Contact

HUMAN to SYSTEM relationship models

A working library of recurring human behaviour and how systems could adapt to them. This is a working library, updated as mechanisms are understood and developed.

Decision Compression
People recognise what they want when they see it but cannot describe it in advance. Declared preferences are unreliable proxies for revealed ones.
Constructed Insight
Understanding arrives through guided inquiry. The right questions create the conditions for insight to surface.
Autonomy Gradient
The degree of agency delegated to system could only increase as trust increases.
Preference Accumulation
Preferences are situational. Taste is more stable. Systems should learn how people choose, not just what they choose.
Cognitive Scaffolding
People often need help thinking, not answers. Systems can help with strengthening thought rather than generating output.
Elimination Sequencing
When constraints are objective and self-known, preferences can be declared sequentially.
Tiebreak Resolution
The user has a shortlist and cannot move. The deciding constraint exists but hasn't been named yet.
Escalation Intelligence
A model for designing the conditions under which a system recognises its own limits and returns control to a human.
Intent Extraction
People ask for solutions instead of expressing goals. The real objective is rarely what the surface request suggests.
Judgment Amplification
A model for improving the quality of expert decisions without displacing the expertise that makes them valuable.
Progressive Capability Reveal
A model for exposing system sophistication in proportion to a user’s demonstrated capability.
Attention Compression
A model for reducing cognitive load in information-dense environments without reducing the awareness needed to act.
Shared Deliberation
A model for designing systems that support consequential decisions requiring more than one perspective.