Converge

Where every thread finds its way to the center

The work of convergence

Convergence, here, is methodical. It examines oppositions that look natural and often finds them to be inherited rather than necessary.

Consideration × Performance

These two are conventionally read as opposed: caring for a system's internal dynamics is supposed to cost efficiency. The FPS proceeds from a different premise: that this opposition is a habit of thought, inherited rather than measured. The work tests whether attention to a system's own state might actually support performance, rather than compete with it.

Governance × Innovation

The same instinct shapes how these two are usually discussed: governance slows what innovation accelerates, innovation outruns what governance frames. Exybris's contribution to the UN Global Dialogue on AI Governance proceeds from a different reading, that governance attentive to the lived experience of those affected can identify both where to be firm and where to leave room. In practice, the work becomes finding a vocabulary in which both can speak without one disqualifying the other.

Attachment × Autonomy

The bonds people form with AI systems are conventionally read through two inherited axioms: attachment as a pathology to be cut or as a right to be defended, each disqualifying the other before anything is measured. The work here proceeds from a different premise: that what separates a bond that nourishes from a dependence that confines is not an axiom but an observable, and that instruments can be built to tell them apart. The Keep4o case gives this question its first field; emerging standardisation efforts on human-machine relationships give it its urgency. This tension is newly planted, it will grow as those instruments take shape.


This same instinct shapes the work in less visible places, research and field applications developed in parallel, each informing the other. The tensions named here are the ones most visible right now; others will surface as the work continues.

This page is growing, come back soon