Skip to content
Personal AI Labs

Technology

Reading the structure
of a mind.

From scattered signals of meaning to a connected personal graph — and from that graph to a learning path no one else could walk.

MEMORYATTENTIONIDENTITYCURIOSITYPRACTICECOREHOVER A NODE — DENSITY = PERSONAL WEIGHT

fig. 04 — anatomy of a semantic core

01 · The unit

What is a semantic core?

A semantic core is a dense cluster of connected meaning inside a personality: a value with the experiences that shaped it, an interest with the questions it keeps asking, a goal with the identity it protects.

Cores differ in weight. Some organize half a life; others are still forming. Our models attend to that weight — because a learning path routed through heavy cores holds, and one routed around them decays.

02 · The structure

One person, one graph.

Connected cores form a personal semantic graph — a structure as individual as a fingerprint, but alive: it shifts as the person grows. Everything we build reads from, and writes back to, this graph.

fig. 05 — a personal semantic graph (anonymized structure)

Nodes
Semantic cores — values, interests, cognitive patterns, goals — each carrying a personal weight.
Edges
Measured relationships between cores: reinforcement, tension, dependency. The bridges a path can use.
Weights
How much of the personality a core organizes. Heavy cores anchor the graph; paths respect them.

03 · The route

From graph to learning path.

A path is computed the way a good mentor works: start where the person is strong, cross by meaning, stretch just past the edge of the known.

fig. 06 — anchor → bridge → stretch

α

Anchor

The path starts at a heavy core — territory the learner already owns. New knowledge gets something to hold on to.

β

Bridge

Each next concept is reached over an existing edge: an analogy, a shared question, a value it serves. No cold jumps.

γ

Stretch

Only then does the path leave the map — into genuinely new ground, at a distance the graph says this mind can cross.

04 · The loop

The path learns back.

People change while they learn — that is the point. So the graph is never frozen: every step a learner takes updates the map, and the map re-routes the path.

Observe, re-map, re-route, learn. The loop is slow, deliberate and private — adaptation in service of the person, not of engagement metrics.

fig. 07 — the adaptivity loop

05 · The instrument

The pipeline, end to end.

Five stages between a person expressing themselves and a learning path that fits them — each one auditable, each one revisable, none of them a black box we ask you to trust blindly.

fig. 08 — from signal to living route

  1. S01

    Signal

    Language, choices, questions — the raw expressions of a mind, read with consent and attention to meaning rather than metrics.

  2. C02

    Cores

    Models surface the dense clusters of meaning: values with their histories, crafts with their metaphors, goals with the identities they protect.

  3. G03

    Graph

    Cores and their measured relations become a live personal graph — reinforcements, tensions, bridges — one structure per person.

  4. P04

    Path

    A route is computed through the graph: anchor in strength, bridge by meaning, stretch just past the edge of the known.

  5. L05

    Loop

    Every real step the learner takes flows back: cores re-weighted, edges revised, the route re-drawn. The map stays honest.

06 · Foundations

Engineered on old truths,
not new hype.

None of our core claims are novel — the research traditions below have argued them for half a century. What is new is the ability to act on them for one specific person at a time.

Semantic networks

Knowledge lives as relations, not entries.

Decades of work on how concepts connect in memory — spreading activation, associative structure — give us the vocabulary for what a personal graph is and why attachment beats repetition.

Learning sciences

Growth happens just past the edge of the known.

From the zone of proximal development to mastery learning: the stretch that teaches is the one calibrated to the individual, not the cohort. Our path grammar is this insight made computable.

Memory research

What connects, consolidates.

Spacing, retrieval, elaboration — memory science consistently rewards knowledge that is woven into existing meaning. Routing new material through heavy cores is elaboration by design.

Psychology of meaning

People are organized by what matters to them.

Values, narrative identity, intrinsic motivation: the study of what makes things matter to a particular person is what separates a semantic core from a keyword cloud.

05 · The stance

Personal AI is only possible with radical respect for the person.

  • 01

    A person's semantic graph belongs to that person — not to us, and never to advertisers.

  • 02

    We collect the minimum the technology needs, and we say plainly what that is.

  • 03

    No selling of personal data. No shadow profiles. No dark-pattern consent.

The fine print matches the promise — read our Privacy Policy.

Appendix

A short glossary of the atlas.

terms as we use them — precisely

Semantic core01
A dense cluster of connected meaning inside a personality — a value with its history, a craft with its habits, a question with its pull. The unit our maps are made of.
Weight02
How much of the personality a core organizes. Heavy cores anchor learning; ignoring them is why generic content decays.
Edge03
A measured relation between two cores: reinforcement, tension, or dependency. Edges are the bridges a path can legally cross.
Personal graph04
All of a person's cores and edges held as one live structure. As individual as a fingerprint; unlike one, it changes as you grow.
Anchor05
The first move of a path: begin at a heavy core the learner already owns, so new knowledge lands on solid ground.
Bridge06
The second move: reach the next concept over an existing edge — an analogy, a shared pattern, a value it serves. No cold jumps.
Stretch07
The third move: leave the map for genuinely new territory, at a distance this mind has shown it can cross.
Learning path08
A route computed through the personal graph by the anchor–bridge–stretch grammar, re-drawn as the person changes.
Adaptivity loop09
Observe → re-map → re-route → learn. The cycle that keeps the map honest and the path personal.
Super-personal AI10
Our term for systems built around one person's structure of meaning — as opposed to personalized delivery of the same content to everyone.

Want to go deeper into the research?