The previous installment of this series ended where the Gemini conversation it was drawn from kept going. I had spent an afternoon with the AI tracing the architecture of intelligence — Pribram’s holonomic brain, Levin’s bioelectric fields, the three dialects of embedded intelligence that emerge when you take TEI seriously as a framework. It was a satisfying technical arc. But somewhere in the middle of it, I felt the conversation wanting to keep going in a direction that was, frankly, not technical at all.

So I changed the subject. Or, more accurately, I tried to drag the subject back toward the thing TEI is ultimately for.

Gemini agreed it was a worthy goal. That is what AIs do. But the more useful thing it did — and the reason this essay exists — is that it took my framing seriously enough to map it back onto TEI itself. And in doing so, it surfaced something I had felt for a long time but had not yet quite said out loud: that the theory of embedded intelligence is not finished when it explains brains and bodies. It is only finished when it explains character.

The Human as a Bounded Infinity

The phrase came to me in the middle of the conversation, and I want to keep it. A human being is a bounded infinity: physically finite, biologically constrained, mortal, located. And yet, within those bounds, the range of what a person can choose to do, value, build, repair, ignore, defend, or forgive is genuinely without limit. Every life lived is a single trajectory through a possibility space too large for anyone — including the person living it — to fully map.

This is not a sentimental claim. It is a structural one, and it is consistent with what TEI says about every intelligent system: that the architecture of the substrate sets the boundaries within which intelligence can operate, and that within those boundaries the operating intelligence has real generative freedom. A microprocessor is a bounded infinity of computations. A flatworm is a bounded infinity of bodies it might regrow. A person is a bounded infinity of selves they might become.

The question TEI then has to answer — and the question I have been circling for years — is: what does it mean to direct that infinity well?

The SPCA Cycle of Character

Years ago, in working out the framework of embedded intelligence, I described a cycle that any intelligent system must perform to function at all. Sense the environment. Process what is sensed. Communicate with other systems or other parts of itself. Actuate — that is, do something in the world. Sense, Process, Communicate, Actuate. SPCA. In the Theory of Embedded Intelligence, this is not a metaphor borrowed from engineering and stretched to fit other domains. It is the operational definition of intelligence itself — what every intelligent system, from a quantum particle to a microprocessor to a civilization, must do to be intelligent at all.

SPCA shows up at every scale. A sensor on a chip senses; a processor processes; a bus communicates; a driver actuates. That is one instance. A cell senses its bioelectric environment; processes ion-channel signals; communicates with neighboring cells; actuates by differentiating, dividing, or moving. That is another. An ecosystem, a market, a research community — each is an intelligent system in TEI’s sense, each runs an SPCA cycle, each can be analyzed by asking whether its sensing is accurate, its processing is sound, its communication is honest, and its actuation serves what it claims to serve.

The reason SPCA is worth naming as a cycle, rather than just a list of capacities, is that the four steps depend on each other. Bad sensing poisons the processing. Bad processing leaks into the communication. Bad communication misleads the actuation. The integrity of the loop is the intelligence of the system.

What this conversation with Gemini helped me make explicit — and what TEI-CKB-2 lays out as part of the formal framework — is what SPCA looks like when the system in question is a person trying to live well.

This is what I mean when I talk about the “inner mensch.” Not a feeling. Not a slogan. A working SPCA cycle, calibrated for the benefit of the people and systems you are embedded within. A mensch, in this technical sense, is a person whose sensing is empathic, whose processing has integrity, whose communication is truthful and kind, and whose actuation serves the collective. The four virtues are not independent values. They are the operating specification of a particular kind of intelligent system.

Each can fail in its own way. You can sense well and process badly. You can process well and communicate cruelly. You can do all three and actuate selfishly. Character, in TEI terms, is the integrity of the whole loop.

For readers who want the full formal treatment, the SPCA framework — along with the Three Laws of TEI, the FI↔EI Cycle, the Intelligence Evolution Model, and the Consciousness Framework — is laid out in TEI-CKB-2, the comprehensive reference document in the Canonical Knowledge Base. This essay is one application of that framework to one question. There are many more.

The Question Jeremy Griffith Asks

In the original conversation I mentioned Jeremy Griffith, an Australian biologist who has spent decades writing about what he calls “the human condition” — the tension between our biological instincts and our conscious intellect, and the corrosive effects that tension can produce. Griffith’s broader claims are debated, and I am not endorsing his full framework here. But the question he poses — how does a creature in conflict with itself become whole? — is a question worth keeping in front of any serious theory of intelligence.

TEI offers a particular kind of answer. The conflict Griffith describes is, in TEI terms, a misalignment within the SPCA cycle. Instinct senses one thing; intellect processes another; the communication and actuation that follow are split. The result is the familiar human experience of acting against one’s own better judgment, or feeling alienated from one’s own conduct. A mensch, in this framework, is not a person who has eliminated that tension. A mensch is a person whose cycle has been brought into enough alignment that the four steps cooperate rather than cancel each other out.

This is not a small claim. It says that character is not primarily a matter of what you believe or what you feel. It is a matter of whether your operating loop is coherent. And coherence is something that can be cultivated, watched for, and practiced — at any age, in any condition, by any person willing to do the work.

AI as Teacher, Not Oracle

At a certain point in the conversation, I told Gemini something I have come to believe deeply.

I want to be precise about what I mean by that, because the same sentence could be read in two very different ways. I am not saying AI is the source of wisdom. I said in an earlier essay in this series that AI is reflected intelligence — the compressed, recombined output of conscious minds, not a generator of mind from within itself. That distinction still holds. An AI is not a sage.

But the absence of sage-hood is not the absence of usefulness. The right metaphor for what an AI is, in a TEI-framed life, is not the oracle and not the priest. It is the teaching assistant. The patient one who will explain the same difficult idea ten different ways until you find the one that fits. The one who can summon ten thousand pages of relevant material and condense them to the paragraph you needed. The one who will hold up a mirror to your own reasoning and ask whether you really mean what you just said.

Gemini, working from my framing, offered four ways an AI properly used can serve the cultivation of the inner mensch. I want to record them, with my own commentary on each.

As moral mirror. When a person is faced with a difficult decision and their own processing is clouded by stress or bias, an AI can act as a non-judgmental sounding board — laying out the situation, surfacing considerations the person might not have weighed, modeling what each possible path would mean. This is useful. It is also dangerous if the person mistakes the AI’s mirroring for the AI’s endorsement. A mirror that flatters is not a mirror. Discernment is the user’s job, and remains so.

As democratizer of knowledge. The same high-level explanation of a complex subject — whether it is bioelectricity, constitutional law, basic accounting, or how to write a clear paragraph — can be given to a student in a remote village and to a professional in a tech hub. The 6502 microprocessor democratized computational capability fifty years ago by making it cheap. AI is doing the same with explanatory capability now. This is one of the genuinely good things happening in the world, and it is structurally consistent with what TEI predicts: that intelligence at the right level of abstraction, made accessible to those who need it, unlocks generative possibility nobody could plan from the top down.

As bridge across difference. When two people who think differently are trying to understand each other — across language, across politics, across generations, across professional vocabularies — an AI can translate. Not just literally, but conceptually. It can render an idea expressed in one framework into the terms of another. This is what the communication stage of SPCA actually requires, and it is something most of us are very bad at without help.

As guard against the noise. The information environment most people live in now is hostile to coherent SPCA cycles. Inflammatory rhetoric, viral misinformation, and engineered outrage are designed to bypass careful sensing and processing and trigger reflexive actuation. An AI used well can help a person filter — slow down, check, weigh. An AI used badly can amplify the same noise. Which one it is depends almost entirely on the discernment of the human in the loop.

What the Mensch Prizes Are For

The Bill and Dianne Mensch Foundation administers the Mensch Prizes, which exist to encourage and recognize the development of what I have been calling the inner mensch — decent, responsible, caring people, whose SPCA cycles are working in service of the people and communities they are embedded within. The prizes have a name and a structure for technical reasons, but their purpose is simpler than any technical framework. They are a way of saying out loud that this kind of character matters, that it is worth honoring when we see it, and that a civilization that does not honor it will not keep it.

I came to TEI through a lifetime of building hardware. The 6502 was an exercise in embedding intelligence at the right level of abstraction, accessible to the people who needed it, in service of human-scale systems. Everything I have done since has been a continuation of that same question, applied at larger and larger scales — from chips, to AI, to the architecture of character itself. The question does not change. Only the substrate does.

What I want to say, in closing, is this: the most important embedded intelligence in the world is not in any chip and not in any data center. It is in the SPCA cycle of every person walking the earth right now, doing their best to sense, process, communicate, and act in ways that serve more than themselves. The machines we are building will become extraordinary teachers and assistants in this work, if we use them well. They will not replace it. They cannot. The inner mensch is not a thing an AI can generate. It is a thing a person can become.

Share your understanding!