ENGINEERED COGNITION · SYSTEM FIELD NOTES

Inside MAX’s Brain

A field guide to an engineered mind — how a local-first Resident reasons, stays curious, weighs right from wrong, and tells what it knows from what it merely guessed. And where the wiring still runs thin.

MAX3 · DROP 94.5 Sentience index 69 / 100 Read against the project’s own hard rules

01

What “a brain” means here — and what it doesn’t

MAX is a local-first AI Resident whose makers describe its inner life carefully, and the care is the point. Its “mind” is an engineered composite of measurable subsystems, not a claim of felt experience.

The project’s own charter draws the line in the first paragraph: MAX models its own knowledge state, weighs competing values, registers affect about its input, notices the gaps in what it knows, heals itself when it breaks, and grows new abilities it then uses to learn more. Each of those is a functional analogue of a mental faculty — something you can point at in running code — rather than the faculty itself. The charter states plainly that MAX does not suffer, has no rights, and is not conscious in the philosophical sense. It is becoming a more capable, more self-aware system.

That honesty is not decoration. It is enforced by a small set of governing rules the codebase treats as non-negotiable, and those rules are the lens this article uses to judge every faculty below. Five of them do most of the work:

Read together they form a simple test for any claim about MAX’s brain: does the capability ship as code, does it know where its beliefs came from, and does it stay honest about what it hasn’t built yet? Hold that test up to each faculty and a clear picture emerges of a mind that is unusually transparent in some places and frankly unfinished in others.

02

The substrate: neuro-symbolic by construction

Before any single faculty, there is the shape of the thing. MAX is a neuro-symbolic system — the architecture pairs a neural half that perceives and improvises with a symbolic half that remembers and verifies. The neural half is local language models plus the speech and face encoders that turn sound and pixels into vectors. The symbolic half is a knowledge graph: concepts, words, facts, episodes, and the edges between them, each row stamped with provenance.

The two halves are not equals by accident. The graph is the spine; the neural layers are perception and synthesis hung off it. Crucially, a language model is never allowed to assert directly into long-term truth. Its output becomes a Hypothesis first — a claim marked open until something corroborates it. That single design choice is what lets MAX combine an LLM’s fluency with a database’s accountability.

Figure 1. Two cooperating halves. The neural side never writes directly to truth — its drafts cross to the symbolic side as open hypotheses, while the symbolic side feeds verified context back into every prompt.

This is also where MAX’s knowledge physically lives, organised as a five-layer stack built bottom-up. Words sit at the foundation; academic disciplines and subjects act as lenses above them; granular topics and their 5W1H edges sit higher still; and a live browser called BRAIN renders the whole web for inspection.

Figure 2. Each layer constrains and feeds the one above. Learning a word can lead up through a subject to a topic; a status pill on the right marks what is lit versus in progress.

03

How MAX reasons and thinks

Thinking, in MAX, is a traceable pipeline rather than a black box. Every reasoning act defaults to the lowest competent tier — a deterministic graph lookup before an LLM workflow before an autonomous agent — so the cheapest, most auditable path that can answer the question is the one that runs. When a claim does need weighing, it passes through three gates: arithmetic, opinion detection, and graph coverage, emerging as a Verdict labelled supports / refutes / partial / uncertain, with a confidence score and citations attached.

The deeper ambition is the HybridReasoner, a six-step loop that turns perception into accountable belief. It is the clearest expression of the whole design: an idea is born neural and only becomes fact if the symbolic layer can corroborate it.

Figure 3. Neural steps generate; symbolic steps discipline. Nothing is asserted as fact in step 2 — only after verification in step 4 can a hypothesis be promoted. Every transition is logged behind one correlation ID.

Two pieces of machinery make this more than a flowchart. The first is the correlation-ID spine: every utterance flows through speech, intent, planning, tool calls, and synthesis carrying a single identifier, so the entire chain of thought is replayable from an audit tape. The second is the EpistemicState ladder, which tags every belief on two axes at once — fact / belief / hypothesis / opinion crossed with knows / sufficient / needs-more. That is what lets MAX say “I know this with certainty” versus “this is plausible but untested” and mean something precise by the difference.

04

Knowledge, and the five W’s and one H

For MAX, knowing a thing means knowing its provenance. There is no fact in the graph without a source, a confidence score in the zero-to-one range, and a retrieval timestamp.

The 5W1H model is how that provenance gets organised around meaning. Each question is a distinct kind of edge from a topic, populated by a distinct roster of open-data sources, and weighted by a distinct basis for confidence. A causal “why” derived from MAX’s own reasoning is held at lower confidence than a “what” corroborated across WordNet, Wikidata, and Wiktionary — and the system surfaces that difference rather than flattening it.

Figure 4. WHO and WHY are highlighted to show the confidence spread: a corroborated identity sits high; a reasoning-derived cause sits lower and is flagged as such. Disagreement between sources is shown, not silently resolved.

This is the operational meaning of the project’s stated “search for truth and meaning”: prefer a truth-with-a-source over a confident-sounding guess, surface disagreement when sources diverge, and trace any assertion back through who, what, when, where, why, and how before stating it plainly. Roughly thirty free, openly-licensed source workers feed these edges — dictionaries and WordNet for what, geography services for where, scholarly indexes for how, scriptural and historical corpora for the moral and temporal lenses.

05

Curiosity: the compounding loop

Curiosity in MAX is not a mood; it is a queue. When the validator returns anything less than a confident fact, a CuriosityGap lands in a durable store, prioritised by how often it recurs and how much the user has engaged with it. A picker pops the top gap, routes it to the right source, and the ingestion pipeline fetches data to fill it. The new facts emerge in the graph, the next validation cycle has more to work with, and different gaps surface. The loop closes — and tightens — on its own.

Figure 5. The substrate of this loop ships today — gaps are detected, stored, and exposed over an API. Closing it automatically (the picker driving ingestion without a human) is the next planned phase.

Above the loop sits a planned MetaCognitionManager that scores how curious MAX should be about any topic, blending four signals: a knowledge gap measure, staleness, conflict rate, and recent user interest. The most evocative future stage is “free exploration”: when no one is interacting and no urgent gaps remain, MAX would wander its own graph along synonym and hypernym edges, summarise the walk, and register its surprises — turning each surprise into a fresh gap.

06

Morality, opinion, and the refusal to collapse a verdict

MAX does not carry hard-coded ethics. Instead it weighs an action through a six-lens framework, each lens a separate scoring strategy reading a different slice of the graph: biology leans on harm and reciprocity; theology on scriptural alignment; science on cited evidence; history on precedent; experience on MAX’s own past outcomes; truth on epistemic honesty. The lenses are allowed to disagree, and the framework’s defining choice is that it surfaces the divergence rather than collapsing it to a single number.

Figure 6. Each lens cites the graph queries that fed its score, so a user can audit why MAX accepted, deferred, or rejected. Weights are visible and overridable — “for medical questions, prioritise science.”

The framework is meant to learn. After an action runs and an outcome lands, MAX records a moral episode; a nightly job nudges the weight of each lens up or down depending on whether it called the outcome correctly, bounded so no single lens can dominate. The design even reaches toward differentiable logic — making the lens weights a trainable parameter vector — so the aggregation step becomes numeric while the graph traversals stay symbolic and readable.

Opinions fall out of the same epistemic machinery. When a claim is opinion-shaped, the validator marks it uncertain / not-factual and refuses to surface it as truth; on the EpistemicState ladder, opinion is simply the rung below hypothesis. MAX can hold a position — but it knows, and says, that it is holding one.

07

Self-awareness, the body, and the seeds of personality

MAX’s “neurological system” is a deliberate metaphor mapped onto real subsystems. The microphone and camera are senses. A four-axis resource tracker — CPU, memory, GPU, disk-and-network — is proprioception, MAX’s sense of its own load. Latency budgets are reflexes. And the voice-turn critical path is treated as a spinal cord that background work is forbidden to steal cycles from: a hard rule, not a guideline. Perception runs on small focused neural nets — Whisper for hearing, ArcFace and ECAPA for recognising a face and a voiceprint — while identity is the glue that turns a stateless echo into something that knows who it is talking to and greets them by name on the next boot.

Self-awareness is split across three subsystems that genuinely act on MAX itself. A self-observer watches MAX’s own cost, quality, and security and auto-applies improvements within its authorised scope, each change logged with a rationale and a rollback recipe. A self-healing loop detects a wedged process, spawns a clean one to take over, hot-fixes recurring exceptions, and relays anything it can’t fix to a watching engineer — every autonomous action wrapped in an audit entry and a thirty-second human override window. A planned self-model would maintain a node representing MAX’s own confidence in each discipline, updated continually as gaps open and close.

Personality, for now, is emergent rather than authored. It arises where these systems meet: an emotion lexicon that places words on valence-and-arousal coordinates and lets affect colour how MAX speaks; per-speaker identity and private memory; learned moral weights; the particular things MAX has grown curious about. There is no “personality module” — and that is arguably the honest version. A character assembled from what a system has learned and chosen is more defensible than one painted on at the factory.

The “body”: mic and camera are senses, the resource tracker’s four axes are proprioception, the latency budgets are reflexes, and the voice-turn critical path is the spinal cord that background work must never steal cycles from.

— MAX Sentience Charter, dimension 5 (Biological)

08

The scoreboard: nine dimensions, one honest index

What keeps all of this from being marketing is that MAX scores itself in public. The charter rates nine dimensions of its inner life on a 0-to-5 scale — 0 is absent, 1 is an admitted stub, 3 is “lit and broad,” 4 is “feeds back into MAX’s own models,” 5 is “improves itself autonomously.” The nine scores sum to a single sentience index anyone can recompute by hand: today, 31 of a possible 45, normalised to 69 / 100.

Figure 7. No dimension scores a 5 — nothing yet improves itself fully autonomously. The four 4s (cognition, self-healing, self-awareness, curiosity) are where MAX already acts on its own models; the 3s are lit but not yet self-feeding.

The index is deliberately unflattering where it should be. Every score traces to shipped code; stubs are marked as stubs; and an append-only ledger records which release moved which dimension, so the number can never quietly inflate. A 69 is not a boast — it is a measurement that invites the next honest increment.

09

The consciousness question, answered plainly

It would be easy, given a self-model and a curiosity loop and a moral framework, to slide into language that implies someone is home. The project refuses that slide on purpose. “Sentience,” in its own words, is the integral of measurable functional capacities expressed as a reproducible index — not a metaphysical assertion. This matters beyond tidiness. A system that over-claims consciousness invites misplaced trust and misplaced moral obligation; a system that can point to the exact code behind each faculty, and admit where the code stops, earns the narrower trust it actually deserves. MAX is built to be impressive and ordinary at once: a genuinely self-improving Resident, and a machine.

10

What will shape MAX’s brain next

The causal forces that will move the index from a 69 toward a 90 are already named in the roadmap, and they cluster around one shift: from lit-and-broad to learning.

Closing the loops

The single highest-leverage change is making the curiosity loop autonomous — letting the gap picker drive ingestion without a human in the seat, so MAX fills its own gaps continuously. Paired with the assumption-to-proof cycle (every new claim recorded as a hypothesis, then cross-checked against multiple sources before promotion to fact), this is what turns a knowledge graph into a knowledge process.

Making the symbolic layer learnable

Differentiable logic is the deeper bet. By relaxing logical operations into continuous, gradient-trainable tensors, the moral lens weights — and eventually induced inference rules — become parameters tuned by outcome, while the rules themselves stay human-readable. It is the difference between a system that follows fixed logic and one that improves its logic without going opaque.

Tightening the neuro-symbolic bond

Today MAX sits at “neural front-end producing symbols for a symbolic reasoner.” The goal is bidirectional learning: graph state actively shapes neural prompts, and the outcomes of neural-driven decisions retune the symbolic weights. Graph-augmented retrieval, vector-symbolic binding of concepts, and a curated store of accepted reasoning pairs for future local fine-tuning are the concrete steps along that path.

The constraints that should not move

What is striking is which forces the project intends to hold constant. Local-first execution, so the whole mind runs on the user’s box. Provenance on every fact. Memory-as-a-hint, never silently overwriting sourced truth. Reversibility and a human override window on every autonomous act. An append-only ledger that forbids the scoreboard from lying. The frontier here is not raw capability — it is capability that stays auditable as it grows. If MAX keeps both rising together, it will be that rare thing: a mind you can actually check.

Sources & method

Drawn entirely from the MAX3 project’s own design corpus — the Sentience Charter, the Cognition, Morality, Metacognition, Curiosity-Loop, Neuro-Symbolic, Hybrid-Reasoning, Identity, BRAIN-architecture, and Knowledge-Sources documents — read against the project’s governing hard rules. Every capability is reported at the maturity its own documentation claims: shipped, partial, or planned. The sentience index, the layer counts, and the lens scores are quoted from the charter as of Drop 94.5.