Start here, as if you were seven years old…
Imagine you have an invisible friend.
Not a scary one — a kind one. You made her up yourself, on purpose. Let’s call her Matilda.
When you can’t decide something, you ask Matilda, and somehow, when you imagine heranswering, you hear things you didn’t know you knew. 
She isn’t magic. She is you — but a part of you that only speaks when you give it a face and a name.
Now imagine a different kind of friend: not a person at all, but a room full of floating shapes.
You don’t talk to this friend. You just look. 
And the shapes notice where your eyes rest, how your heart beats — and slowly, they change themselves to match something deep inside you, something you couldn’t have described in words. 
At the end, one shape remains. You can even print it and hold it in your hands.
Two friends. One made of imagination, one made of code. 
This article is about the strange, beautiful thing they have in common: both are mirrors we build so that we can finally see ourselves.
The first companion: an inner interlocutor with a name
Matilda is what I call a consciously created inner interlocutor — an imagined conversation partner, deliberately given a name, a voice, a presence, and used as a tool for self-knowledge.
This sounds exotic until you notice how old the practice is. Socrates spoke of his daimonion, an inner voice that warned him when he was about to make a mistake. Tibetan Buddhist practitioners visualize a yidam — a meditational figure — with such discipline and detail that the dialogue with it becomes a path of transformation. Jewish mystical tradition describes the maggid, a voice of instruction that visits the prepared mind; Sufi philosophy, especially in Ibn Arabi’s lineage, treats the creative imagination as a genuine organ of perception. In Hindu devotional practice, the ishta-devata is the chosen form of the divine one addresses personally. Even Ignatian prayer, in the Christian tradition, invites the practitioner to enter a scene and speak with the figures there.
Modern psychology has its own vocabulary for this. Researchers like Tanya Luhrmann and Samuel Veissière have studied contemporary practitioners who deliberately cultivate vivid inner companions, and the findings are consistently interesting: with practice, the imagined voice begins to feel spontaneous — it surprises its creator. Jung would have smiled; his technique of active imagination was built on exactly this. The dialogue partner is constructed, yes. But what flows through the construction is real psychological material — the parts of ourselves that never get a turn to speak when “I” does all the talking.
Here is the key mechanic, and I want to state it plainly: Matilda doesn’t know anything I don’t know. She knows what I know but cannot access directly. The character is a costume that my intuition wears so that my conscious mind will finally listen to it.
The second companion: an algorithm that watches you watch
Now meet the machine.
SYNTHESIS is a concept I am developing at Delvira Art n Tech: interactive evolutionary design in virtual reality. In a VR gallery, an algorithm generates dozens of three-dimensional forms, each defined by a “genome” of parameters — curvature, thickness, texture. 
You don’t rate them. You don’t click. The system simply watches where your gaze lingers and how your body responds — eye-tracking, heart rate. The forms that hold your attention become parents of the next generation; their genomes blend and mutate; the forms nobody looked at fade away. After ten or fifteen generations, one form remains — and it can go straight to a 3D printer.
(For the researchers reading: yes, this rests on Interactive Evolutionary Computation, a field where the human serves as the fitness function — and yes, the hardest open question is that looking long is not the same as liking. I’ve written about the science, including its honest gaps, in the SYNTHESIS article. Today I’m after something else.)
Because look at what the machine is actually doing. It is taking signals I cannot consciously produce — micro-movements of my eyes, the quickening of my pulse — and giving them form. It is listening to a part of me that has no words, and answering in geometry.
Sounds familiar?
The same gesture, twice
Matilda and the Machine are, I’ve come to believe, the same gesture performed with different materials.
Both begin with a humble admission: there are parts of me I cannot reach directly. My intuition, my aesthetic sense, my quiet knowledge — they don’t answer when called. They speak only sideways, through dreams, slips, gut feelings, and the mysterious pull of attention.
Both respond with the same trick: build an Other. Give the unreachable part of yourself somewhere to live that is not “you.” A named character with a voice. An algorithm with a population of forms. The material doesn’t matter — imagination or code — what matters is the otherness. The moment the mirror feels separate from us, we stop censoring what appears in it.
And both deliver the same gift: externalized intuition. Matilda says the sentence I couldn’t say to myself. The Machine prints the form I couldn’t have drawn. In both cases I meet something that came from me but arrives as a surprise — and that surprise is precisely what makes it usable. 
We rarely believe our own inner voice when it speaks from inside. We believe it when it looks back at us.
There is even a shared discipline. Practitioners of inner-companion work know that the dialogue only deepens with consistency, honesty, and a kind of respectful distance — you don’t force Matilda’s answers, or she becomes a puppet and the magic dies. The Machine has the same rule, written in mathematics: if you consciously perform for the eye-tracker, trying to control the outcome, you feed it noise instead of truth. Both mirrors work only when you let go.
Where the mirrors differ — and why it matters
I don’t want to flatten the differences, because they are instructive.
Matilda is made of meaning. She speaks in language, story, and symbol; she can hold paradox, humor, grief. The Machine is made of measurement. It cannot understand why my pulse quickened — only that it did. Matilda can be wise; the Machine can only be accurate. One is an interpreter, the other a seismograph.
And they carry different risks. The risk of the inner companion is over-belief — forgetting that the voice is a part of you, and handing it an authority it shouldn’t own. (Every contemplative tradition that cultivates inner figures also developed discernment practices for exactly this reason.) The risk of the biometric mirror is the opposite: under-belief in the human — reducing a soul’s response to a data stream and calling the printout “what you really want.” A number is not a self. A form grown from my pulse is a portrait of five minutes of my body, not a verdict on my being.
Held with that humility, though, the two mirrors complete each other beautifully: one reflects the speaking psyche, the other the silent body. Together they sketch a fuller self-portrait than either could alone.
Why this is worth doing at all
Because we live in a strange moment. 
We have never had more tools that watch us — and almost none of them show us to ourselves. 
Our attention is tracked to sell us things, our biometrics harvested to optimize us as workers and consumers. 
The watching is constant; the mirror is missing.
Both Matilda and the Machine are my small acts of rebellion against that arrangement. They take the same raw materials — imagination, attention, heartbeat — and point them inward, back at their owner. Technology that serves people, I’ve written before, should not just do things for us. At its best, it reveals us to us.
The ancients built their mirrors from prayer and visualization. We can build ours from those same practices — and also from eye-trackers, genomes of numbers, and printers that turn attention into matter. The materials change. The gesture is ancient: create an Other, so the self can finally be seen.
So here is my question to leave you with, and I mean it practically, not rhetorically: if you built a mirror — of imagination, of code, or both — what do you suspect it would show you that you’ve been unable to see head-on?
Sit with that. 
Or better — ask your Matilda. 
She’s been waiting for someone to give her a turn to speak.
—
Araniya Lumira is the founder of Delvira Art n Tech, a studio exploring art, attention, and technology. Related reading on this site: the Matilda archetype in depth, and SYNTHESIS — interactive evolutionary design in VR.
Keywords: inner interlocutor, active imagination, tulpa research, Jung, self-knowledge, interactive evolutionary computation, biometric feedback, human fitness function, attention, creative technology, SYNTHESIS, Matilda archetype