Build for Human.

A product is an instrument of exchange. A person gives it some combination of their attention, their trust, and their money. In return, the product changes something. Every product therefore makes a promise, whether or not anyone wrote it down: give me these things, and you will be different afterward in a way you'll be glad about.

Most products never state the promise, which is convenient, because most products could not survive stating it. At Antropod, stating the promise is the first act of building. Everything else in this document is machinery for checking whether the promise is honest and whether it was kept.

An outcome is the change in a person's condition that a product causes and that persists after the session ends. Three things an outcome is not, because these are the standard substitutions the industry makes:

It is not an output, what the product did: lessons delivered, notifications sent, features shipped. A language app can deliver ten thousand lessons to a person who still cannot order coffee in the language.

It is not an experience, what using it felt like. Experience matters, but a slot machine is also an experience. The feeling during the session is not the change after it.

It is not engagement, the person returning. People return to things that harm them; return is evidence of pull, not of benefit.

The outcome is what remains in the person when the screen goes dark. That is the unit we underwrite, build for, measure, and answer to.

Tangible outcomes are changes you can count directly in the world: money saved or earned, weight lost, hours returned to a person's week, a job secured, a debt cleared, a skill certified. The world does the measuring for us.

Intangible outcomes are changes in a person's inner condition or capacities: confidence, understanding, peace of mind, connection, restored energy, self-command. A person who gained confidence makes different decisions for the rest of their life. The intangible kind is often the deeper kind; much of what makes a life good is intangible.

Two nearby words point in opposite directions. Intangible means cannot be touched; it says nothing about importance. Immaterial means of no consequence, the word the industry would love to pin on everything that doesn't show up in a spreadsheet.

Our claim is the reverse: every outcome we build for, tangible or intangible, must be material, substantive, consequential, the kind of change whose absence would be noticed in the person's life. Materiality is our floor. A tangible outcome can fail it (a badge, a streak counter) and an intangible one can clear it easily (the confidence to ask for a raise touches everything).

And materiality implies measurability, at least by proxy, because a change of consequence leaves evidence. We do not demand that intangibles be reduced to a single number; we demand that they leave footprints we agreed in advance to look for. If no conceivable evidence could ever show the change occurred, what we have is not an outcome; it is a sentiment.

So the full standard for an Antropod outcome: definable before we build, material in the person's life, and verifiable after we ship, whether tangible or intangible.

A product does not produce one outcome; it produces a ledger, outcomes summed across all its users and across time, including second-order effects. Some products can point to a satisfied individual while running a ledger that cannot go positive.

Gambling is the instructive case. Any given user may win; the ledger cannot. The house edge is the business model, which means the expected value of the users' collective attention and money is negative by construction. Dosage does not change the sign of the expected value.

Time matters on the ledger too. A product can deliver a small positive every session and compound a large negative across years. We account at the horizon where the person actually lives their life, not the horizon of the session. Our requirement: the ledger must be net positive across users, across time, with second-order effects included.

The process is the path by which the outcome arrives: what the person understands, consents to, experiences, and becomes along the way. Two products can deliver the identical outcome through opposite processes, one through clarity and dignity, one through shame and manipulation, and they are not the same product.

An honest process meets three conditions. The person understands what the product does and how it makes its money, well enough to explain both accurately to a friend. Nothing in the interface manufactures urgency, exploits confusion, or says one thing while doing another. And the person emerges stronger and clearer for the journey, more capable, more agent, not merely delivered somewhere while sedated.

The process is where a promise is kept honestly or kept by force, and only the first counts as kept.

Attention is a portion of a finite life. When a person gives a product an hour, they are spending an hour of the only life they have. This makes attention the most serious thing a product ever asks for, more serious than money, which is replaceable.

But attention is not a cost to be minimized, and requiring a great deal of it is not a mark against a product. A piano teacher who demanded less of your attention would be a worse teacher. A serious course, a hard training program, a book you cannot put down: these absorb people completely, and the absorption is the product working. Flow is among the best states a human can occupy.

The discriminator is not volume, it is what the attention becomes. Attention as investment converts into something that persists: skill, health, understanding, wealth, connection, or legitimate restoration. Attention as consumption terminates in the product itself; the hours are spent and nothing persists, or what persists is negative.

And here is what makes this load-bearing: the consuming products are often light on attention, and the investing products are often heavy. A Ponzi scheme's entire pitch is minimal attention for maximal return; the language course asks for two hundred hours. A test that certifies the fraud and flunks the teacher is a broken test. Duration was never the sin. Conversion is the test.

The most important definition, because it sets the reference point for every judgment above: better by whose lights? Two tempting standards fail. The first is the person's desires, but people demonstrably want, and pay for, things that diminish them; desire is evidence, never verdict. The second is the builder's sentiment, build what you would give someone you love. This was, for a time, our own final test, and we retired it, because love without wisdom hands people harm sincerely.

The standard must sit outside anyone's feelings: the measurable condition of the person afterward, judged against criteria named before we started. Not what they wanted, not what we fondly imagined for them, but what changed. (The love question survives in a demoted role, as a fast human gut-check. It screens; it does not judge.)

A note on foundations. These definitions carry a conviction: serve the person in front of you, not the aggregate; tell the truth; restore agency rather than capture it. That conviction has a source, and we know what it is. We express it through behavior, not branding, in language anyone can hold us to.