A New Era of Longevity Science: Models of Aging, Rapamycin Trials and Biological Clocks
Longevity scientist Brian Kennedy joins Peter Attia to unpack what is actually driving aging: a slow, linear accumulation of damage plus a faster-moving oscillation that lifestyle and molecules can shift. He argues today's interventions mostly improve the oscillation — worth five to ten extra healthy years — while changing maximum lifespan would require a different class of discovery. Along the way: rapamycin, alpha-ketoglutarate, urolithin A, spermidine, and why a clinical-chemistry clock beats methylation clocks at predicting mortality.
Overview
Brian Kennedy, former CEO of the Buck Institute and now leading a large aging programme in Singapore, describes a research pipeline that runs from yeast and worms through killifish and mice into human clinical studies — the logic being that an intervention working across a billion years of evolutionary distance is more likely to work in people.
He pushes back on the popular reading of the hallmarks of aging: they are outputs and readouts, not twelve separate switches to be fixed one by one, because a single intervention such as rapamycin moves nearly all of them. His preferred picture is a homeostatic network that keeps a young body in a deep valley of health, with the surrounding hills of resilience gradually flattening as damage accumulates. That model reconciles a linear-looking accumulation of damage with the exponential rise in mortality described by Gompertz.
Superimposed on the linear line is an oscillation reflecting how well someone functions at their current damage level — and this, Kennedy suspects, is what sleep, exercise, nutrition and most supplements are actually moving. He details a six-month rapamycin study in Singapore in adults aged 40 to 60, an already-completed placebo-controlled trial of time-released alpha-ketoglutarate, and unpublished mouse work on urolithin A that sharply reduced frailty in males.
On measurement, he is blunt: several first-generation methylation clocks predict mortality worse than a passport does, whereas a clinical-chemistry clock built from roughly fifty routine markers outperforms them and — crucially — points to things a clinician can actually act on. The conversation closes on combining interventions, the fitness costs of long-lived mutants, and his caution that stacking twenty compounds is like mixing twenty colours of paint.
Key quotes
5All of the interventions that extend lifespan reduce chronic inflammation, or almost all of them.
Healthy aging is about maintaining homeostasis — maintaining a responsive network in your body that keeps you in equilibrium. And that network is highly malleable.
If we can give everybody five or ten years of extra health span, that's a major breakthrough in medicine.
Some of the first-generation clocks are worse than chronologic age. Your passport is better than they are at predicting mortality.
If you're taking 20 pills, it's like mixing 20 colours of paint together. You're going to get some ugly grey outcome.
Key ideas
9Aging still has no agreed definition
Decades of conferences have failed to produce a working definition of aging, which is why Kennedy helped organise a first international meeting on the physics of aging. His bet is that the answer will ultimately have to be written in equations, because biologists bring intuition where modellers bring physical principles.
Hallmarks are outputs, not twelve separate switches
Because a single intervention like rapamycin shifts nearly every hallmark at once, Kennedy reads them as readouts of an underlying network rather than independent causes. The idea that you fix all twelve and live forever, he argues, is completely wrong.
Restoring dynamic range, not switching pathways off
With aging, baseline mTOR signalling creeps up so the pathway can no longer be fully switched off. The goal of an intervention is not a supraphysiologic change but the restoration of the youthful on-and-off rhythm — mTOR on after a wound, an infection or a large meal, and off the rest of the time.
Valleys of health and flattening hills of resilience
When young you sit in a deep valley and get pulled back to health almost regardless of what you do. With age the surrounding hills flatten, so a viral infection or a fall can push you over into a failure state you cannot climb back out of.
Linear damage, exponential mortality
Large data sets suggest damage accumulates roughly linearly, while mortality rises exponentially. Flattening hills reconcile the two: a linear drop in hill height produces an exponential rise in the chance of going over the edge.
The oscillation is what most interventions move
A second component in biological clocks oscillates around the rising damage line and reflects how well you function at your current damage level. It explains why two fifty-year-olds can look decades apart — and why today's tools work around the edges rather than bending the slope.
A six-month rapamycin study in healthy adults
Roughly 150 to 200 participants aged 40 to 60, without diagnosed disease, take about five milligrams once weekly for six months. Outcomes include clocks, inflammatory cytokine panels, pulse wave velocity, DEXA, strength and cognition — with Attia openly sceptical that six months can produce a signal in most of them.
Alpha-ketoglutarate, urolithin A and spermidine in mice
Time-released alpha-ketoglutarate produced a modest lifespan gain but a dramatic reduction in frailty in mice; unpublished urolithin A work sharply reduced frailty in males but not females; spermidine both suppressed high-fat metabolic dysfunction and extended lifespan in an experiment not designed to measure it.
A clinical-chemistry clock that clinicians can act on
Built from around fifty routine markers in the NHANES dataset, it predicts mortality better than methylation clocks and better than cardiovascular risk scores. Its real advantage is that its principal components map onto treatable things — lipids, blood pressure, metabolic markers — rather than an uninterpretable number.
Practical takeaways
6- 1
Treat commercial biological age tests with caution 1:12:00
Four leading consumer tests run in duplicate at the same moment disagreed with each other — and often with themselves. Treat any single number you buy online as entertainment rather than as a measurement of how you are aging.
- 2
Aim for the best version of the state you are in 1:00:30
Rather than chasing 150 years, the realistic and still remarkable goal is seven to ten more years of far higher quality life. Sleep, training, nutrition and judicious medical management are the levers that move that oscillation.
- 3
Add one thing at a time and measure around it 2:04:20
Kennedy adds a single compound to a stable baseline, measures before and after, then removes it. Twenty simultaneous additions leave you unable to tell what helped, what did nothing, and what cancelled something else out.
- 4
Lean mass and fitness are the highest-leverage assets 1:56:40
Hazard ratios for low strength, low muscle mass and low cardiorespiratory fitness dwarf many familiar risks. Kennedy, a lifelong runner, added serious resistance training for precisely this reason.
- 5
Timing may matter as much as the molecule 1:41:00
Kennedy notices poorer runs within twenty-four hours of taking rapamycin and better training three to four days later, consistent with needing the pathway available around exercise. If you take a weekly compound, look at where your hardest sessions sit relative to it.
- 6
Hormone conversations deserve to be reopened 1:58:20
Kennedy suggests the question is not whether a woman should consider hormone replacement but whether there is a reason not to, and notes how strikingly low uptake remains worldwide. This is a discussion to have properly with your own physician.
Topics & chapters
16The questions that frame the field
Can aging be slowed, stopped or reversed — and what would count as proof? A cold open on rapamycin, inflammation and bending the slope.
The Buck Institute years
Kennedy on running the first institute devoted solely to aging, the lean funding years, and the star faculty who built the field.
The 2017–2018 inflection point
Why philanthropy and Silicon Valley suddenly discovered longevity, and how earlier hype cycles slowed investor confidence.
Singapore and the cross-species pipeline
Yeast, worms, flies, killifish, mice and humans in one programme — validating interventions across a billion years of evolution before taking them to clinical study.
What is actually causing aging?
The failure to define aging, the first conference on the physics of aging, and the hope that theoretical modellers can supply what biologists' intuition cannot.
Hallmarks, pillars and the network view
Why Kennedy sees the hallmarks as entrained outputs of one homeostatic network, and why ranking them by causality remains unresolved.
Inflammation as the recurring signal
Every new clock and nearly every life-extending intervention points back to chronic inflammation — readout, driver, or both.
mTOR from yeast to humans
The genome-wide yeast screen, the nutrient-sensing kinase it uncovered, and the creeping baseline that erodes youthful dynamic range.
Rapamycin: from Easter Island to geroprotection
How an antifungal became an immunosuppressant, why weekly trough-based dosing is a different proposition, and why it remains the small-molecule gold standard.
Resilience, valleys and failure states
The dynamic-systems picture of aging, why treating one disease still leaves fifty other hills to fall over, and what the hills actually are.
The epigenetic reprogramming thought experiment
If you reset one twin's epigenome to age twenty, what happens? Kennedy is unconvinced the epigenome holds primacy over DNA damage and mitochondrial change.
Which diseases are most inevitable
Cardiovascular and metabolic disease as the most preventable, cancer as the most stubborn, dementia as underdetermined — and the neglected role of immunity in all of them.
Designing the Singapore rapamycin trial
Dosing, participants, outcome measures, the cost of running such a study, and honest uncertainty about whether six months is long enough.
AKG, urolithin A, spermidine and NAD
How natural products enter the pipeline, the frailty results in mice, sublingual NAD, and one researcher's n-of-one exercise observations.
Clocks that clinicians can use
Why several methylation clocks underperform a passport, how a clinical-chemistry clock beats them on the NHANES data, and why actionability matters more than the number.
Combining interventions and staying safe
Fitness costs, cancelling effects, unproven gene and stem cell therapies abroad, and the case for scientists engaging with longevity clinics rather than ignoring them.