Behind the archetypes

Most genetic reports hand you a list. Variant X associated with trait Y. Variant Z associated with trait W. Pages of individual associations, each with a tiny effect size and no connection to any other.

It feels like information. It isn't, really. It's data with no frame.

The list model made sense when genetics worked a certain way: one gene, one outcome. It still works for a handful of conditions — cystic fibrosis, sickle cell, Huntington's. These are high-penetrance traits where a single variant largely determines what happens. Consumer genetics inherited that framework, then tried to apply it to everything else.

Almost nothing people actually care about works that way. Metabolism, body composition, stress resilience, sleep timing, sensory sensitivity — these are polygenic traits. Dozens or hundreds of variants, distributed across biological systems, interacting in ways that no single-variant report can capture. The associations are real. But stripped of each other, they don't tell you much.

This is the monogenic fallacy: using single-gene thinking on biology that doesn't work that way. It's why your genetic report felt underwhelming. Not because the data was bad — because the abstraction was wrong.

A different model

We start with a different question. Not "what does each variant say?" but "given everything we can measure about how your biology is wired, what kind of system are you?"

To answer that, we use two complementary methods. For traits driven by a small number of high-impact variants — drug metabolism is the clearest example — we score those variants directly. Your CYP enzyme variants determine how quickly you clear caffeine, alcohol, and certain medications. The mechanism is understood. The effect is large. This is as close to deterministic as genetics gets outside clinical disease.

For traits distributed across hundreds of genomic regions — stress recovery, circadian timing, body composition — a handful of candidate variants can't carry the load. Here we draw on genome-wide data from population studies involving hundreds of thousands to over a million participants, using statistical methods that aggregate signal from across the entire genome rather than from a curated shortlist. We also use imputation to expand from hundreds of thousands of directly measured genetic markers to millions of inferred data points, increasing the resolution of every profile.

The result isn't "gene X says you're fast." It's the integrated signal from every relevant genetic marker we can measure or infer, placing you relative to the population on seven dimensions of how your body and brain are wired.

The seven axes

We organise the seven dimensions into layers based on how strongly genetics determines the trait — because not all genetic signals are equal, and we think you deserve to know the difference.

Your chassis — Metabolic Pacing and Physical Architecture. These are your structural constraints, the highest-confidence signals in the model. Your CYP enzyme variants directly control how you clear caffeine, alcohol, and medications. Your muscle fibre composition is anchored by ACTN3, one of the most replicated performance-associated variants in human genetics. If your genome has something clear to say, it's here.

Your sensors and fuel — Sensory Gain and Fuel Partitioning. Real genetic signal for how your brain processes sensory input and how your body partitions energy between storage and expenditure, both enriched by genome-wide data from large population studies. Moderate confidence — meaningful, but modulated by environment in ways your chassis traits are not.

Your timing — Circadian Pacing and Stress Recovery. How your system regulates itself across daily and stress cycles. Confirmed by genome-wide studies involving hundreds of thousands of participants. Chronotype especially has one of the strongest genome-wide signals we work with — though modern light environments and social schedules can override what your biology prefers.

Your tendencies — Drive Profile. Genome-wide data on self-regulation, attention, and approach behaviour across hundreds of thousands of participants. Individual-level prediction is modest. We put this at Layer 3 for a reason: it's a starting point for self-experimentation, not a verdict about who you are.

Beyond individual axes, we compute cross-axis profiles — composites that capture how your dimensions interact. Sleep architecture, emotional bandwidth, motivation style — these emerge from combining multiple genome-wide signals into dimensions you can actually recognise in yourself.

Your archetype code

Every archetype is expressed as a code built from the seven axes. Each axis has two poles, each with a letter. If you lean toward a pole, its letter appears in your code. If you're strongly differentiated, the letter doubles. Mid-range axes don't appear at all. Letters always follow a fixed axis order.

FFPHHMDB reads: strong fast metabolism, power-built, strongly high sensory gain, morning-wired, driven, energy-burning. HD means most of your biology is mid-range with a couple of sharp edges. FFPPHHMRRDDBB means you're differentiated across the board.

Most people land in the 4–8 character range. Code length reflects differentiation, not quality. There's no better or worse architecture — only more or less self-knowledge about the one you have.

These aren't natural types, by the way. Human biology varies continuously. We impose operational categories on continuous distributions because they're easier to remember, easier to act on, and — most importantly — they make the next part possible.

The actual point: your biological peers

Genetics is the foundation. But a foundation alone isn't a building.

The missing context is phenotype — what actually happens in real bodies with real lives. And phenotype only becomes useful when it's sorted by biology.

Population-average recommendations fail individuals because they ignore biological variation. A supplement that works for one genetic architecture may do nothing for another, or cause problems. A training protocol that transforms one body type may injure another. Generic sleep advice may fight your circadian architecture rather than work with it. You've probably felt this: recommendations that work brilliantly for other people and do nothing for you.

The reason isn't that you're doing it wrong. It's that you're biologically different, and the recommendation wasn't built for your biology.

The only way to surface what actually works, for people like you, is to group people by genetic similarity and let signal emerge from their collective experience. That's what cohorts do. Within a cohort, patterns emerge: which supplements produce measurable effects for people wired like you, which protocols fail consistently for your architecture, what blood biomarker ranges are normal for your biology, what your biological peers have learned through trial and error that no clinical study has thought to ask.

This is observational data, not randomised trials. But it's observational data stratified by genetic similarity — something that doesn't exist anywhere else. Every user who logs their experience against their archetype makes the dataset more valuable for everyone who shares that architecture. Early users build the signal. Later users benefit from it. That flywheel is the core of what Humankind is.

Your genes tell us how you're wired. Your biological peers tell you what to do about it. Your code is how you find them.

What this is — and isn't

This is a probabilistic framework for understanding biological individuality. A cohorting mechanism that connects you to your biological peers. A platform where phenotypic data — what actually works — is sorted by genetic architecture rather than lumped into population averages.

It is not medical diagnosis. Not disease prediction. Not genetic determinism. Not a replacement for healthcare providers.

Your genes describe probabilities and tendencies, not destinies. Environment, lifestyle, development, and chance all matter. What we're offering is self-knowledge that enables better decisions — and a community of biological peers whose experience sharpens yours over time.

That's Humankind.

Next: The seven biological axes — a deep dive on each dimension, how we measure it, and what the confidence levels actually mean.

See also: The discovery network — how cohorts, expression tracking, and biomarker validation turn genetics into measurable, evolving insight.