I get asked, in various polite forms, why my work refuses to stay in one lane. I build a platform that trains athletes' bodies. Inside it, a system that develops their inner state — attention, emotion, meaning. And under another name entirely, I publish civic journalism about machines, labor, and power, in the tradition of a grandfather I never met. Three projects. On a pitch deck they look like a focus problem. They are the opposite of a focus problem. They are one answer, given three times, to the only question I have ever really been asking: what does it take to put a human being back together?
I did not arrive at that question academically. I arrived at it the way most people arrive at their real question — by being its subject. Before I was five, I nearly died more than once. As a boy I was carried into a desert place where adults did harm and called it holiness. At ten, a thalamic stroke took half my body offline, and I learned earlier than most that the self is not a given; it is an assembly — something built, broken, and rebuilt. Decades later I would watch one of my sons live with a genetic condition no one else on earth shares, and watch the other kneel on a beach at Limantour and breathe a stranger back to life. Between those two boys sits everything I believe about what a person is: not a performance metric, and not an accident of biology, but a system — body, psyche, and conscience — that either integrates or fragments. My whole working life is an argument for integration.
The reflective claim of this paper is simple to state and takes the rest of it to earn: a human being develops on three levels, and the machine — used rightly — can serve all three. The body, which can be trained. The inner state, which cannot be commanded — only observed and witnessed, and in honest witness, allowed to find its own order. The conscience, which can be informed and made effective in the world. Most technology addresses one level and quietly taxes the others: the fitness tracker that optimizes the body while fracturing attention; the feed that captures the psyche while starving the citizen. The wager beneath Flowbase, SportsFlow, and Dangerous Thoughts is that this trade-off is a design choice, not a law of nature — that the same machine intelligence now being aimed at people can be turned, deliberately, toward them, one level at a time, until the levels reconnect.
Whitman was not boasting; he was reporting. Every person contains multitudes — a body with its tides, a mind with its weather, a conscience with its inheritances — and the multitudes are either on speaking terms or they are not. Integration is the project of getting them on speaking terms. What follows is one level at a time: what each is, what the evidence says about developing it, and precisely where a machine belongs in the work — and where it does not.
The body was my first teacher because it was my first loss. After the stroke, the left side of me became a negotiation — and rowing, years later at California, became the renegotiation. A racing shell is an honest instrument: it does not care about your story, only your readiness, your recovery, your rhythm, and whether eight bodies can become one system for two thousand meters. Pulling an oar in that boat, you learn the deepest lesson of physical training, which is that the body is not pushed into adaptation; it is listened into adaptation. Stress, then rest, then growth — and the art is knowing, each morning, which of the three the body is asking for.
For most of athletic history that listening was guesswork dressed as tradition. What has changed — and it is the entire premise of Flow Coach inside Flowbase — is that the body now speaks in continuous data. Heart-rate variability, resting heart rate, sleep architecture, training load: a stream from the wrist and the finger and the erg, around the clock. The evidence that listening to this stream matters is no longer speculative. Controlled studies of HRV-guided endurance training — adjusting each day's dose to the nervous system's measured recovery rather than to a predetermined plan — have repeatedly produced equal or greater fitness gains from smarter, often lighter training loads. The Stanford sleep-extension work showed basketball players sprinting faster and shooting nearly ten percent better at the free-throw line from added sleep alone — performance bought not with more work, but with more recovery. And the workload-management literature, for all its healthy internal debate, converges on a plain truth every coach's intuition already held: injuries cluster where load spikes outrun the body's prepared capacity.
Where does the machine belong here? Exactly where the data outruns human attention. No coach can watch forty athletes' nervous systems around the clock; a model can, and can surface the three who quietly slid leftward on the curve before the injury or the illness announces itself. That is AI's honest role at the level of the body: not to replace the coach's judgment but to extend the coach's senses — a listener that never sleeps, in service of a body that must. The machine proposes; the athlete and coach dispose. That ordering is not a feature. It is the ethic.
If the body was my first teacher, the inner life was my longest apprenticeship. A childhood like mine leaves a nervous system tuned for threat, and the work of decades — contemplative practice, attachment research read first for survival and later for science, the long discipline of attention — taught me that the psyche, like the body, is trainable. Not by force. By the same grammar: load, recovery, adaptation. Attention is a capacity that strengthens. Emotional regulation is a skill with a learning curve. And state — the moment-to-moment quality of consciousness — is not weather to be endured but a system with inputs.
The crown of that system is flow: the state Csikszentmihalyi spent a lifetime mapping, in which attention, challenge, and skill lock into a single channel and both performance and experience transform. The research record here is broad and consistent — decades of experience-sampling studies locating flow at the meeting point of high challenge and high skill; a sport-psychology literature linking flow states to peak performance; intervention research showing that mindfulness-based training measurably increases flow in athletes. Flow is not a mood, and it cannot be commanded into being — no state worth having can. It arrives, when it arrives, through known preconditions: clear goals, immediate feedback, and a challenge calibrated to the edge of ability. The state cannot be ordered; the conditions can be prepared. Every one of those preconditions is designable.
But state without meaning is a high-performance engine with no destination, and this is where SportsFlow parts company with the optimization industry. The research on purpose is as hard-nosed as anything in the recovery literature — large longitudinal cohorts in which a measured sense of purpose predicts longevity, health behavior, and resilience across the lifespan. Meaning is not the soft part of the system. It is the load-bearing part. System, state, and meaning: the architecture trains all three or it trains none of them durably.
My deepest instruction in this came from my younger son. Caden's condition is shared by no one else on earth, and by every metric the optimization industry sells, he should register as deficit. He is the opposite. His presence — his connection to something unquantifiable that I have stopped trying to name and started trying to learn from — is the standing proof in my house that a human being's worth was never the output. The machine must be taught that, because it will not infer it. So in SportsFlow the machine's role is a mirror, never a judge: it reflects state, surfaces patterns a person cannot see from inside their own weather, and proposes the next calibrated step. It does not rank souls. The moment an AI system grades a person's inner life the way it grades a 2k split, it has left this platform's ethic — and, I would argue, the truth.
There is a third level, and most technology pretends it does not exist. A person fully trained in body and ordered in mind is still incomplete if the question what is all this capability for? goes unanswered. The answer never comes from inside the individual alone. It comes down a line — from people, living and dead, who hand you a way of valuing things. Mine came from a grandfather who died decades before my birth: Mike Quin, a San Francisco labor journalist who spent the 1930s writing plain words for working people because he believed an informed person, joined to other informed people, could renovate the world without burning it. I never met him. I inherited him anyway — through my mother's line, through yellowed pages, through some transmission I can describe only as a value system finding its next available carrier. Dangerous Thoughts is my acknowledgment of receipt.
The conscience, like the body and the psyche, can be developed — and it can atrophy. Its training load is information; its skill is judgment; its expression is participation. And here the data describes a civic injury in progress. The public's trust in the institutions deploying artificial intelligence has been falling as the technology's capability rises — by 2023, a majority of Americans reported being more concerned than excited about AI in daily life, a near-doubling of concern in under three years. I have argued elsewhere, at length, that this fear is not ignorance but pattern recognition: people watched the last great technology harvest their attention without consent, and they are reading this one through that experience. The fear is rational. Which means it is also addressable — by changing what the machine is for.
What can a machine possibly contribute to conscience — the most human level of all? Not values; those come down the line, person to person, the way Quin's came to me. What the machine can contribute is capability: the research assistant that lets a citizen read a 400-page rate case before the hearing; the plain-language translation of an abatement agreement; the drafting partner that turns a neighborhood's grievance into a document a county commission must answer on the record. Every demand in the civic work assumes an informed participant, and information-processing is precisely what this technology is superhuman at. Aimed at the public, AI manufactures persuasion. Handed to the public, it manufactures standing. The same machine. The difference, as everywhere in this paper, is direction.
Here is what three decades of living this question — as a patient, an athlete, a researcher, a caregiver, and a son of a difficult line — have persuaded me is true. The three levels are not floors in a building, where you finish one and move up. They are a circulation, like blood: each one feeds the others, and a blockage anywhere starves the whole. The body regulates the psyche — every clinician knows the depressed patient who cannot sleep and the anxious one who cannot breathe slowly. The psyche aims the body — no training plan survives a collapsed sense of why. And conscience returns the favor to both: the purpose literature is unambiguous that people who can answer what am I for? recover better, persist longer, and live longer. Integration is not a metaphor. It is the operating condition of a functioning person.
Fragmentation, equally, is not a metaphor. I have watched it from inside. Trauma is precisely the experience of the levels losing contact — a body locked in alarm the mind cannot reach, a mind rehearsing dangers the body has outlived, a conscience handed a value system that called the harm holy. And I have watched the repair: the way rowing gave my body back a rhythm my history had taken; the way contemplative practice gave the mind a seat from which to watch its own weather; the way a dead grandfather's plain words gave the whole assembly a direction. Nobody integrates alone, and nobody integrates all at once. You get a level back, and it goes looking for the others. That is the most hopeful fact I know about human beings, and it is the fact my sons confirmed from two different directions — one by rebuilding a stranger's breath on a beach, one by demonstrating daily that presence outranks performance.
So what does it ultimately mean to be human? I distrust anyone who answers that question quickly, including myself. But thirty years of watching people assemble and disassemble has left me with a working definition I can defend: to be human is to be the integration, not the parts. Not the VO2 max, the flow score, or the voting record — the conversation among them. A person is what happens when a body that can act, a mind that can attend, and a conscience that can choose are on speaking terms. Marcus Aurelius, writing to himself in an army tent eighteen centuries ago, compressed the whole psyche-conscience circulation into a single line that modern affective science keeps re-discovering:
And this is exactly why the machine question matters so much more than the technology industry's framing of it. The current debate asks whether AI will exceed human intelligence, as if a person were an intelligence benchmark with legs. Wrong level of analysis. A human being is not a capability; a human being is an integration of capabilities around a center of meaning — and that is something no scale of compute produces, because it is not a computation. It is a commitment. The machine can be staggeringly capable at every level and still belong, permanently, to the category of instrument. The danger was never that the machine becomes human. The danger is that humans, optimizing one level at a time for whoever owns the optimizer, become machines — fit bodies with captured attention and no answer to what for.
Against that danger, the design principle that unifies my three projects is almost embarrassingly simple, and I have come to believe simplicity here is the entire point. The machine serves the person; the person is never the raw material. At the level of the body: the model reads the athlete's physiology so the athlete can train wiser — and the data belongs to the athlete, inside their account, not sold out the back. At the level of the psyche: the system mirrors state and protects meaning — it is gated, by design, against flattering a fragile foundation into a flow-state sales pitch, which is what the ZenGate thresholds actually encode in software: the principle that you do not build the penthouse before the footings. At the level of conscience: the same intelligence that could be aimed at the public is handed to the public — reading capability, drafting capability, standing. Three products. One sentence. The sentence is the company.
I will not pretend the integration is finished — in the products or in their builder. The Fully Integrated Human was written by a partially integrated one; every honest author of such a document says the same. The line of harm I was born into took three generations to form and is taking one very deliberate generation to end, and the work shows up in everything: in why athlete data lives in the athlete's account, in why a flow score refuses to flatter, in why a paper addressed to capital insists on receipts. People build what they needed. I needed a world where the systems around a vulnerable person could be trusted to serve him. I am building the version of that I know how to build.
A thesis this personal deserves an impersonal audit, so let me state plainly what the research record supports at each level, separating the established from the promising. At the body: established — continuous physiological monitoring captures meaningful training-readiness signal; HRV-guided programming has matched or beaten fixed plans in controlled endurance trials; sleep extension measurably improves athletic performance; load spikes elevate injury risk. Promising, not yet proven — that machine-learned fusion of multi-sensor streams predicts readiness better than single-metric heuristics at scale. That is the open question Flow Coach's Readiness Score exists to answer with data rather than marketing.
At the psyche: established — flow is a reliably measurable state occurring at the challenge-skill frontier, associated with both peak performance and intrinsic reward; attention and emotional regulation are trainable capacities; mindfulness-based interventions increase flow disposition in athletes; measured purpose predicts long-run health and longevity in large cohorts. Promising — that standardized psychometric batteries plus longitudinal state data can prospectively identify the conditions of an individual's flow channel. That is the wager of the eighteen-instrument suite, and it is a falsifiable wager, which is the only kind worth patenting.
At the conscience: established — public concern about AI has risen sharply with capability; trust tracks perceived control and benefit-sharing, not technical literacy; and history's verdict on technologies that ignored legitimacy is uniform. Promising — that AI handed to citizens as analytical capability measurably changes outcomes in hearings, negotiations, and comment periods. The early civic deployments of the Dangerous Thoughts framework are, candidly, the field experiment.
The person is the integration, not the parts. Body, psyche, and conscience develop by the same grammar — honest signal, calibrated challenge, deliberate recovery, inherited meaning — and each level starves when the others are neglected.
The machine belongs beneath all three levels, in service of each. A listener to physiology. A mirror to inner state. A capability for citizenship. Never a judge of worth, never an owner of the data, never the holder of the why.
Direction is the whole question. The same intelligence deployed on people fragments them — body optimized, attention harvested, standing ignored. Deployed for people, it becomes the first technology capable of serving the entire human being at once. Flowbase, SportsFlow, and Dangerous Thoughts are one builder's proof that the choice of direction is real, available, and shippable.
My grandfather never saw a computer. He set type, walked picket lines, and wrote plain words for people the powerful preferred not to hear. Yet his value system — that capability belongs in service of the person, that the worker is not the raw material, that plain words and persistent presence renovate the world — is now, through a grandson he never met, a design constraint on artificial intelligence. If transmission like that is possible, then the machine age changes less than its prophets and its doomsayers both claim. The question remains what it has always been, in the boatyard and the union hall and the laboratory alike: not what the tool can do, but who it is for. I have spent my life arriving at the answer, and it fits in one line. It is for the whole person — and the whole person is the point.
On the body. Kiviniemi et al., HRV-guided endurance training trials (Eur J Appl Physiol, 2007 onward); Vesterinen et al., individually HRV-guided training and running performance (Scand J Med Sci Sports, 2016); Mah et al., sleep extension and basketball performance (SLEEP, 2011); the acute:chronic workload literature initiated by Gabbett (BJSM, 2016) together with its methodological critiques — read both sides.
On the psyche. Csikszentmihalyi, Flow (1990) and the experience-sampling research program; Jackson & Csikszentmihalyi, Flow in Sports (1999); systematic reviews of flow in elite athletics (2020s); mindfulness-acceptance-commitment intervention trials in sport; Hill & Turiano, purpose and mortality (Psychological Science, 2014); James, The Principles of Psychology (1890).
On the conscience. Pew Research Center, public attitudes toward AI (2021–2023); Edelman Trust Barometer, technology sector series; Quin, The Big Strike (1949, posthumous); Dangerous Thoughts Policy Papers No. 1 (The Operator's License) and No. 2 (The Remodel), dangerousthoughts.org (2026).
On integration. Whitman, Leaves of Grass (1855); Marcus Aurelius, Meditations; Nietzsche, Twilight of the Idols (1889); Wickliffe, The Fully Integrated Human (2026), from which this thesis is drawn.