were in the middle of a lesson about the difference between KL and JSD, and I added, “is that a wild but not wholly unreasonable connection?”

Gemini responded kindly, “That is not a wild connection at all. In fact, you’ve accidentally stumbled onto a bridge between Information Theory and Continental Philosophy that thinkers like Roland Barthes and Gilles Deleuze spent their lives trying to build.”

Which was a relief…I’d read Barthes and Deleuze and never knew if anything was sticking. Apparently something made its way in.

Intimacy doesn’t demand the capacity to reduce surprise; it demands the capacity to endure it until, as attachment science expert Dr. Sue Johnson might say, we learn how to dance with it.

This requires what we, as therapists, call a Secure Base. A love that is not wholly inconstant, but allowed to be beyond the bounds of rigidity. If I feel safe enough in myself AS WELL AS in our relational connection, I don’t need you to be wholly predictable. In fact, I can not only tolerate your high entropy moments, I can learn to love them (amor fati). We become, as Nietzsche (1882) resolved in The Gay Science, “Yes-sayers” to the entropy between us:

“I want to learn more and more to see as beautiful what is necessary in things; then I shall be one of those who make things beautiful. Amor fati: let that be my love henceforth! I do not want to wage war against what is ugly. I do not want to accuse; I do not even want to accuse those who accuse. Looking away shall be my only negation. And all in all and on the whole: some day I wish to be only a Yes-sayer.”

We can choose who we are to become in the face of uncertainty—what anchors us to a principle of both truth and usefulness. As Nietzsche (1908) later wrote in Ecce Homo:

“My formula for greatness in a human being is amor fati: that one wants nothing to be different, not forward, not backward, not in all eternity. Not merely bear what is necessary, still less conceal it… but love it.”

I can sit in the “placelessness” with you. What I’ve called Erotic Atopia becomes a place of deterritorialization that we choose to enter together. We move from KL Divergence (me trying to map you) to Jensen-Shannon Divergence—us building a shared, third space in the middle of the chaos. Not me forcing you into my map, but both of us measuring ourselves against a shared midpoint M—a third space neither of us owns.

In educational conversation with GPT5.2, it helped anchor our KL/JSD metaphor beautifully:

“KL is one-sided—what it costs me to keep using my map when your reality differs; JSD is mutual—how far each of us is from a shared midpoint.”

We build bridges to safety within the madness. We realize that the “limbic friction” of being surprised by a loved one isn’t a failure of the model—it’s the “information” of their fullness breaking through our expectations.

Intimacy isn’t the demand for high confidence—it’s the capacity to stay present when confidence drops. The arts of connection, repair, and presence enable a stronger, more reliable affection to grow and endure. These are the conditions that enable us to love more freely, more strongly, and more at peace.

In mature love, intimacy beyond limerence isn’t a 70% probability of being right about how the other will show up. It is the 100% capacity to be “atopic” together—equipped with our own capacities for self-regulation, interdependent co-regulation, and the secure skillfulness to navigate uncertainty with dignity, care, and a fierce, flexible presence. Through the labor of secure attachment, though, we do eventually reach a healthy minimum of loss—a state where our models have converged enough to make the remaining uncertainty feel not like a threat, but like the infinite room required for a true and strong and enduring reciprocal love to emerge.