Show the roots,
not just the fruit.

The app may use the WHO adult guideline as a broad reference point, while scaling recommendations to the member’s capacity and Life Season.

Population

Adults, older adults, and named subpopulations covered by the WHO guideline.

Limits

A population guideline is not an individual prescription. Disability, pregnancy, chronic conditions, symptoms, and clinical advice can change what is safe.

Expected error

Activity entered manually or inferred from devices may be incomplete or misclassified.

What would change this

A superseding WHO guideline or strong evidence changing population recommendations.

For adults 18–60, seven or more hours is a consensus population recommendation; the Protocol does not treat a single short night as failure.

Population

Generally healthy adults aged 18–60.

Limits

Duration alone does not establish sleep quality or a disorder. Individual needs differ, especially during recovery or illness.

Expected error

Self-reported sleep and consumer-device estimates can disagree with clinical measurements.

What would change this

A newer consensus statement or a member-specific clinical plan.

Time of day and chronotype may influence perceived effort and performance, but findings are heterogeneous and should be treated as context—not certainty.

Population

Studies of athletes and physically active participants included in the review.

Limits

Small, heterogeneous studies; physiology and response vary. The app must not promise a fixed performance percentage.

Expected error

A modeled chronotype from behavior may differ from a validated questionnaire or laboratory assessment.

What would change this

Larger preregistered trials showing consistent individualized timing effects.

Most complex traits reflect many variants plus environment. A variant or risk score is context for a conversation, not destiny, diagnosis, or treatment advice.

Population

People represented by the underlying genomic studies; portability can vary by ancestry and dataset.

Limits

TLP does not perform genetic testing, diagnose disease, or validate uploaded raw files for clinical use.

Expected error

Consumer raw data can omit variants or contain errors; risk estimates depend on study population and model.

What would change this

Updated NHGRI guidance, improved ancestry portability evidence, or validated clinical integration.

Living Score is a product-specific summary of recorded protocol activity. Forecasts are scenario models, not validated measures of health or lifespan.

Population

Signed-in members with sufficient recorded protocol data.

Limits

The score is not a biomarker, diagnosis, medical-device output, or comparison of human worth.

Expected error

Missing logs, stale data, and changing weights can materially change the result.

What would change this

A scoring-version change, calibration study, or evidence that the presentation causes harmful interpretation.

Method

Weighted normalized activity signals by current protocol phase; exact implementation is versioned in source.

AI may summarize records, propose reversible actions, or explain a model. It may not fabricate source data, diagnose, prescribe treatment, or silently change deterministic truth.

Population

Eligible adult and teen members; additional feature restrictions apply to child accounts.

Limits

Language models can be wrong. Confidence and citations describe the available context, not a guarantee.

Expected error

Generated language can omit nuance or overstate causality; members can reject, modify, flag regret, or revoke learning consent.

What would change this

Model/provider changes, a safety incident, new evaluation evidence, or regulatory guidance.