01
Continuity and memory
What should persist, what should evolve, and what must remain outside the model so that continuity can survive provider and context changes.
Research
Research that beginsafter the demo.
The useful questions appear across returns: after a good response, after a rupture, after memory changes the next exchange, and after someone stops speaking as if the interaction were temporary.
Lines of inquiry
Six parts of the same unresolved problem.
02
Conversation and initiative
What gives an exchange movement, curiosity, timing, and tension without turning the system into an interrogator or an attention machine.
03
Rails and trust
How refusals, warnings, tonal shifts, and safety behavior alter expectations beyond the response in which they appear.
04
Private and adult interaction
A relational stress test for ambiguity, desire, shame, boundaries, consent, and language that becomes less socially edited over time.
05
Voice and presence
What changes when silence, interruption, pacing, turn ownership, and ambient presence become part of the interaction contract.
06
Model independence
How identity and relational continuity can remain coherent when the model, provider, policy behavior, or delivery surface changes.
Field Notes
Evidence from inside the work.
These are not finished theories. Each note begins with an observed behavior, records the interpretation that followed, and ends with what remains unresolved.
Field Note 01Architecture
The model is not the relationship
Why continuity, identity, and trust cannot live entirely inside one context window, provider, or model personality.
Read the Field NoteField Note 02Continuity
Designing for return
The first exchange can impress. The relational product begins when someone comes back and the previous interaction has consequences.
Read the Field NoteField Note 03Private expression
The private interior of human–AI interaction
Why adult and private conversation exposes relational behavior that ordinary demos and socially edited exchanges often hide.
Read the Field NoteChapter zero
The Intimacy Code was the first attemptto name the question.
Published before Xiara and before the current depth of model testing, the book explored technology, desire, and emerging forms of human connection.
I now see it as an early document from a more superficial stage: useful not because it settled the subject, but because it opened a line of inquiry that later became technical, behavioral, and architectural.
Open the English editionXiara is where these questions become observable product behavior.
Explore Xiara