A living genealogy of ideas.
This isn't a manifesto. It's a working record of things discovered through designing, making, teaching, observing and conversation—where they came from, what they connect to, and what they become.
Some may grow into principles. Some may contradict each other. Some may go nowhere.
One occurrence is an observation. Repeated occurrences become a pattern. A pattern that survives different contexts may eventually become a principle.
Ideas that have appeared often enough to be worth carrying forward—and testing again.
Begin with why something should exist and what should happen. Hardware, software, AI and physical material are implementation choices.
Really important, somewhat important, and nowhere near as important should not receive equal treatment. Hierarchy is not decoration; it communicates judgment.
Attention spent on one thing is unavailable to another. Asking someone to transcribe exactly what is being said while simultaneously reasoning about it exposes the constraint immediately.
A notification, paragraph, choice, animation, explanation or interruption makes a claim on cognition. Attention should be proportional to importance.
A system may know ten thousand things while asking a person to reason about three. Keep the active working set graspable.
People already point, talk, arrange, hand things to one another, squeeze, hesitate, gather and compare. When intelligence can understand those behaviors, don't require people to encode intent into computer vocabulary first.
Effort isn't automatically failure. Resistance can establish intent, make consequence perceptible, or strengthen judgment.
Trivial actions should be easy. Consequential actions can deserve deliberation, resistance, ceremony or even irreversible change.
The current state tells you what is. History tells you why. Whenever possible, don't make people remember what the system thinks happened.
Build the smallest thing that exposes the truth. A prototype is not merely a cheaper product; it is an instrument for answering a question.
Externalize enough of a system that you can point at it, compare it, reconstruct it, question it and let the representation change your understanding.
A feature says what the product can do. Capability asks what the person can now do that they couldn't do before.
A good idea proves that an idea exists. It doesn't prove it is the only idea. Generate enough possibilities to see relationships, alternatives and frames that weren't visible at the beginning.
Exploration has value until further exploration stops producing value. Sometimes an idea resolves the problem strongly enough that continuing is hesitation rather than discovery.
Integration does not require sameness. Combine things enough that they can work together while preserving the differences that made each useful.
Give enough understanding to let someone continue without unnecessarily taking discovery, judgment or agency away from them.
Not everything needs to be either visible or gone. Important things can come forward; less relevant things can recede while remaining in awareness. Reduce attention without destroying awareness.