The Practice
Running yourself as a system
Your body already runs itself. Temperature, heartbeat, sleep — regulated without your permission, better than you could ever do deliberately.
One layer is left to you. Energy. Tension. Clarity. The state you act from — the condition underneath every decision, every hour of focus, every conversation you handle well or badly. It is the most consequential system you own, and most people run it on guesswork.
We manage everything else. Calendars, budgets, codebases, teams — instrumented, reviewed, steered. The system that operates all of them goes unmanaged. Not because it doesn't matter. Because until now there was no good way.
Managing your state is not a feature, and not a hack. It is a practice — an old one. Slow breathing, posture, directed attention: people have used these to steer their own condition for longer than we've had writing. Modern research caught up and confirmed what held. The methods were never the missing piece.
What was missing was memory. A practice needs a mirror — something that sees your state relative to you, not to an average. And it needs a ledger — something that remembers what you tried and what actually moved you. A teacher used to be that memory. Most people never had one.
That is what AI changes. Not new methods — the old ones, finally personalised. A system that reads where you are, holds your baseline, and learns, over months, what works for the one person it serves.
So the practice is simple to state. Read your state before you fight it. Steer it with methods that have earned their place. Keep the ledger. Let it compound.
It comes with rules. Your state is functional, not clinical — no diagnoses, no predictions, no fear. Signals are physical — no machine should tell you what you feel. And the model built from your practice belongs to you.
We build Deywos for this practice. But the practice is bigger than any product — it is a way of running yourself. If you start today with nothing but your own attention, you have started.
If you want a companion that keeps the ledger — Deywos is here.
— Manuwey · 2026
A note from the founder
People think environment shapes the person. My experience taught me the opposite: environment sets the conditions — state dictates what you do with them. Your reactions, your decisions, and in the end, the changes you make around you — all of it runs on the state you're in.
I didn't learn this from books. I learned it living through war, through losing people I loved, through poverty and the kind of loneliness that sits next to despair. In every one of those situations there was exactly one thing still mine to govern: the state I met it from. That choice — break, or use what was breaking me — is the only reason I'm here building this.
So I built what I needed and couldn't find: a guide and a working base of methods, ancient and modern, for managing your own state. Deywos won't make you a different person, and it won't promise you an easy life — mine wasn't. It helps you act from the state you'd choose, more often than chance would allow. The rest of life stays what it is: grief and joy, barely glowing and burning bright. You get to be present for all of it.
— Ivan Nevidomskyi
Definition
What makes an AI state-aware
A state-aware AI is an AI system that takes a person's current functional state — energy, activation, tension, clarity — as a first-class input, models it against that person's own baseline, acts to shift it, and measures whether the shift happened. Language models process what you say. A state-aware AI works with the condition you say it from.
- It reads state. Self-reported scales, body signals, physiological measures — some structured read of where the person is right now. An AI that only infers state from word choice is guessing, not reading.
- It models the individual. State means little in the absolute. A state-aware AI holds a personal baseline and interprets every reading relative to that one person — not a population average.
- It acts. Reading without response is tracking. A state-aware AI adapts its behaviour to the state it reads — in conversation, or through guided practice designed to shift the state.
- It measures outcomes. After acting, it asks what changed, records the answer, and lets the record sharpen its next response. Without a closed loop, there is no learning — only content.
- It respects the line. State is functional, not clinical. A state-aware AI describes what signals show — activation, recovery, tension — and never claims to read emotions from physiology, never diagnoses, never predicts illness. Systems that cross this line aren't state-aware; they're guessing with confidence.
By these criteria, most current tools fall outside the category. Wearables read and model, but don't act. Chatbots act, but don't read or measure. Meditation apps deliver practice, but hold no model of the individual and close no loop. A state-aware AI is defined by doing all five — the loop is the category.
First published by Manuwey, 2026 · This definition will evolve; changes are logged.
