ORBI+ / RESEARCH
Working notes on what ORBINTU is trying to solve: context, wrong work, the human–machine boundary, verification, persistent intelligence, and learning from outcomes.
Filter by area. Open an entry for why it matters and why it is hard.
What must a system persistently know about a person’s working world to reduce repeated explanation, without absorbing private state it should not hold?
How should a Need encode current reality, desired reality and constraints so work can be checked against it later?
How does a system decide which part of a person’s world is relevant to a Need without over-collecting?
How can intelligent systems measure whether the work being executed should have existed in the first place?
How should a requirement be bridged to a model, tool, browser, API, database, software, person or future capability?
When does a request contain enough understanding to become bounded, executable work?
Separating capability from participation preference in personalized work systems.
How can delegation preferences and verification patterns be learned while staying inspectable and revocable?
How can shared work coordinate several personal intelligences while keeping private state distinct?
How can autonomy increase only where trust has been earned, and how is that measured?
What counts as enough evidence that reality actually changed?
No entries in this area yet.