Your life stays yours.
Your original private records belong in a vault you control. Permission to help with a task should be specific: what may be read, learned, shared, and done.
The philosophy behind HiNet
Intelligence belongs
to people.
Your AI, Everyone’s possibilities.
We are building toward a world where the intelligence you cultivate stays yours—and where choosing to connect it opens possibilities for others.
Individual agency. Shared possibility.
A connection should
expand your world.
A life produces more than records. It builds judgment, working habits, relationships, and experience. Our starting point is that people should guide the intelligence shaped by that life, decide how it participates, and share in the value it creates.
HiNet brings that principle into a network of personal iCores. Each is meant to learn with its owner. When help is needed, willing specialists can form a quorum around the question. Cooperation begins with a choice.
No permanent center.An illustration of the network we are designing.
What we stand for
Six principles.
A human foundation.
Your judgment matters.
A personal intelligence should learn from your corrections and the outcomes of your work. Our research calls this owner-bound learning Atomic Alignment: a continuing relationship between a person and their iCore.
Connect by choice.
A quorum should form from willing participants, with a purpose and a limit. Useful relationships can grow through shared work. Each iCore should learn when a peer can help—and when to keep a question local.
Make room for many minds.
The vision is a network of distinct specialists shaped by different lives. A good answer can bring their perspectives together while preserving disagreement, uncertainty, and the ability to decline.
Share in what you create.
Experience can help someone else without handing over the records behind it. We want useful, permitted contributions to earn recognition and, through a planned marketplace, payment for their owners.
Keep people in control.
Authority should remain close to each person, with enforceable limits and a way to withdraw future participation. Respect for an owner must also include responsibilities toward the other people a system can affect.
Many languages. Different traditions.
Words for
coming together.
Community, harmony, unity, and reciprocity have rich histories of their own. These words offer points of reflection for our project; each carries meanings that a short English gloss can only begin to describe.
संघ
saṅgha · Sanskrit / PaliCommunity
Saṅgha can denote a gathering or community; in Buddhist contexts it has specific meanings, including the community of monks and nuns.
Context: Cambridge University Pressubuntu
Nguni languagesShared humanity
Ubuntu concerns humaneness and the recognition that our humanity is bound up with other people’s. It carries ethical meaning beyond simply being connected.
Context: Dictionary of South African English和
wa · JapaneseHarmony
One meaning of wa is harmony: people coexisting, getting along, and helping one another. The character also has other meanings and uses.
Context: Government of Japankotahitanga
MāoriUnity
Kotahitanga includes unity, togetherness, solidarity, and collective action. In Māori contexts it speaks to people coming together around a shared purpose.
Context: Te Aka Māori Dictionaryayni
Quechua · Andean traditionsReciprocity
Ayni describes ongoing reciprocity in Andean communities, including the exchange of comparable work or goods. Giving and receiving sustain a relationship over time.
Context: Smithsonian National Museum of the American Indianconcilium
LatinAssembly
Concilium can mean a gathering, assembly, or council convened for consultation. It names a setting in which people come together to deliberate.
Context: Lewis & Short Latin DictionaryThese are related ideas, not interchangeable translations or claims that these traditions describe AI. The connection to HiNet is our interpretation; the sources above provide cultural and linguistic context.
From belief to practice
Principles must
shape the system.
For us, ownership means designing local storage and explicit permissions. Personal learning means developing feedback and consolidation around each owner. Cooperation means learning useful peer connections, with checks on what crosses them. Reciprocity means designing ways for contributors to share in the value of their expertise.
These are design commitments, and parts of the network remain research and proposed work. Contributor payments are not live. Learned alignment needs independent permission enforcement; shared abstractions can still reveal private information, and information already shared cannot be recalled.
Decentralization does not by itself guarantee better answers or safer behavior. Our papers make the conditions and limits explicit. The ambition is human-owned intelligence with benefits that can travel further than any one person’s machine.
Your AI,
Everyone’s possibilities.