Continuous Biological Evolution: The Case Against Static Health Plans

Why the Plan Model Persists

Static plans dominate for reasons that have nothing to do with biology.

They are legible. A protocol on a page can be read, sold, followed and adhered to, and it produces a satisfying sense of completion. They are commercially convenient: a plan can be a product with a fixed price and a defined delivery. And they map onto how people prefer to think about problems, as things to be solved and then finished.

The health domain also inherits the clinical episode model, in which a person presents with a complaint, receives an assessment and a course of action, and the episode closes. That model works well for acute problems with defined endpoints. It works poorly for the management of a trajectory over decades, which is what longevity practice actually is.

The Three Timescales

One reason a single plan cannot hold is that biological change happens on at least three separate timescales, and interventions on each behave differently.

Fast, meaning hours to days. Glucose responses, sleep quality, heart rate variability, inflammatory response to a hard session, hydration and mood. These fluctuate constantly and are readable in near real time through wearables and continuous monitoring. Reacting to individual readings here is almost always an error, because the noise dominates.

Medium, meaning weeks to months. Vitamin D concentration, HbA1c, membrane fatty acid composition, training adaptations, body composition, inflammatory markers, muscle mass. This is the timescale at which interventions can be evaluated fairly, and it is why quarterly review matches the biology.

Slow, meaning years to decades. Bone density, arterial calcification, cognitive trajectory, biological age composites, cumulative exposure. Interventions here cannot be evaluated on their own timescale within a useful period, which means they have to be justified by mechanism and by medium-term proxies.

A static plan implicitly picks one timescale and ignores the others. Daily trackers chase fast noise. Annual check-ups miss medium-scale drift. Longevity protocols justified only by slow-scale reasoning never get evaluated at all.

What Changes and How Fast

Cataloguing the perturbations makes the case concrete.

• Seasonal. Vitamin D status, inflammatory markers in some populations, activity patterns, sleep timing with daylight, food availability.
• Cyclical. Menstrual cycle effects on temperature, sleep, training capacity and several laboratory values. Training periodisation. Work cycles.
• Situational. Travel across time zones, altitude, heat, acute illness, an intense work period, poor sleep for a week.
• Developmental. Age-related changes in absorption, anabolic sensitivity, hormonal output and recovery capacity. Menopausal transition.
• Behavioural. A new training programme, a dietary change, starting or stopping alcohol, a new sleep schedule.
• Medical. New medication, a new clinical finding, a procedure, an injury.
• External. New evidence, changed regulatory status, a product reformulation or quality issue.

Each of these changes what the correct protocol is. Most of them are ordinary rather than exceptional, which means the state of being out of date is the normal condition of a static plan rather than an unusual failure of one.

Adaptive Does Not Mean Reactive

The obvious failure mode of rejecting static plans is to replace them with constant reaction, and that is worse.

A person who adjusts their protocol daily on the basis of a wearable readiness score is responding to noise. Heart rate variability varies substantially night to night for reasons including position, alcohol, room temperature and measurement artefact. Nightly stage classification from wrist devices carries real error. A single laboratory value can move twenty percent within normal biological variation.

Reactivity also has a psychological cost. Sleep tracking anxiety is documented, and a person whose behaviour is governed by a daily score has traded agency for a metric that may not reflect anything.

The distinction that matters is between signal and state. Fast-timescale data is useful in aggregate and as trend, not as instruction. The right response to a run of poor recovery data across two weeks is different from the right response to one bad night, and confusing the two produces either paralysis or churn.

What Replaces the Plan

Four properties define a process rather than a plan.

A current best protocol, held provisionally. There is still a set of actions, and it is understood as the best available answer for now rather than as a solution.

Explicit objectives and windows per element. Each element carries what it is for, when it will be judged and what would end it. This is what makes the protocol falsifiable rather than perpetual.

A review cadence matched to the timescales. Fast data read as trend, medium review quarterly, slow measures annually or biennially, with the understanding that slow measures cannot validate interventions on their own timescale.

Trigger-based revision. Discrete events, new medication, relocation, illness, a new finding, revise the protocol outside the schedule, because waiting for the next quarter is sometimes unsafe.

Under this model, the deliverable is not a document. It is the maintained alignment between what a person is doing and what their current state and the current evidence support, and the value accumulates through persistence rather than through any single decision.

Why This Is Hard to Do Alone

None of the above is conceptually difficult. It is operationally demanding in ways that predict failure for most people attempting it unaided.

It requires recording objectives and windows at the point of decision, when motivation is highest and documentation feels unnecessary. It requires a checklist for context change, because a new prescription or a relocation does not feel like a protocol event. It requires tracking an evidence base that publishes faster than anyone can read. It requires distinguishing signal from noise in fast data, which demands statistical intuition most people have not been given. And it requires the discipline to remove things, which is psychologically harder than adding them.

These are not moral failings. They are the predictable result of asking an individual to run a continuous process with no supporting infrastructure, which is precisely the gap that a platform is for.