On two of the hardest stages of this year’s Tour, riders on UAE Team Emirates-XRG drank sodium lactate. Deliberately, stirred into their carbohydrate, confirmed by their own nutritionist while the race was still running. The scientist behind the drink is George Brooks, who has spent most of a career arguing that lactate was filed wrong from the beginning.

Now think about what you were taught. Lactate is the waste. It’s what stings in the last thirty seconds of an interval, it’s what the cool-down is for, it’s the thing an ice bath supposedly chases out of your legs. That story sat in textbooks for decades and the people teaching it weren’t being sloppy. It’s still wrong where it counts.

Lactate is a fuel. Your body is making it right now, sitting still, with plenty of oxygen around, and trading it between cells so your heart and your slow fibres and your brain can burn it. It also carries information, changing which genes a muscle cell switches on.

The narrow claim: lactate doesn’t cause the burn, it isn’t evidence that anything failed, and the threshold everyone trains against is a reading rather than a wall.

This is an educational and strategic perspective, not personal medical advice. The views are the author’s own and not statements by Atlas Cove Lda.

The molecule you are making right now

Oxygen gave the old story its trigger. Muscle runs clean, demand outruns supply, an emergency pathway opens, lactate spills out as the residue. Tidy. It explains the sting and it certifies that you went past something.

The measurements don’t cooperate. Lactate turns over continuously across all sorts of cell types with oxygen everywhere and nothing failing anywhere (Brooks, 2018), which means the millimole per litre circulating in you as you read this isn’t left over from anything you did earlier in the week, it’s the standing balance of a system that has been making the stuff and burning it every minute you’ve been alive.

What’s actually going on is duller. Breaking down glucose gives you pyruvate, the enzyme that carries pyruvate to lactate is quick and sits close to equilibrium, so lactate turns up whenever glucose is going through fast, oxygen or no oxygen. A rising lactate number tells you how fast you’re burning carbohydrate, not the moment your oxygen ran short.Those two happen to move together when you go hard. That’s the whole reason a wrong explanation lasted so long.

Subscribe now

Muscle that sells and muscle that buys

Lactate becomes an economy because it moves. Cells push it across their membranes through the monocarboxylate transporters, and the two that matter here have opposite jobs. MCT4 holds lactate loosely and shifts a lot of it, which suits a fast glycolytic fibre making more than it can use. MCT1 sits in oxidative tissue, the slow fibres and the heart muscle, where lactate gets pulled in to be burned instead. Neither door runs one way. The same muscle can be selling one minute and buying the next (Brooks, 2018).

Under exercise that becomes constant trade, with a fast fibre letting go of it, the slow fibre next door taking it up and oxidising it, your heart burning it in preference to glucose, and your liver quietly rebuilding a share of it back into glucose again. Three quarters to four fifths of production never accumulates anywhere.

The brain is the part people refuse to believe. Van Hall and colleagues infused labelled lactate into six healthy volunteers and got most of the label back as labelled carbon dioxide, which is what oxidation looks like when you can follow the atoms. Lactate covered about 8 percent of the brain’s energy at rest, 19 percent once blood lactate was raised, and 27 percent during cycling at 75 percent of maximum oxygen uptake (van Hall et al., 2009). Six people. The lactate came from a drip rather than from effort. So treat the figures as indicative and the direction as solid: when you’re working hard, better than a quarter of what runs your brain is the stuff you were told to clear out.

The word that did the damage

The mistake lives inside one expression. Nobody argues about the burn, and the acidity of hard work is real and measurable. What “lactic acid” does is settle the question of cause by grammar, welding an acid and a lactate ion into a single object and handing you a suspect before anyone has looked at any evidence.

Follow the protons instead. Acidity comes overwhelmingly from spending ATP, because a proton gets released every time a cell splits ATP for energy, and hard work has the cell turning ATP over outside the mitochondria at a furious rate. The reaction that makes lactate does the opposite of what its name implies. Converting pyruvate to lactate takes a proton out of solution, which makes that step mildly alkalising (Robergs et al., 2004). Both things are true at once. That’s why the shorthand lasted.

Here’s the part that usually gets skipped. This correction gets overstated nearly every time somebody repeats it. Böning and Maassen have argued in print that “lactic acid” survives as a fair summary, since one lactate ion and roughly one proton come out of the same overall process, and they agree explicitly on the mechanism, that the proton only shows up when the fresh ATP gets spent (Böning and Maassen, 2008). Robergs has revised his own arithmetic since 2004.

So the defensible version is narrower than the popular one and quite a lot more interesting. Lactate isn’t the proton donor, both camps agree where the protons come from, and the argument that remains is about bookkeeping rather than chemistry. Nobody publishing in this field still claims lactate burns your legs. That’s the bit almost everyone got told.

The ache that turns up two days late

Second charge against lactate: the stiffness that arrives a day or two after unfamiliar work. The clock alone should have cleared it. Lactate is back at resting values inside the hour, and the soreness hasn’t even started yet.

The cleanest evidence here is old. Schwane and colleagues ran people downhill and on the flat and watched the two come apart completely, because flat running drove blood lactate up and left almost no soreness behind it, while downhill running produced pronounced delayed soreness in people whose blood lactate never rose at all (Schwane et al., 1983). The intensities weren’t matched, so this separates the two rather than disproving a link. The condition that hurt for days is the one that never raised lactate.

What downhill running supplies instead is eccentric loading, muscle making force while it lengthens, which is where the mechanical damage and the repair afterwards actually come from. Soreness after a new eccentric stimulus is closer to a receipt than to a warning that something needs clearing. It’s damage you asked for, and the repair is the point.

A fuel that also talks

Lactate’s newer job is carrying information. Hashimoto and colleagues put lactate on cultured muscle cells and watched expression change across six hundred and seventy-three genes, MCT1 up inside an hour, cytochrome c oxidase up inside six, the whole response converging on the machinery that builds mitochondria (Hashimoto et al., 2007).

The limits belong in the same breath, because results shaped like this travel faster than their evidence. Cultured rat cells, not a person. Ten and twenty millimoles per litre, and the higher figure sits above anything you would reach in blood. No exercise anywhere in it.

It doesn’t show that lactate drives your adaptation. What it makes plausible is that part of what a hard session buys you arrives through lactate talking rather than lactate burning, which would put the product of hard work inside the decision to get better at hard work. Still a hypothesis in humans. I like it more than I can currently defend it.

What the needle in your earlobe is reading

A lactate test is a run of small blood samples taken at rising intensities, and the curve appeals because it looks like it fixes a line you can train against.

The literature is messier than the curve. Faude and colleagues counted twenty-five separate definitions of the lactate threshold, which means one athlete tested on one afternoon can walk out with materially different numbers depending on which convention the lab happens to prefer (Faude et al., 2009). What the physiology holds is a transition with two edges: the point where lactate lifts off baseline, and the hardest intensity at which production and removal still cancel out.

Be fair to the test, though. That same review found thresholds correlate strongly with endurance performance, so they measure something real. The error is reading the number as a wall inside you rather than a report on how the system is running today. A threshold captures the balance between how fast you make lactate and how fast you clear it, and both sides move with training, sleep, heat and what you ate that morning.

I learned that the slow way, and not through sport. For a couple of years I had winter and spring episodes of severe migraine, flu-like aching and a fatigue that took the whole day with it. Every blood marker came back unremarkable. I kept waiting for one of them to go abnormal, because that is how I assumed bodies worked: something breaks, a number moves, you go and fix the thing. Nothing was broken. The problem was regulation, and no single marker reports on regulation. A lactate curve is honest in the same shape. It describes a balance, and it will not tell you which part to blame.

Subscribe now

Why drinking it is not the same as using it

Back to the bottle. If your body already trades lactate, putting some in by mouth looks like a way to add fuel without asking more of a gut that’s already saturated with carbohydrate. That’s the team’s reasoning and it’s a fair hypothesis.

Dose is the problem. Péronnet and colleagues fed people labelled lactate and labelled glucose together and followed both. The ingested lactate supplied about two and a half percent of the energy used. The glucose supplied about eight. The gap is mostly tolerance, because there’s a hard ceiling on how much lactate anyone can swallow before the gut objects (Péronnet et al., 1997).

I know that ceiling from the other side. On race day at Cascais I was taking in roughly 100 grams of carbohydrate an hour, which came out as 8 to 10 bars and 5 gels across the bike leg alone. That is already at the edge of what a stomach will accept while you’re working, and it is not a pleasant thing to do to yourself for five and a half hours. Whatever else you want to add has to fit inside what’s left of that budget. There isn’t much left.

The outcome trials have been cool as well. Sixteen trained cyclists on a race-like protocol got better buffering and lower perceived effort, with performance unchanged (Bordoli et al., 2024). So the reframe stands up and the drink doesn’t, at least not yet. Nobody involved is claiming otherwise, and UAE’s own nutritionist said during the race that the performance effect remains an open question. A better formulation might lift the tolerable dose and change all of this. Worth watching. None of it touches the biology underneath.

Three sessions a week and no laboratory

You train three or four times a week around a job, and you’re never going to have a needle in your earlobe. So what actually changes?

First, how you read getting fitter. Bergman and colleagues took nine untrained men through nine weeks and then measured lactate kinetics two different ways, and the interesting part is that the two ways disagree: held at the same absolute power as before training, blood lactate fell 41 percent because the men were simply producing less of it with clearance unchanged, while held at the same relative intensity, meaning a harder effort scaled to their new fitness, it was clearance and oxidation that went up instead (Bergman et al., 1999). Nine men, one cycling protocol. Treat the numbers as illustrative. The shape is the point: fitness is two things at once, a production side and a disposal side, and a session improves whichever one you loaded.

Second, how much authority the zones on your watch have earned. Zones are a usable model and I use them, because a continuous variable has to be cut somewhere before anyone can plan a week. Across the months before Cascais mine came out at 72 percent low intensity, 27 percent threshold and 2 percent high, and that distribution did its job. Or rather, the distribution did its job while the boundaries between the zones stayed an estimate the whole time. A line taken off a formula, or off a test you did in March, is a guess about you today.

Two checks beat any single measurement. Has your pace or power at a fixed easy heart rate moved across the last eight weeks? That asks the same question a lactate curve asks, and it costs nothing. Then the one that needs no equipment at all: could you repeat today’s easy session tomorrow without dreading it? An easy day only works if it’s genuinely easy, and the usual way a good week rots is that the bottom end creeps upward while nobody is looking at it.

Where this leads

Which is close to why an Atlas Cove week treats a measurement as something obliged to answer a question you already have, rather than as a score to be collected and filed, because a number earns its place on the page only when you can say in advance what a high reading and a low reading would each make you do differently. A number you can’t act on differently is decoration.

The Atlas Cove Method

Lactate shows how long a wrong explanation survives when it’s vivid, teachable and roughly lines up with the sensation it claims to explain. Forty years. Plenty of cool-downs are still being done for it.

You’re making some right now. Your heart is buying it. So, apparently, are the fastest cyclists in the world, though on that last one the evidence hasn’t caught up with the enthusiasm yet.

This is an educational and strategic perspective, not personal medical advice. The views are the author’s own and not statements by Atlas Cove Lda.

Sources

1. Brooks, G. A. (2018). The science and translation of lactate shuttle theory. Cell Metabolism, 27(4), 757-785. DOI: 10.1016/j.cmet.2018.03.008

2. van Hall, G., Strømstad, M., Rasmussen, P., Jans, O., Zaar, M., Gam, C., Quistorff, B., Secher, N. H., & Nielsen, H. B. (2009). Blood lactate is an important energy source for the human brain. Journal of Cerebral Blood Flow & Metabolism, 29(6), 1121-1129. DOI: 10.1038/jcbfm.2009.35

3. Robergs, R. A., Ghiasvand, F., & Parker, D. (2004). Biochemistry of exercise-induced metabolic acidosis. American Journal of Physiology - Regulatory, Integrative and Comparative Physiology, 287(3), R502-R516. DOI: 10.1152/ajpregu.00114.2004

4. Böning, D., & Maassen, N. (2008). Point: Lactic acid is the only physicochemical contributor to the acidosis of exercise. Journal of Applied Physiology, 105(1), 358-359. DOI: 10.1152/japplphysiol.00162.2008

5. Schwane, J. A., Watrous, B. G., Johnson, S. R., & Armstrong, R. B. (1983). Is lactic acid related to delayed-onset muscle soreness? The Physician and Sportsmedicine, 11(3), 124-131. DOI: 10.1080/00913847.1983.11708485

6. Hashimoto, T., Hussien, R., Oommen, S., Gohil, K., & Brooks, G. A. (2007). Lactate sensitive transcription factor network in L6 cells: activation of MCT1 and mitochondrial biogenesis. The FASEB Journal, 21(10), 2602-2612. DOI: 10.1096/fj.07-8174com

7. Faude, O., Kindermann, W., & Meyer, T. (2009). Lactate threshold concepts: how valid are they? Sports Medicine, 39(6), 469-490. DOI: 10.2165/00007256-200939060-00003

8. Péronnet, F., Burelle, Y., Massicotte, D., Lavoie, C., & Hillaire-Marcel, C. (1997). Respective oxidation of 13C-labeled lactate and glucose ingested simultaneously during exercise. Journal of Applied Physiology, 82(2), 440-446. DOI: 10.1152/jappl.1997.82.2.440

9. Bordoli, C., Varley, I., Sharpe, G. R., Johnson, M. A., & Hennis, P. J. (2024). Effects of oral lactate supplementation on acid-base balance and prolonged high-intensity interval cycling performance. Journal of Functional Morphology and Kinesiology, 9(3), 139. DOI: 10.3390/jfmk9030139

10. Bergman, B. C., Wolfel, E. E., Butterfield, G. E., Lopaschuk, G. D., Casazza, G. A., Horning, M. A., & Brooks, G. A. (1999). Active muscle and whole body lactate kinetics after endurance training in men. Journal of Applied Physiology, 87(5), 1684-1696. DOI: 10.1152/jappl.1999.87.5.1684

Nine volunteers. Eight hours in bed for three nights, then four hours a night for four nights running, with bedtime held at eleven while the alarm walked back to three, which is roughly what a hard fortnight does to anyone holding down a job and an early start.

Brunner, Dijk and Borbély ran it in 1993. Slow wave sleep came through the short nights largely untouched. Stages one and two went, and REM went with them (Brunner et al., 1993).

So a night doesn’t shrink evenly. Take a fifth away and you don’t lose a fifth of everything in it, because the body triages, defends the front of the night, and charges almost the whole shortfall to the back end where REM lives. The order of a night decides what you get out of it, and total hours are a much weaker predictor than nearly everyone assumes.That’s the narrow claim. Duration still counts, and a genuinely short night is short in every compartment at once.

This is an educational and strategic perspective, not personal medical advice. The views are the author’s own and not statements by Atlas Cove Lda.

Why the night is built in that order

Two systems decide when you sleep, and they answer to different things. The first is homeostatic: pressure builds the longer you stay awake and drains once you’re under. The second is circadian, run by an internal pacemaker on roughly a twenty-four hour cycle, and it doesn’t care how tired you are. Borbély and colleagues set out the interaction of those two, Process S and Process C, as the arrangement that best reproduces both when human sleep happens and how deep it goes, which is a model that has survived thirty-four years and a formal reappraisal by its own authors (Borbély et al., 2016). The messenger on the homeostatic side is adenosine, which accumulates across a waking day. Caffeine sits on its receptors. Hold that thought.

Two clocks instead of one has an awkward consequence. Seven hours starting at eleven and seven hours starting at two aren’t the same purchase, because the circadian system is somewhere different in its cycle and permits different things.

A night flight and an ordinary night don’t do the same work. Same duration, different night.

Subscribe now

Deep sleep goes early, REM waits until morning

The cleanest evidence comes from forced desynchrony. It’s a slightly brutal protocol: you put people on a day length their body clock can’t follow, so their sleep episodes land at every circadian phase in turn and the two systems come apart where you can watch them separately.

Dijk and Czeisler kept eight men on twenty-eight hour days for a month with no time cues. The split was clean. REM sleep carried a strong circadian rhythm whose crest sat just after the low point of core body temperature, while slow wave activity fell across every single sleep episode and showed only a small circadian wobble that didn’t track sleep propensity at all (Dijk and Czeisler, 1995).

Then comes the detail that turns this from a curiosity into something you can act on, which is that REM also responded to how long the person had already been asleep and reached its highest values of the whole protocol when the scheduled sleep episode happened to end at the hour those men would normally have woken up anyway.

The hours where REM is most available are exactly the hours an early alarm removes. Structural fact. Nothing to do with your discipline.

The part I had backwards

I assumed for years that deep sleep was the thing to defend after forty. Everyone says so. It’s where the repair happens, and the repair is what you want.

The first half of that is right. Slow wave sleep carries the largest growth hormone pulse of the day, tissue repair runs hardest there, and it’s when parasympathetic activity peaks, which ties it straight back to the autonomic account part one of this series was about.

The second half is where I was wrong. Van Cauter, Leproult and Plat pooled records from 149 healthy men aged sixteen to eighty-three. Deep slow wave sleep filled 18.9 percent of the night in the sixteen to twenty-five year olds. In the thirty-six to fifty year olds it filled 3.4 percent. The difference had been absorbed by lighter sleep, and REM hadn’t significantly dropped yet at that point. Growth hormone output fell alongside it and tracked slow wave sleep independently of age (Van Cauter et al., 2000).

Those were cross-sectional comparisons between age bands, all of them men, so nobody was followed as they aged and none of it licenses a claim about women. It still reorders the priority. By the time you’re forty-five, deep sleep has mostly already done its declining, and REM is the compartment still substantially there and still being cut every morning. I’d been guarding the part that had largely gone and spending the part that hadn’t.

Nothing announces a missing REM

Losing REM produces no symptom you’d notice. That’s most of why it disappears without complaint, and it’s what makes the trade feel free.

Wagner, Gais and Born used the night’s own structure as their instrument and tested recall of emotional against neutral material across an early-sleep window rich in slow wave sleep and then across a late-sleep window rich in REM, and the emotional advantage turned up across the late one, which they read as consistent with a supporting role for the REM that dominates the second half of a night (Wagner et al., 2001).

The popular version of that finding gives REM a good deal more than the evidence does, motor skill learning in particular, where the lighter non-REM stages have the better claim. Worth knowing if you’ve read otherwise.

Kept at its real size the point holds. Cutting the tail off a night removes something specific rather than a slice of everything, and what goes is emotional processing, which no training log and no blood panel is built to see. Week after week it shows up as a slightly shorter fuse. Almost nobody blames the alarm clock.

Subscribe now

Falling asleep is a heat problem

Everyone says keep the bedroom cool. The specifics are better than that.

Kräuchi and colleagues ran a whole field of candidate predictors against how fast people actually dropped off, and the one that won was the distal to proximal skin temperature gradient, roughly how much warmer your hands and feet run than your trunk, which beat core body temperature, the rate at which core temperature was changing, heart rate, melatonin onset and how sleepy people said they felt (Kräuchi et al., 2000).

So you aren’t simply getting colder. You’re opening the taps at the extremities and dumping heat, and the cool room matters because it gives that heat somewhere to go. Cold feet and a cold bedroom fight each other. Socks with the window open. Took me an embarrassingly long time to accept that one.

Two things that rearrange a night without shortening it

Alcohol is the clearest case, and the honest version is more interesting than the disapproving one. Twenty-seven studies pooled in a 2025 systematic review and meta-analysis found that falling asleep faster, which is the effect the drink is actually bought for, only registered at around five standard drinks, while total sleep time, sleep efficiency and night-time waking all stayed too uncertain to call and the one thing that came out clearly was a dose-response effect on REM, delayed and reduced at roughly two standard drinks and worsening as the dose climbed (Gardiner et al., 2025).

None of which says the wine at dinner has to go. It says the thing it’s bought for is largely unevidenced, while the cost lands on the compartment that’s already most exposed.

Caffeine works on the other system. It occupies adenosine receptors without switching them on, so the pressure keeps building while your ability to feel it drops away, and at a half-life around five to six hours the afternoon cup is still substantially present at midnight. Drake and colleagues gave 400 mg, about two strong coffees, at zero, three and six hours before habitual bedtime in a randomised placebo-controlled crossover, and every one of those timings cost more than an hour of sleep, with deep sleep at the six hour point falling to roughly forty-nine minutes against seventy-one on placebo (Drake et al., 2013).

Here’s the part that explains the behaviour. At six hours the instruments caught it and the participants’ own diaries didn’t. The cost wasn’t absent. It was invisible from the inside, which is a fair explanation for why people ignore this advice in perfectly good faith. Twelve participants though, and a home EEG headband rather than a laboratory rig. Hold it loosely.

The number on your wrist that it can't actually see

Most people will check all of this against a wrist, and the devices deserve neither the dismissal nor the trust they usually get.

Chinoy and colleagues put seven consumer trackers against laboratory polysomnography across three nights in thirty-four healthy young adults, one night deliberately disrupted. Every device detected sleep well, sensitivity 0.93 or better. Specificity landed between 0.18 and 0.54. Most devices left thirty to fifty percent of both deep and REM sleep unidentified on average, usually relabelling it light sleep, and they got worse on the poorer nights (Chinoy et al., 2021).

So last night’s stage chart is an estimate, it degrades on exactly the broken nights that made you open the app in the first place, and averaging across a week doesn’t rescue it either, because a weekly figure sitting near the truth may have arrived there by cancelling errors in both directions.

This one irritates me. The least trustworthy number on the screen is the one everybody reads, and the most trustworthy one is the timestamp nobody looks at. Your device knows when you went to bed and when you got up. That’s a timing measurement, and timing is where the evidence actually is.

The lever is free, and you've probably already given it away

Windred and colleagues built a sleep regularity index out of more than ten million hours of accelerometer data across 60,977 UK Biobank participants, mean age sixty-three, then followed deaths for an average of a little over six years, and against the least regular fifth of that cohort the four more regular fifths carried twenty to forty-eight percent lower all-cause mortality risk. Then they put regularity and duration into equivalent models. Regularity won, minimum hazard ratio 0.52 against 0.69 for duration before adjustment, 0.70 against 0.76 after (Windred et al., 2024).

It’s observational. The authors say plainly that regularity might be a cause of mortality risk or simply a marker of it, nobody has shown that fixing your schedule lengthens your life, and somebody drifting into illness would produce this same signature. Read at its proper strength it still says something worth having: the best-evidenced input here costs nothing, needs no equipment, and is the first thing a calendar takes.

Now the bill. I live in Portugal, where dinner starts at an hour I was raised to consider bedtime, so holding a wake time inside a one-hour band across all seven days costs me evenings I’d rather have and I don’t always pay it. That’s the real trade. Anyone telling you it’s free hasn’t tried to hold it through a Portuguese summer.

Two weeks and no hardware will tell you where you stand. Write down when you actually get up, then compare the weekday average against the weekend, and read a gap above an hour as one instruction handed to your circadian system every Friday and reversed every Sunday. Then take your last coffee and add six hours. If that lands after your intended lights out, you’ve got something to change tonight.

Where this leads

The strange thing about the Ironman I raced at Cascais in October 2025 is how much of it I could account for afterwards. Training distribution to the percent. Heart rate by discipline, 143 in the water, 142 on the bike, 144 on the run. Carbohydrate to the gram. My sleep timing across that whole build sits in no file anywhere, not in the training log, not in a spreadsheet, not on a watch, and of everything I recorded that year it’s the one variable carrying the mortality evidence I’ve just spent two thousand words on.

An Atlas Cove week puts sleep timing in the same column as the training load and the autonomic data, because a variable that decides what a night returns shouldn’t be filed under background conditions. Guests leave knowing which hour of their own night is the one to defend.

Atlas Cove - The Return

Eight hours describes an opportunity. It says nothing about whether the night contained the deep sleep that repairs, the REM that processes, or the regularity that outpredicts both. Almost none of what's available here needs more time in bed, which nobody has going spare. It needs the same time, held.

This is an educational and strategic perspective, not personal medical advice. The views are the author's own and not statements by Atlas Cove Lda.

Sources

1. Borbély, A. A., Daan, S., Wirz-Justice, A., & Deboer, T. (2016). The two-process model of sleep regulation: a reappraisal. Journal of Sleep Research, 25(2), 131-143. DOI: 10.1111/jsr.12371

2. Dijk, D. J., & Czeisler, C. A. (1995). Contribution of the circadian pacemaker and the sleep homeostat to sleep propensity, sleep structure, electroencephalographic slow waves, and sleep spindle activity in humans. The Journal of Neuroscience, 15(5 Pt 1), 3526-3538. DOI: 10.1523/JNEUROSCI.15-05-03526.1995

3. Brunner, D. P., Dijk, D. J., & Borbély, A. A. (1993). Repeated partial sleep deprivation progressively changes the EEG during sleep and wakefulness. Sleep, 16(2), 100-113. DOI: 10.1093/sleep/16.2.100

4. Van Cauter, E., Leproult, R., & Plat, L. (2000). Age-related changes in slow wave sleep and REM sleep and relationship with growth hormone and cortisol levels in healthy men. JAMA, 284(7), 861-868. DOI: 10.1001/jama.284.7.861

5. Wagner, U., Gais, S., & Born, J. (2001). Emotional memory formation is enhanced across sleep intervals with high amounts of rapid eye movement sleep. Learning & Memory, 8(2), 112-119. DOI: 10.1101/lm.36801

6. Kräuchi, K., Cajochen, C., Werth, E., & Wirz-Justice, A. (2000). Functional link between distal vasodilation and sleep-onset latency? American Journal of Physiology-Regulatory, Integrative and Comparative Physiology, 278(3), R741-R748. DOI: 10.1152/ajpregu.2000.278.3.R741

7. Gardiner, C., Weakley, J., Burke, L. M., Roach, G. D., Sargent, C., Maniar, N., Huynh, M., Miller, D. J., Townshend, A., & Halson, S. L. (2025). The effect of alcohol on subsequent sleep in healthy adults: a systematic review and meta-analysis. Sleep Medicine Reviews, 80, 102030. DOI: 10.1016/j.smrv.2024.102030

8. Drake, C., Roehrs, T., Shambroom, J., & Roth, T. (2013). Caffeine effects on sleep taken 0, 3, or 6 hours before going to bed. Journal of Clinical Sleep Medicine, 9(11), 1195-1200. DOI: 10.5664/jcsm.3170

9. Chinoy, E. D., Cuellar, J. A., Huwa, K. E., Jameson, J. T., Watson, C. H., Bessman, S. C., Hirsch, D. A., Cooper, A. D., Drummond, S. P. A., & Markwald, R. R. (2021). Performance of seven consumer sleep-tracking devices compared with polysomnography. Sleep, 44(5), zsaa291. DOI: 10.1093/sleep/zsaa291

10. Windred, D. P., Burns, A. C., Lane, J. M., Saxena, R., Rutter, M. K., Cain, S. W., & Phillips, A. J. K. (2024). Sleep regularity is a stronger predictor of mortality risk than sleep duration: a prospective cohort study. Sleep, 47(1), zsad253. DOI: 10.1093/sleep/zsad253

“Should I take creatine?” is the question I get more than any other, and I’ve stopped answering it straight.

Not out of caution. The honest answer turns on a number nobody in the conversation has, me included.

Here is how it looks like. Creatine adds no fuel like the way a carbohydrate adds fuel, and it doesn’t make fibres pull harder. What it does is hold your ATP concentration steady across the short window where you’re spending ATP faster than you can remake it. That window opens early in any hard exertion. It’s why the fourth heavy set feels like a different exercise from the first.

So, the narrow claim: the effect you can rely on lands on repeatable work rather than on peak output. That’s not a claim about potency. Plenty of people take it and notice nothing at all, and they aren’t doing it wrong.

This is an educational and strategic perspective, not personal medical advice. The views are the author’s own and not statements by Atlas Cove Lda.

What actually runs out

Your muscle carries almost none of the molecule it spends.

Cytosolic ATP sits at around three to six millimolar depending on the cell type. The phosphocreatine pool beside it runs as high as thirty. That imbalance is deliberate (Wallimann et al., 2011).

Creatine kinase works the exchange both ways. Wherever ATP is being burned it strips a phosphate off phosphocreatine and hands it to spent ADP, and near the mitochondria, where ATP gets made, it runs the same reaction backwards. Wallimann and colleagues describe the result as a buffer doing two jobs at once, holding the ATP concentration steady across seconds and also shuttling high-energy phosphate between the places it’s produced and the places it’s needed.

Phosphocreatine gets spent so ATP doesn’t have to fall. That’s how a muscle goes straight into near-maximal output without every enzyme that assumes stable ATP falling over in the first second.

What the arrangement can’t do is last. The pool drains fast under real load, and once it’s low you’re running on slower machinery that brings its own costs. That’s the part of a set which stops feeling like the start of it.

Subscribe now

Five of eight

If the buffer is what empties, the interesting question is how quickly it refills.

Greenhaff and colleagues went at that directly in 1994. Eight men, twenty grams of creatine a day for five days, then biopsies of the vastus lateralis at rest and at twenty, sixty and a hundred and twenty seconds into recovery from a single intense electrically evoked isometric contraction (Greenhaff et al., 1994).

Five of the eight raised total muscle creatine by about a quarter. In those five, and only in those five, phosphocreatine refilling across the second minute improved by roughly thirty-five percent. The other three gained five to seven percent and got nothing measurable.

Three things about that study should travel with the number. The effect sat in the second minute rather than the first, which points at the slower phase of refilling. The contraction was evoked electrically, so this is cleaner physiology than it is sport. And there was no placebo arm, with the supplemented trial always run second, which is the sort of detail that should make you hold the figure loosely and the mechanism firmly.

A molecule working on the gap between efforts isn’t working on the effort. Nearly everything else follows from that. It won’t add weight to your one-rep max in any way you’d notice. It might bring your fifth set closer to your first.

Most of the tank is already full

Muscle creatine gets measured in millimoles per kilogram of dry muscle, and untrained men sit somewhere near a hundred and twenty.

Hultman and colleagues ran thirty-one men through four intake schedules, with results tidier than nutrition research usually manages (Hultman et al., 1996). Six days at twenty grams a day lifted total creatine by about twenty percent, roughly twenty-three millimoles per kilo of dry mass. Two grams a day after that held the level flat for four more weeks. The group that stopped was back at baseline by day thirty-five. A separate group took three grams a day with no loading phase at all and landed in the same place by day twenty-eight.

So the loading week is optional. It buys you three weeks.

Read all of that as a record of what a trial gave people, never as a prescription. The shape matters more than the schedule: there’s a ceiling, most of us start fairly close to it, and the gap that’s left decides whether anything happens.

Which is where diet walks in, since you get creatine out of meat and fish as well as making your own. Burke and colleagues found vegetarians at roughly 117 millimoles per kilo of dry mass where non-vegetarians sat at 130, and the vegetarians on creatine gained more muscle creatine than any other arm of the trial (Burke et al., 2003). Across the whole sample, the size of the gain ran inversely to where people started, at r equals minus 0.77.

Put that beside Greenhaff’s five and three and the inconsistency stops being mysterious. I eat meat most days and I lift three times a week, so my own headroom was probably small. No complaint there. It’s the molecule doing precisely what the papers say it does.

The dull mechanism is the one that holds

What connects a fuller phosphocreatine pool to a bigger, stronger muscle is disappointingly ordinary, and it’s worth stating plainly, because the flashier explanations get more airtime than they’ve earned.

More buffer lets you tolerate slightly more training volume. More volume is more mechanical stimulus, and more stimulus is more adaptation. The compound builds nothing itself. It makes the building cheaper.

Chilibeck and colleagues pooled twenty-two trials covering seven hundred and twenty-one participants, all of them drawn out of studies with a mean age of fifty or above, resistance training two or three times a week for anywhere between seven and fifty-two weeks. Creatine alongside training gave a mean difference of 1.37 kg of lean tissue over training alone, confidence interval 0.97 to 1.76, with standardised mean differences of 0.35 for chest press and 0.24 for leg press (Chilibeck et al., 2017).

Now the figure that should actually shape what you expect. Ten of those twenty-two trials found no creatine effect at all. A pooled estimate describes a population. You’re one person carrying an unknown amount of room.

There’s a second explanation in circulation that I treat more carefully than most people seem to. Creatine is osmotically active, it pulls water into the cell, and cell swelling has been put forward as an anabolic signal in its own right. Reasonable idea. Awkward measurement problem, though: intracellular water counts toward lean tissue mass on a DXA scan, so a trial reporting extra lean tissue can’t cleanly separate contractile protein out of fluid. Burke’s correlation of 0.61 between the creatine gain and the lean tissue gain, plus 2.4 kg against 1.9 kg in the two supplemented arms, fits real hypertrophy and fits hydration equally well. I’d call it unsettled, and I’d be glad to be shown otherwise.

Subscribe now

The brain takes the harder route

Neurons run the same creatine kinase system and meet demand spikes that outrun oxidative supply for the reasons muscle does. Getting the creatine in there is the problem.

Roschel and colleagues put the brain rise at roughly five to ten percent under supplementation, about half what muscle manages, with the protocol that would reliably move it still undetermined, and at least one study measuring both tissues found muscle up and brain unchanged (Roschel et al., 2021).

What that rise does splits along one line, and the line is strain.

McMorris and colleagues gave nineteen people creatine or placebo for seven days, then kept them awake for twenty-four hours with light exercise scattered through it. The creatine group fell off less on random movement generation, choice reaction time, static balance and mood. Verbal and spatial recall showed nothing. Neither did plasma catecholamines or cortisol (McMorris et al., 2006).

That’s a smaller loss under load, and it isn’t a gain in a rested person. The difference matters more than the marketing lets on, because trials in rested, well-slept adults come back inconsistent, and Roschel’s review declines to conclude anything firm about them.

One appealing bridge collapses when you look at it properly. The best-known cognitive trial recruited only vegetarians and vegans, chosen precisely because their baseline sits low, and Roschel’s review notes both that a comparable trial in meat eaters found nothing and that brain creatine looks similar in vegetarians and omnivores anyway. The diet argument that works downstairs doesn’t obviously work upstairs. I’d rather say so than let two findings lend each other weight neither has earned.

The blood test that looks worse than you are

This one shows up in a doctor’s office instead of a gym, and it’s the part I’d most want you to have read.

Creatinine, the marker routine panels use to stand in for kidney function, is what creatine breaks down into. So its concentration reports on your muscle mass and on your dinner as well as on your filtration. Antonio and colleagues note that supplementation and a meat-heavy diet can both raise blood and urinary creatinine, and that rises arriving that way are unlikely to mean your kidneys are failing (Antonio et al., 2021).

The pattern across the studies they gather isn’t uniform. Twelve found no rise at all. Eight found a rise that stayed inside the reference range. Two went past the normal limit.

A number that moves for a non-pathological reason still needs explaining rather than dismissing. If a panel comes back with a raised creatinine or a dropped eGFR, tell the clinician what you’re taking and how much meat you eat, so the value gets read properly instead of argued about. None of that replaces a clinical judgement about your own kidneys, and anyone with existing renal disease is in a different conversation entirely.

Twelve weeks, one number

Almost nothing in a supplement cupboard can be tested by the person who bought it. Creatine can, because the mechanism predicts something specific enough to watch for.

What it predicts is a smaller gap between your first set and your last.

So write down the reps you get at a fixed load across three or four sets. Do it this week, then do it again in twelve. If the drop-off hasn’t narrowed and nothing else in your training changed, whatever room you had was small, and this particular lever belongs to somebody else.

Here’s my confession. I took creatine for years and never once ran that test. I stopped during one Ironman build because nothing in the endurance log had moved, which was a daft reason, since an endurance log records pace and heart rate and volume and none of those are quantities a phosphocreatine buffer touches. The one thing it plausibly affected was the quality of my last set in the gym. That was the session I’d already decided was optional.

There’s an order to this, though, and it comes before any purchase. The benefit travels through training volume you can tolerate, so buffering a stimulus you aren’t applying buffers nothing. The strength work that part two of the applied physiology series called the highest-return hour in a crowded week comes first. A compound that makes it marginally more repeatable comes after. Get the order wrong and the supplement question turns unanswerable, which is roughly why it never gets answered.

Where this leads

Supplements are worth something at the margins and worth nothing underneath. That’s my position, and creatine hasn’t shifted it. What creatine does have going for it is that it’s cheap, heavily studied, and unusually honest about its own limits once you read past the label.

An Atlas Cove week starts by measuring the things a later decision has to rest on, rather than handing anyone a list of what to take. Whether a molecule like this does anything for you turns on your headroom, your diet, your training and what your bloods already say. A shelf category answers none of that.

The Atlas Cove Method

The loudest version of a supplement’s story almost never matches the version the physiology will carry. Here the supported version is small: a buffer that shortens the refill between hard efforts, in the people who had room for it, which shows up as slightly more tolerable volume and eventually as slightly more muscle and slightly more strength, with a brain case that’s real under strain and unresolved at rest.

Small, and better evidenced than most of what sits beside it on that shelf. Whether any of it is true for you is a question your own notebook answers in twelve weeks, and nobody selling it to you can answer that at all.

This is an educational and strategic perspective, not personal medical advice. The views are the author's own and not statements by Atlas Cove Lda.

Sources

1. Wallimann, T., Tokarska-Schlattner, M., & Schlattner, U. (2011). The creatine kinase system and pleiotropic effects of creatine. Amino Acids, 40(5), 1271-1296. DOI: 10.1007/s00726-011-0877-3

2. Greenhaff, P. L., Bodin, K., Söderlund, K., & Hultman, E. (1994). Effect of oral creatine supplementation on skeletal muscle phosphocreatine resynthesis. American Journal of Physiology, 266(5 Pt 1), E725-E730. DOI: 10.1152/ajpendo.1994.266.5.E725

3. Hultman, E., Söderlund, K., Timmons, J. A., Cederblad, G., & Greenhaff, P. L. (1996). Muscle creatine loading in men. Journal of Applied Physiology, 81(1), 232-237. DOI: 10.1152/jappl.1996.81.1.232

4. Burke, D. G., Chilibeck, P. D., Parise, G., Candow, D. G., Mahoney, D., & Tarnopolsky, M. (2003). Effect of creatine and weight training on muscle creatine and performance in vegetarians. Medicine & Science in Sports & Exercise, 35(11), 1946-1955. DOI: 10.1249/01.MSS.0000093614.17517.79

5. Chilibeck, P. D., Kaviani, M., Candow, D. G., & Zello, G. A. (2017). Effect of creatine supplementation during resistance training on lean tissue mass and muscular strength in older adults: a meta-analysis. Open Access Journal of Sports Medicine, 8, 213-226. DOI: 10.2147/OAJSM.S123529

6. Roschel, H., Gualano, B., Ostojic, S. M., & Rawson, E. S. (2021). Creatine supplementation and brain health. Nutrients, 13(2), 586. DOI: 10.3390/nu13020586

7. McMorris, T., Harris, R. C., Swain, J., Corbett, J., Collard, K., Dyson, R. J., Dye, L., Hodgson, C., & Draper, N. (2006). Effect of creatine supplementation and sleep deprivation, with mild exercise, on cognitive and psychomotor performance, mood state, and plasma concentrations of catecholamines and cortisol. Psychopharmacology, 185(1), 93-103. DOI: 10.1007/s00213-005-0269-z

8. Antonio, J., Candow, D. G., Forbes, S. C., Gualano, B., Jagim, A. R., Kreider, R. B., Rawson, E. S., Smith-Ryan, A. E., VanDusseldorp, T. A., Willoughby, D. S., & Ziegenfuss, T. N. (2021). Common questions and misconceptions about creatine supplementation: what does the scientific evidence really show? Journal of the International Society of Sports Nutrition, 18(1), 13. DOI: 10.1186/s12970-021-00412-w

Now hold that next to your last five years.

If you’re the kind of person who reads this newsletter, you probably have a folder. Panels going back years. A wearable on your wrist since 2021. Maybe a scan. You’ve listened to the podcasts and you can hold your own on ApoB. And if you’re honest about it, the file has grown considerably more than you have.

That’s not a personal failure. It’s the predictable output of a model, and the model has a name.

Camp one made a promise the trials never made

Camp one says: measure more. The implied theory is that health behaviour is bottlenecked on information, and that once you can see the number, you’ll act on it. It’s a seductive theory because it’s true of almost everything else in a competent adult’s life. It’s mostly false here.

The most-cited evidence in camp one’s favour is the wearable literature, and it’s worth being precise about what it shows. An umbrella review in Lancet Digital Health in 2022, led by Ferguson at the University of South Australia, pooled 39 systematic reviews and meta-analyses. Activity trackers did move behaviour: roughly 1,800 extra steps a day, about 40 minutes more walking, around a kilo of body weight. Those are real effects and I’m not going to pretend otherwise.

Look at what happened downstream, though. In the same review, effects on blood pressure, cholesterol and glycated haemoglobin were typically small and often not statistically significant. The device moved the behaviour it counts. It did not reliably move the physiology you bought it for.

Then there’s the trial nobody in camp one quotes. The IDEA study, published in JAMA in 2016, put 470 young adults with overweight or obesity through six months of intensive diet and exercise counselling, then randomised them: half added a wearable with a web interface, half continued with web-based self-monitoring alone. Over 24 months, the group with the wearable lost less weight. One trial, one population, and I’d caution anyone against building a worldview on it. But it should at minimum retire the assumption that adding measurement to an intervention makes the intervention stronger.

Scans deserve the same honesty. A 2018 BMJ systematic review by Gibson and colleagues pooled 32 studies of brain and body MRI in apparently asymptomatic adults and found potentially serious incidental findings in 3.9% of scans, rising to 12.8% once findings of uncertain seriousness were included. Roughly half of the serious ones were suspected cancers. A separate 2019 systematic review in the Journal of Magnetic Resonance Imaging put the pooled false-positive proportion for whole-body screening MRI at around 16%, with wide uncertainty around it. Both reviews report substantial heterogeneity, so treat the point estimates as ranges rather than facts. The direction is clear enough: in a population without symptoms, a very sensitive test mostly finds things that turn out to be nothing, and each one costs a follow-up, a wait, and a few weeks of a person’s peace.

So camp one gave a generation of careful, well-resourced adults a folder, a monthly subscription, and a low-grade dread. The one thing it didn’t reliably give them was a different body.

The gap is not made of information

Here’s the mechanism, and it’s been measured.

Webb and Sheeran’s 2006 meta-analysis of experimental studies found that a medium-to-large change in people’s intentions (d = 0.66) produced only a small-to-medium change in their behaviour (d = 0.36). Reviews of health behaviour more broadly find intention accounts for somewhere around a fifth of the variance in what people actually do. You can move someone’s intention a long way and move their behaviour barely half as far. Every intervention built on informing, persuading and alerting is fishing in that shallow half.

This is where I have to be careful, because the sloppy version of my argument is that measurement doesn’t work, and that’s not what the evidence says.

Harkin and colleagues, writing in Psychological Bulletin in 2016, meta-analysed 138 experimental studies covering 19,951 people. Prompting someone to monitor their progress towards a goal did promote goal attainment, d = 0.40, and the change in monitoring frequency mediated the effect. Monitoring works.

But read the moderators, because they’re the whole argument. The effect was larger when the outcome was physically recorded, and larger when it was reported or made public. In other words, monitoring works when it’s attached to a goal, written down, and seen by someone else. That is a very specific behaviour. It is not the same activity as a device silently accruing readings you glance at on the train.

Camp one borrowed the credibility of progress monitoring and shipped something structurally different: data collection with no goal fastened to the other end, no accountability, no consequence for the reading being bad. That version has no evidence behind it, and the reason is legible in the mechanism rather than mysterious.

Camp two is treating an illness you don’t have

The backlash arrived on schedule. Put the wrist device in a drawer. Switch off. Rest. Nervous system regulation. The unspoken claim is that the missing input was recovery.

For a subset of people that’s exactly right, and I want to be careful here because burnout is real and the people carrying it are not helped by a founder telling them to try harder.

But the evidence on what a restful break does over time is not ambiguous. A more recent meta-analysis in European Psychologist, by Wendsche and colleagues, pooled 13 studies covering 1,428 employees, with an average holiday length of 11 days. Well-being improved, d = 0.25. Then people went back to work, and after the first post-holiday week the differences from pre-holiday levels were no longer significant, all effect sizes at or below 0.12. The earlier meta-analysis by de Bloom in 2009 found the same shape: a positive effect during the break, a fade after work resumed. Longer holidays didn’t protect the gain.

Rest is necessary. Six days of it is not a stimulus, it’s a pause, and a pause does not change a slope. If your problem is that you already know what to do and don’t do it, a week of switching off returns you to exactly the life you left, slightly better rested, with the same architecture waiting.

Camp two isn’t wrong about rest. It’s answering a question most of camp one’s refugees weren’t asking.

What has actually moved outcomes, in trials, at scale

The Diabetes Prevention Program is the study I’d put in front of anyone who thinks structured doing is the soft option.

3,234 adults at high risk of type 2 diabetes, randomised to placebo, metformin, or an intensive lifestyle programme, published in the New England Journal of Medicine in 2002. Over a mean 2.8 years, the lifestyle arm cut diabetes incidence by 58%, with a confidence interval of 48 to 66. Metformin cut it by 31%. The trial was stopped early because the answer was that clear.

Now look at what the lifestyle arm actually consisted of, because this is the part that gets lost when people summarise it as “diet and exercise”. Every participant had an individual coach. Frequent contact. A structured 16-session core curriculum teaching behavioural self-management. Supervised activity sessions. A maintenance phase with restarts built in for the people who fell off. The published goals were modest, 7% weight loss and 150 minutes a week of brisk walking, and the participants hit them because the programme was engineered around the doing rather than the knowing.

Nobody in that trial needed new information. They needed the thing built.

The smallest version of the same principle has its own literature. Gollwitzer and Sheeran’s meta-analysis of 94 independent tests found that forming an if-then plan, specifying when, where and how, produced a medium-to-large effect on goal attainment, d = 0.65, over holding the goal alone. One sentence, written in advance, moves behaviour more than most of what camp one sells for four figures.

The one number worth your attention

If you’re going to keep something, keep the number that responds to work.

The strongest candidate is cardiorespiratory fitness. Mandsager and colleagues at the Cleveland Clinic followed 122,007 adults who underwent treadmill testing between 1991 and 2014, publishing in JAMA Network Open in 2018. Over 1.1 million person-years, fitness was inversely associated with all-cause mortality, and the association had no observed ceiling. Elite performers, two standard deviations above the age and sex mean, had roughly an 80% lower adjusted mortality risk than the lowest group.

Be precise about what that is. It’s a cohort study of people referred for stress testing, not a randomised trial, and reverse causation is a live concern: unwell people perform worse on treadmills. So this is a strong association, not a proven cause. But it points the same direction as decades of exercise trials, and unlike almost every other marker in your folder, it’s a number you can personally move within weeks by doing something difficult on purpose.

That’s the test I’d apply to everything you currently measure. Does this number change when I do the work? If the answer is no, or no one has ever checked, it’s not a priority, it’s a subscription.

Assessment still matters. It just isn’t a lifestyle. Measure properly, at a point in time, so the work can be built from where you actually are. Then close the file and go and do the work. That’s the shape of the six days we run: assess on day one, build a protocol from those numbers, then implement it with people beside you and measure again before you leave. It’s also why the week doesn’t end at the gate: your team stays on your protocol for the first month home, which is where most retreat gains historically go to die.

I built Atlas Cove as an outsider to this industry, in Portugal, and the reason we run one small cohort a month with ten rooms rather than something scalable is unglamorous. Coaching density is not a luxury feature. In the DPP it was the intervention.

We don’t diagnose, we build.

Our Method

The exit

If you’ve spent five years in camp one, the instinct when it fails is to conclude you need better data. That’s the trap, and the industry is happy to sell you the upgrade.

You don’t have an information problem. You’ve got a folder that proves it. What you have is an energy account running a quiet overdraft, a body that has been described in detail and asked to change nothing, and a set of intentions that have never once been converted into a structure with a coach, a schedule and a witness.

So the question isn’t which panel to run next.

What’s the last thing you measured that changed what you did on a Tuesday? And if the honest answer is nothing, what would it take to spend the next six days building instead of watching?

If this is the argument you’ve been waiting for someone to make, subscribe. The next essays go deeper into what actually moves a baseline: what the training evidence supports, what the recovery science does and doesn’t show, and how to tell a marker worth your attention from one worth ignoring.

Subscribe now

And then there’s the most audacious version. In April, Biohub, the research organisation Zuckerberg founded with the paediatrician Priscilla Chan, committed $500 million over five years to something called the Virtual Biology Initiative: an attempt to build AI models that simulate a human cell accurately enough to predict how disease begins and how to stop it. The stated goal, in their own words, is to help cure or prevent all disease. Not a subset. All of it. Their head of science, Alex Rives, came from an AI lab, not a hospital.

Notice the pattern. The people pouring capital and talent into rebuilding health are, overwhelmingly, not from medicine. They’re from software, hardware, consumer products. And I don’t think that’s a fad. I think it’s a correction.

My brother Tom and I are part of this wave, at a much smaller scale. We spent our careers in tech. Now we run Atlas Cove, an intimate health retreat in Sesimbra, Portugal. This essay is about why outsiders keep showing up in health, what we found when we mapped the retreat market, and how we actually build what we sell, including the part most companies in this space would rather you didn’t ask about.

Medicine solved the wrong half of the problem

Medicine is extraordinary at keeping you alive. It is far less good at keeping you well.

A 2024 cross-sectional study from the Mayo Clinic, published in JAMA Network Open and covering 183 WHO member states, put a number on this. Globally, people now live an average of 9.6 years longer than they live in good health. That gap grew 13% between 2000 and 2019. In the US it’s 12.4 years, the widest in the world. And women carry a gap 2.4 years larger than men, driven by a heavier burden of chronic, noncommunicable disease.

Read that again. Nearly a decade of life, for the average person on earth, spent unwell. Lifespan went up. Healthspan didn’t follow.

This is not because clinicians are failing. The system is built for a different job: identifying and treating disease once it exists. Building health before anything breaks, the daily work of strength, sleep, recovery and capacity, has no department, no billing code, and no owner. That vacancy is exactly the kind of thing tech people are trained to spot. An enormous, universal problem. No incumbent defending the territory. So they walk in.

The retreat market has a top, a bottom, and a gap

So why retreats, of all things?

Because when we mapped the market, we found a strange shape. Wellness tourism is enormous: the Global Wellness Institute measured it at $894 billion in 2024. But almost all of it sits at two extremes.

At the bottom: one-off yoga and pilates weeks, hosted by a single facilitator, built around one modality. Lovely, often. Personal, rarely. Measured, never. You leave relaxed and, within a fortnight, unchanged.

At the top: the medicalised estates. German diagnostic clinics. Ayurvedic hospitals in India. Longevity clinics with full imaging suites, where a week can cost more than a car. Their diagnostics are genuinely excellent. But they inherit medicine’s blind spot. They are superb at telling you what’s wrong and thin on making sure you do anything about it once you’re home.

Between the two sits a gap, and the gap is not really about price. It’s about what the week is for. Almost nobody has built a stay where implementation is the product: your own baseline measured on arrival, a protocol built from those specific numbers, the work done under supervision, the same markers measured again before you leave, and then a year of somebody staying on it with you.

That absence is absurd, because implementation is the entire problem. Most people who could afford either extreme already know what to do. They have the podcast knowledge, sometimes the bloodwork, occasionally a full-body scan. What they don’t have is the doing. A 99% effective health protocol is 0% effective if you never do it.

Why us

I won’t turn this into a memoir, because the point of this essay isn’t my biography. But you should know why two people from tech care this much.

Tom and I have both lived with chronic illness for most of our adult lives. Mine are Hashimoto’s, PCOS and endometriosis. Tom has Gilbert’s syndrome, cold urticaria, migraine, elevated autoimmune reactivity and inflammatory joint pain. None of these kill you. All of them shape every week of your life, and several sit in the badly researched part of medicine, particularly the ones that mostly affect women.

Nobody handed either of us a protocol. We built our own, over years, out of published research, trial and error, and a long series of appointments that ended with a shrug and a prescription. Everything we now know about making a body with a chronic condition strong, functional and full of energy, we had to assemble ourselves. That’s not a complaint about our doctors. It’s a description of what the system is and isn’t for.

Atlas Cove is that assembly, turned into a product. Six days on the Arrábida coast. Your baseline measured on day one: strength, fitness, sleep, recovery, blood pressure, stress, on validated instruments we bring. Overnight, those numbers become your personal protocol. You do the work, coached daily. On day six the same markers are taken again, side by side with day one, so whatever changed is yours to see rather than ours to describe. Then the year after: a month of guided return, included, and a membership that keeps your numbers and your protocol alive between visits. A week that doesn’t survive contact with your real life is theatre.

We don’t diagnose, we build.

Read about our Method

The part nobody in this industry wants to say out loud

Our protocols are generated by AI.

I’ll say it plainly, because most of this sector won’t. The method is ours: the logic of what to do with a given set of numbers, built out of published research and about two decades of managing our own conditions. That method is encoded into a system, and the system turns your day-one assessment into your week. Before anything reaches you, a doctor reviews it remotely against your health history and your assessment results, and approves it. Nothing runs on you that a clinician hasn’t signed off.

We kept quiet about this for a while, and I’ve decided that was the wrong call. The reason we kept quiet is that “AI built your health plan” makes people flinch, and I understand why. But hiding it would mean arguing in public that AI belongs in health while being coy about using it, and that position doesn’t survive a single hard question.

So here’s the case, honestly. The strongest evidence I know is the MASAI trial in Sweden: a randomised controlled trial in over 105,000 women, the first of its kind, with final results published in The Lancet in January. When AI supported radiologists reading mammograms, cancer detection rose 29% with no increase in false positives, screen-reading workload fell 44%, and fewer aggressive cancers surfaced between screening rounds.

What matters most is what that trial did not do. It did not remove the radiologists. It gave them better instruments, and the combination beat either alone. That’s the structure we copied, scaled down to ten rooms. A system that never gets tired at eleven at night and never forgets a contraindication buried on page four of your history, checked by a human who is accountable for the result.

Worth noting too: Biohub, the most ambitious AI project in this whole field, is co-led by a practising paediatrician. The serious version of this always has a clinician in it.

What outsiders actually bring

The advantage of coming from outside medicine isn’t thinking you can do everyone’s job. It’s knowing precisely which job is yours.

Ours is the system, and standing behind it. Tom and I host every cohort ourselves, alongside specialists in the disciplines the week runs on, because a method nobody shows up for is just a document. What the system can’t do, and what we can’t do, goes to people who trained for it.

We also brought an instinct medicine’s structure suppresses: build it, measure it, change it weekly. Ek is doing that with scanners. Biohub is doing it with cells. Function is doing it with bloodwork. We’re doing it with the unglamorous last mile, the distance between knowing and doing. I’ve written before about what that distance costs, in energy and mental load, and I’d argue it’s the most expensive gap in health. Not the missing scan. The missing follow-through.

Now the part where I have to be straight with you, because an essay arguing for measurement that fudges its own would deserve to be ignored.

We have not proven this yet. September is our first full cohort. The week is designed so that the same markers are taken on day one and day six, on the same instruments, and every guest sees their own two columns. But I have no outcome data to show you, because we haven’t collected it at any meaningful scale. The figures published on our own method page are estimates drawn from research on comparable programmes, and they’re labelled as exactly that. I’d rather say this plainly in an essay than have someone discover it and assume we were hoping they wouldn’t look.

What we do have is a demand signal, and it’s the part that convinced me the gap is real. More than a thousand people are on our waitlist before a single full-price cohort has run. That’s not proof the method works. It’s evidence that a lot of people have already tried the bottom of this market and the top of it, and are looking for something neither one sells.

So the honest position is this. The category argument, I’ll defend now. The outcome claim, I’ll have data on by next summer, from cohorts running monthly from November. Ask me then, and hold me to it.

Ten rooms is a strange place to start an argument with an industry, so let me be clear about what this is. Sesimbra is one site, at a partner estate, and it’s deliberately the smallest version of the thing. The plan is a network of our own-operated sites across Europe. We’re starting here because a method that hasn’t survived contact with real cohorts shouldn’t be copied into five buildings, and because the discipline of ten rooms and one cohort a month is what makes the measurement honest. Scale after proof, not before it.

Zuckerberg’s timeline for curing all disease runs to the end of the century, which is a polite way of saying nobody alive will audit it. Ours is six days and then twelve months, which is a much harder promise to make, because you can check it.

What I’m confident about is the direction. Health is being rebuilt by people who were failed by the current version of it, using tools the current version was too slow to adopt, with clinicians as the check rather than the gate.

So here’s what I’d ask you. Where does your own gap sit: in what you know, or in what you do? And if you’re honest with yourself, when did knowing more last change anything?

If this argument interests you, subscribe. The next essays go deeper into what actually moves a baseline, and what a year of holding it looks like.

Subscribe now

Strength work was the first thing I cut. Every time.

Through the years I was building toward long-distance triathlon, the gym session was the one with no race consequence I could point at, so whenever a week got tight it went. The endurance numbers stayed fine. My VO2 max read 51.6 that October in Cascais, down on where it had sat in April, and the race still went the way I wanted it to. What I was quietly giving up across those years was the ability to produce force fast, and I never caught it, because an endurance training log has no column for that.

If you train six or seven hours in a week and work eats the rest, you have to decide what those hours are for. Strength usually loses that argument. It looks like the vanity option, the thing you do for how you look rather than for how long you last, while the endurance work feels like it’s buying something a doctor would recognise.

The physiology disagrees, and the margin isn’t small.

So, the narrow claim: strength training has the widest set of downstream effects per hour you put in. Endurance and mobility work aren’t wasted. They do things lifting can’t touch.

This is an educational and strategic perspective, not personal medical advice.

The fibres you stopped asking for

Ageing doesn’t thin muscle evenly. Everything else here follows from that.

Look at what the biopsies showed. Nilwik and colleagues compared leg muscle cross-sectional area in young and elderly men, and the difference tracked almost entirely to the size of the type II fibres, with the number of fibres largely intact (Nilwik et al., 2013). Sarcopenia, the age-related loss of muscle mass and function, is selective. One compartment deflates while the rest holds.

Type II fibres are the fast, high-force ones, and they sit at the top of the recruitment order, which means your nervous system reaches for them last and only when a movement demands real force or real speed. Sitting comes nowhere near that threshold. Neither does walking, and neither, and this is the part people resist hardest, does a steady aerobic hour. A compartment the body never gets asked for is a compartment it has no reason to keep.

What goes, then, is speed of force. You meet it when a kerb arrives sooner than you thought and you catch yourself, or when the low chair suddenly wants a hand on the armrest that it didn’t want three years ago. Nothing athletic about either. It’s the ordinary business of a day, and the reserve is there or it isn’t.

Underneath that sits the version of muscle that matters most and gets discussed least: the reserve your body draws down during illness, injury and the long immobilities that arrive later.

Subscribe now

Some fibres lose their address

One level up from the tissue, in the wiring, a second loss runs at the same time.

Every muscle fibre answers to a motor neuron. One neuron plus all the fibres it commands makes a motor unit, the smallest thing your nervous system can switch on. Motor neurons die off with age, and when one goes, the fibres it owned are orphaned.

The body does have a repair strategy, in that surviving neighbours sprout and reinnervate the abandoned fibres, so the remaining motor units grow bigger and the muscle holds together with fewer controllers in charge. That strategy has a ceiling. Intramuscular electromyography on the tibialis anterior and vastus lateralis, run across young men, non-sarcopenic older men, pre-sarcopenic men and sarcopenic men, found the expansion happening as you would expect in the men who weren’t sarcopenic. In the sarcopenic men it hadn’t happened, and their motor units measured smaller than the ones carried by the older men without sarcopenia (Piasecki et al., 2018).

Handle that one carefully. It’s cross-sectional, every subject was male, and it locates the failure without showing that lifting reverses it.

What it does clarify is worth carrying. Strength is partly how much muscle you own and partly how much of it you can still reach, and heavy, deliberate, well-recovered efforts are what exercise the reaching. Which is why the first few weeks of sensible training change what you can lift long before they change what you weigh.

Why the same meal does less at fifty

Feed older muscle the same protein and you don’t get the same result, and the reason is more specific than a story about slowing metabolism.

Forty-four healthy young and older men of similar build were given varying doses of essential amino acids, and muscle protein synthesis was measured directly. Basal rates came out indistinguishable between the groups, which kills the intuitive idea that old muscle simply idles. The deficit sat in the response. Older muscle proved both less sensitive and less responsive to the same amino acid signal, and the machinery inside the fibre that senses amino acids and turns them into new protein, mTOR, p70 S6 kinase, eIF4BP-1 and eIF2B, showed lower expression and weaker activation after feeding, alongside a sharp rise in the inflammation-associated transcription factor NF-κB (Cuthbertson et al., 2005).

Call it anabolic resistance. The signal has to be louder before older muscle registers it at all.

The direct reading here is about protein. Sensitivity and responsiveness both fall, so the dose per meal has to be bigger to clear the same bar. Or rather, that’s the reading a dose-response trial is actually built to support, which is worth saying because the inference people usually draw from this paper is a different one.

That other inference travels further than Cuthbertson went. Anabolic resistance does argue for load that genuinely challenges the tissue over the token resistance a general fitness class hands you, because a stimulus below threshold buys you very little of what a stimulus above it buys. But this trial dosed amino acids, not barbells. Take it as a reasonable extension and not as a result.

The organ nobody counts as an organ

Muscle got reclassified over the past two decades and most people were never told.

Pedersen and Febbraio set out the case that skeletal muscle produces, expresses and releases several hundred peptides, the myokines, which act locally on the muscle and also travel in the blood to reach fat, liver, pancreas, bone and brain (Pedersen and Febbraio, 2012). Muscle sits in conversation with nearly every system that fails slowly, including the one between your ears.

On that last one I’ll go further than the evidence takes me, and I want to be visible about it rather than sly. I think strength and endurance work are the base for the mind as much as the body, and that we file as psychological a fair amount of what a badly conditioned body is producing. No paper behind that here. The literature I’d need is associative, the causal arrow runs in both directions, and I’d be doing precisely what I complain about if I dressed a belief up as a finding. So take it as a working position. It’s also the thing I’d most like someone to prove me wrong about, properly, with a trial.

The second function has a hard number on it, and past 40 it’s the one that should get your attention. Under insulin-stimulated conditions, roughly 80 percent of the glucose your body disposes of goes into skeletal muscle, which is why DeFronzo and Tripathy describe skeletal muscle insulin resistance as the initiating defect in type 2 diabetes, sitting there for decades before blood sugar rises far enough for anyone to notice (DeFronzo and Tripathy, 2009).

Put those together and the metabolic case turns concrete. The muscle you carry sets much of the storage capacity available for every meal you eat, which is the unglamorous reason strength work turns up in blood markers that look unrelated to lifting.

Unglamorous is the whole problem. A large, well-evidenced lever that moves at the speed of months will never compete for attention with the toxin of the quarter, and there is always a toxin of the quarter, because that is what the market pays for. Nobody ever built an audience on carry more muscle and go to bed earlier.

Bone listens for one specific message

Bone isn’t maintained by movement in general. That’s why so much well-meant advice about bone density disappoints.

It answers to mechanotransduction, where mechanical strain gets converted into a cellular signal, and the strain has to be high in magnitude and delivered at a high rate before the signal counts as worth building for. A narrow specification.

LIFTMOR tested that specification directly instead of assuming it. A hundred and one postmenopausal women with low bone mass were randomised either to eight months of supervised high-intensity resistance and impact training, twice weekly for thirty minutes, five sets of five reps above 85 percent of one repetition maximum, or to a low-intensity home programme. Lumbar spine bone mineral density rose 2.9 percent in the training group and fell 1.2 percent in the controls. Femoral neck density and cortical thickness both favoured training, every functional measure improved, and across the whole trial there was one adverse event: a minor back spasm (Watson et al., 2018).

Those women aren’t the typical reader here, and heavy loading in osteoporosis was held to be too dangerous to recommend for years. What transfers is the mechanism. Bone answers to load of a particular character, and outside of impact work the only reliable source of it is resistance training, which is why lifting speaks to the skeleton in a language it acts on.

The number that outperformed blood pressure

Grip strength is an odd thing to care about until you see what it predicts. In the PURE study, every five kilograms of grip lost came with a hazard ratio of 1.16 for all-cause mortality, and grip predicted all-cause and cardiovascular mortality more strongly than systolic blood pressure did (Leong et al., 2015).

Now the half that gets dropped every time this finding resurfaces. It’s an association. PURE never tested whether raising anyone’s grip lowers their risk, which would be a different study, and grip is almost certainly standing in for a lot at once: nutritional status, decades of accumulated activity, illness that has started without being diagnosed yet. Someone can be weak because they are becoming unwell. A cohort study can’t tell you which way the arrow points.

Fine. It still justifies a change of procedure. Treat strength as a vital sign, measure it next to blood pressure and lipids, track it across years, and take it seriously when it moves the wrong way.

Subscribe now

What the work costs, and what it costs elsewhere

Two properly heavy sessions a week come to ninety minutes or two hours once you count the warm-up and the rest between sets honestly. Those hours come out of a week that already holds the endurance training you probably care about more.

Wilson and colleagues pooled 21 studies and 422 effect sizes to put a number on the interference. Hypertrophy showed an effect size of 1.23 for strength training alone against 0.85 when the same training ran concurrently with endurance work, power showed a comparable gap, and both the frequency and the duration of the endurance work correlated negatively with gains in size, strength and power. Modality mattered more than most people expect. Running produced significant decrements in hypertrophy and strength; cycling didn’t (Wilson et al., 2012).

The training literature tends to miss a second cost. Part one described the autonomic account that training, work, sleep debt and emotional load all draw on, and heavy lifting is a withdrawal from that same account. Stack two hard sessions onto an unchanged week during a brutal month at work, and the tissue will not answer the way it answers in a quiet one.

Then the pace of the thing. Strength adapts across months, a session that feels like a breakthrough is rare, and the reward for a good year is a slightly heavier bar plus a change in body composition you only really see in photographs taken far apart. I put on about 2 kg across a twelve-week build. Perfectly good result. Looks like nothing.

None of which argues the other work should go. Aerobic training builds the parasympathetic capacity part one was concerned with, in a way lifting doesn’t, and mobility work has its own narrower and frequently misunderstood job, covered separately in the Ironman series. Strength is usually the missing piece. Crowding out the rest would be its own mistake.

The twelve-week question

One question separates strength training from the thing that looks like it from outside. Can you name the load you used on your main lifts twelve weeks ago, and is today’s higher?

If the first half stumps you, the record that progress depends on was never kept, and progress nobody writes down tends not to happen. If you can answer it and the number hasn’t moved, you’ve maintained a habit and not delivered a stimulus. Your type II fibres, your motor units and your skeleton have all been getting a message they finished adapting to a while back.

Second check. Half a minute, no equipment. Stand up from a low chair five times without using your hands, and notice whether the fifth is clearly worse than the first. That tells you where your reserve sits today. It says nothing about your character, and it responds well to being acted on.

Where this leads

This is close to why an Atlas Cove week puts strength work next to the recovery data instead of in its own silo, so the load someone takes on gets decided from their actual state and never from a template. The thing I want surviving the flight home is small: load deliberately, then check whether the loading landed.

The Atlas Cove Method

Muscle is easy to read as decoration and hard to read as an organ. That’s why it goes first when the calendar fills. Considered as an organ it’s your largest site of glucose disposal, a secretory tissue talking to most of the systems that go wrong slowly, and the only mechanical signal your skeleton reliably answers.

The work is modest in hours and dull in feel. Probably the whole reason it keeps getting skipped. Start anyway.

This is an educational and strategic perspective, not personal medical advice.
The views are the author's own and not statements by Atlas Cove Lda.

Sources

1. Nilwik, R., Snijders, T., Leenders, M., Groen, B. B. L., van Kranenburg, J., Verdijk, L. B., & van Loon, L. J. C. (2013). The decline in skeletal muscle mass with aging is mainly attributed to a reduction in type II muscle fiber size. Experimental Gerontology, 48(5), 492–498. DOI: 10.1016/j.exger.2013.02.012

2. Piasecki, M., Ireland, A., Piasecki, J., Stashuk, D. W., Swiecicka, A., Rutter, M. K., Jones, D. A., & McPhee, J. S. (2018). Failure to expand the motor unit size to compensate for declining motor unit numbers distinguishes sarcopenic from non-sarcopenic older men. The Journal of Physiology, 596(9), 1627–1637. DOI: 10.1113/JP275520

3. Cuthbertson, D., Smith, K., Babraj, J., Leese, G., Waddell, T., Atherton, P., Wackerhage, H., Taylor, P. M., & Rennie, M. J. (2005). Anabolic signaling deficits underlie amino acid resistance of wasting, aging muscle. The FASEB Journal, 19(3), 422–424. DOI: 10.1096/fj.04-2640fje

4. Pedersen, B. K., & Febbraio, M. A. (2012). Muscles, exercise and obesity: skeletal muscle as a secretory organ. Nature Reviews Endocrinology, 8(8), 457–465. DOI: 10.1038/nrendo.2012.49

5. DeFronzo, R. A., & Tripathy, D. (2009). Skeletal muscle insulin resistance is the primary defect in type 2 diabetes. Diabetes Care, 32(Suppl. 2), S157–S163. DOI: 10.2337/dc09-S302

6. Watson, S. L., Weeks, B. K., Weis, L. J., Harding, A. T., Horan, S. A., & Beck, B. R. (2018). High-intensity resistance and impact training improves bone mineral density and physical function in postmenopausal women with osteopenia and osteoporosis: the LIFTMOR randomized controlled trial. Journal of Bone and Mineral Research, 33(2), 211–220. DOI: 10.1002/jbmr.3284

7. Leong, D. P., Teo, K. K., Rangarajan, S., Lopez-Jaramillo, P., Avezum, A., Orlandini, A., et al. (2015). Prognostic value of grip strength: findings from the Prospective Urban Rural Epidemiology (PURE) study. The Lancet, 386(9990), 266–273. DOI: 10.1016/S0140-6736(14)62000-6

8. Wilson, J. M., Marin, P. J., Rhea, M. R., Wilson, S. M. C., Loenneke, J. P., & Anderson, J. C. (2012). Concurrent training: a meta-analysis examining interference of aerobic and resistance exercises. Journal of Strength and Conditioning Research, 26(8), 2293–2307. DOI: 10.1519/JSC.0b013e31823a3e2d

“You’re going to tear yourself apart doing it like that.”

A man I used to run with on Sundays said that to me in a car park, and he meant it kindly. He’d watched me finish, sit down on the kerb, pull my shoes off and get straight into the car, and at no point had I stretched a single thing. He’d been running since before I was born. No slogan in it either. He was telling me what nearly everyone in the sport believes, with about thirty years of coaching orthodoxy standing behind him.

He was wrong on the specific claim. That took me a while to say out loud, because he had the majority position and the good intentions, and I had a hunch and a vague memory of a review paper.

So here is that claim, kept small. Static stretching earns its keep for range and for coming down after a hard day. What it has never really shown is that it keeps healthy athletes off the injury list, and that is the exact job most people adopted it for.

That gap matters most for whoever has the fewest spare hours.

This is an educational and strategic perspective, not personal medical advice. The views are the author’s own and not statements by Atlas Cove Lda.

Subscribe now

Thirty years of orthodoxy, and a thin file behind it

The story is easy to picture, and that is why it stuck: a short muscle tugs a joint out of line, tissue further down the chain absorbs force in a direction it was never built for, and lengthening the muscle puts the geometry back where it belongs. Joint angles really do change when you stretch. That part is real. What doesn’t follow is the bit the story promises, once you stop measuring degrees and start counting injuries.

Thacker and colleagues went through the literature on stretching and injury risk and found no reliable drop in injury rates from stretching before or after exercise, in every group they could look at (Thacker et al., 2004). A broader review a decade later arrived from a different angle. Static stretching genuinely does buy a short-term increase in available range, that increase doesn’t turn into fewer injuries in healthy active people, and a long passive hold immediately before a hard effort can temporarily cost you force and power (Behm et al., 2016).

Read that carefully, because the sloppy version of this argument has done real damage to how it lands. Nothing in there says stretching hurts you. Nothing says flexibility is pointless, and nothing says the person who enjoys ten quiet minutes on the floor is wasting their evening. One narrow claim comes apart, which is the idea that holding a muscle long keeps an otherwise healthy athlete intact, and that’s a far smaller statement than the one people usually hear, because nearly all the useful judgement here lives in the space between the two.

Where a stiff hip actually shows up

Some cases need no argument at all. The first is a sport-specific shortfall in range, meaning a movement your sport genuinely demands and your body can’t currently produce. A runner whose ankle won’t dorsiflex. A swimmer whose shoulder won’t rotate through the arc the stroke asks for. That’s a measurable gap sitting directly in the path of something they’ll repeat a few thousand times before lunch, and working on that one joint removes a real limit rather than chasing general suppleness and hoping it helps somewhere.

The second is compensation, where one stiff joint quietly gets paid for by a neighbour badly designed to cover the bill, so that limited hip extension while running tends to get settled by the lumbar spine and the pain then turns up some distance away from whatever caused it. Giving the hip its range back there has nothing to do with being bendy. It switches off a repeated overload upstream of something fragile. And if you sit for nine hours and then run for one, that description fits you far better than it fits whoever is racing next weekend.

There’s also a distinction that gets lost constantly, between a dynamic warm-up and a long passive hold. A dynamic warm-up moves a joint through its working range at rising intensity, in patterns that resemble the session about to happen, and it warms the tissue and wakes the nervous system up for what’s coming. It belongs in anyone’s week. Sitting in a stretched position for sixty seconds beforehand is a different animal, and the passive holding is the part the literature treats with caution. Warming up was never in question. (The foam roller sits somewhere in between and I genuinely don’t have a strong view on it.)

None of which is an argument against mobility work as a category. The argument is narrower and it’s about several undirected hours a week that fix no shortfall you can name and build no capacity, while you go on counting them as insurance.

The Achilles that grumbles for the first ten minutes

To see why loading beats lengthening, be exact about how these injuries turn up. Almost all of them are failures of load tolerance. A tissue got asked, over and over, to absorb more mechanical stress than its current structure could handle, until the accumulated demand found the weakest link in the chain. Running is a long sequence of ground-reaction forces, each footstrike a brief instruction to absorb energy and give it back, and the stretch-shortening cycle that makes this cheap depends on tendon and muscle working as a spring, loaded and released several thousand times an hour. Downhill running raises the eccentric share of that, meaning the muscle produces force while it lengthens under load, and that is exactly the pattern under which tendons accumulate stress.

Seen that way, the complaints you hear in every running club read differently. The Achilles that grumbles for the first ten minutes. The patellar tendon that flares after a hard block. Neither is really a stiffness story. Both are a body handed more than it was ready to carry.

If that reading holds, the thing that answers the problem raises what the tissue can take rather than how long it is, and here the evidence separates most cleanly. Lauersen and colleagues pooled randomised trials of exercise programmes aimed at preventing injury and found the strength-based ones cut injury rates substantially, while the stretching ones managed nothing comparable (Lauersen et al., 2014). That split is hard to explain unless load tolerance is the variable doing the work, and it holds even where people stuck to both equally well.

The mechanism is reasonably well described. Load a tendon progressively and it changes as a structure. It gets stiffer, and it absorbs and returns more force before it starts accruing damage, which is a genuine material change rather than a passing sense of looseness (Bohm et al., 2015). Eccentric work in particular, the controlled lowering under tension that heavy squats and Romanian deadlifts hand you by their nature, is what loading-based rehab for Achilles and patellar tendinopathy is built on (Malliaras et al., 2013). Which is to say it’s already the standard care for precisely the problems endurance athletes most want to avoid.

Stretching changes how compliant a tissue feels this afternoon. Loading changes what it can carry six months out.

Subscribe now

Carrying more muscle up the hill

If strength work only kept you intact it would already earn a slot in a crowded week, and it also tends to make you faster, which tips the trade a long way further.

The standing objection is straightforward enough. Muscle is mass, and power to weight is what endurance performance runs on. I believed it for years. Beattie and colleagues pooled the strength training work in endurance athletes and found close to the opposite, with resistance training improving running economy and the related measures while maximal oxygen uptake and body composition stayed essentially where they started (Beattie et al., 2014).

Running economy is doing most of the work in that result. It’s worth understanding rather than quoting. Economy is the oxygen cost of holding a given pace. It improves for two connected reasons: a stiffer tendon gives back a bigger share of the elastic energy stored at each footstrike instead of losing it, and a stronger muscle works at a smaller fraction of its maximum to make the same force. A tighter meta-analysis in highly trained runners found a large benefit to economy from lifting and plyometric work done a few times a week over a couple of months (Balsalobre-Fernández et al., 2016).

In the last hour of a long race, better economy is fuel and time handed back for nothing, at the one moment neither is available by any other route. If you never race, the same change shows up as an ordinary Tuesday that leaves more in the tank by six.

The first thing off the board

For nearly everyone reading this, the binding limit is neither knowledge nor motivation. It’s hours. Training for a long event already eats most of the discretionary time an adult with a job and a family can find, two or three lifting sessions add to that, and once several more hours of undirected mobility work go on top, something has to give.

I know exactly which thing gives, because for years it was me. Through most of the decade I spent working up to long-distance racing, the gym was the first item off the board whenever a week went sideways. Nothing in my training log went red when I skipped it, so it kept getting skipped, and I filed the whole pattern as a scheduling problem when it was a decision.

What changed my mind was going and reading the thing properly, which is a duller story than a blown Achilles and happens to be the true one. For years I had been disciplined about the hours with the thinnest evidence behind them and casual with the hour carrying the most, and I had never once sat down and worked out which was which, which is a fairly stupid way to spend a decade of Sunday mornings. That one still irritates me.

What came after is the part I would defend. Once the loading stopped being negotiable, the two Ironman builds that followed looked different: three sessions a week at 8 to 12 reps, held in place through the heaviest endurance blocks, zero static stretching, and no overuse injury at either end. Twelve years of consistent training by then, two Ironman finishes and six marathons.

Now the part where I have to be careful with my own story. That’s an n of 1. It proves nothing. It happens to point the same way the pooled literature points, which is a coincidence I enjoy and would not want anyone to lean on, and if I let a personal anecdote stand in for evidence here I’d be making the exact mistake the stretching habit is made of.

Why the pleasant option wins the last free hour is entirely human, by the way. A long hold drops the felt tightness straight away and leaves you calmer than you were, while heavy loading leaves you tired and competes for the same recovery your endurance work is already spending. Low risk feels reassuring right up to the moment you notice it’s also low return for the one job you took it on to do.

One percent a year, and where you notice it

Everything above is about a single goal, which is staying uninjured while training a lot on very little time, and it shouldn’t be allowed to grow into a verdict on mobility in general. Range of motion has value with nothing to do with running injuries at all.

It declines across adult life, unevenly between joints and reliably enough to measure. Araújo went through around 6,000 Flexitest results and found flexibility falling in the region of 1 percent a year, with the hips and ankles among the first to stiffen (Araújo, 2013). None of that turns up in a race result. It turns up in whether you can sit on the floor with your kids without your back filing a complaint, or get off the ground unassisted at seventy. Those are closer to the point of the whole exercise than any finish time I’ve ever recorded. So a working range is a legitimate goal on its own terms, kept for the quality of the next few decades.

The recovery argument deserves the same honesty. A gentle stretch does lower the immediate feeling of tightness, and there’s a real shift toward parasympathetic activity available in a slow, well-breathed few minutes before bed, which sits comfortably beside the case for taking rest as seriously as output. It just shouldn’t be defended as injury prevention. Or rather, it shouldn’t need defending at all, because helping you come down off a day is a good enough reason on its own.

Strength and mobility answer different questions. Both are worth having. Loading raises how much stress a tissue tolerates, mobility keeps the range the joint can move through, and a body that intends to train for another twenty years wants a supply of both. The two only ever conflict when the hours run out.

How to split the hours you’ve got

The division follows without much strain. The dynamic warm-up stays, because preparing tissue for the work about to be done is straightforwardly sensible. Targeted mobility work stays as well, for a sport-specific shortfall or a compensation pattern you can name out loud. Keep some undirected mobility work too, for range and for the wind-down, as long as you’re honest with yourself that this is where its value sits. Then the strength training, two or three sessions a week of compound movements with the load creeping up over the months, and that one isn’t available for trading.

It carries the strongest evidence both for staying intact and for getting faster, and it’s reliably the first thing to vanish when the calendar tightens. I know. I did it for years.

An Atlas Cove week gets built on the same division of hours, so what a particular body is short of gets established before the scarce time gets assigned. A habit can’t do that reading for you.

The man in that car park wasn’t wrong about much. He was wrong about one thing, and it was the thing he was surest of. Stretching isn’t worthless, and if you finish a long ride with fifteen quiet minutes on the floor you’re doing yourself no harm at all. The correction is smaller and more useful than that. Mobility work has good justifications, and staying uninjured isn’t among them, so when time is genuinely short the hour that most keeps you from breaking down is the hour you spend loading the tissue.

This is an educational and strategic perspective, not personal medical advice. The views are the author’s own and not statements by Atlas Cove Lda.

Sources

1. Thacker, S. B., Gilchrist, J., Stroup, D. F., & Kimsey, C. D. (2004). The impact of stretching on sports injury risk: a systematic review of the literature. Medicine & Science in Sports & Exercise, 36(3), 371-378. DOI: 10.1249/01.MSS.0000117134.83018.F7

2. Behm, D. G., Blazevich, A. J., Kay, A. D., & McHugh, M. (2016). Acute effects of muscle stretching on physical performance, range of motion, and injury incidence in healthy active individuals: a systematic review. Applied Physiology, Nutrition, and Metabolism, 41(1), 1-11. DOI: 10.1139/apnm-2015-0235

3. Lauersen, J. B., Bertelsen, D. M., & Andersen, L. B. (2014). The effectiveness of exercise interventions to prevent sports injuries: a systematic review and meta-analysis of randomised controlled trials. British Journal of Sports Medicine, 48(11), 871-877. DOI: 10.1136/bjsports-2013-092538

4. Bohm, S., Mersmann, F., & Arampatzis, A. (2015). Human tendon adaptation in response to mechanical loading: a systematic review and meta-analysis of exercise intervention studies on healthy adults. Sports Medicine - Open, 1(1), 7. DOI: 10.1186/s40798-015-0009-9

5. Malliaras, P., Barton, C. J., Reeves, N. D., & Langberg, H. (2013). Achilles and patellar tendinopathy loading programmes: a systematic review comparing clinical outcomes and identifying potential mechanisms for effectiveness. Sports Medicine, 43(4), 267-286. DOI: 10.1007/s40279-013-0019-z

6. Beattie, K., Kenny, I. C., Lyons, M., & Carson, B. P. (2014). The effect of strength training on performance in endurance athletes. Sports Medicine, 44(6), 845-865. DOI: 10.1007/s40279-014-0157-y

7. Balsalobre-Fernández, C., Santos-Concejero, J., & Grivas, G. V. (2016). Effects of strength training on running economy in highly trained runners: a systematic review with meta-analysis of controlled trials. Journal of Strength and Conditioning Research, 30(8), 2361-2368. DOI: 10.1519/JSC.0000000000001316

8. Araújo, C. G. S. (2013). Age-related mobility loss is joint-specific: an analysis from 6,000 Flexitest results. GeroScience (AGE), 35(6), 2399-2407. DOI: 10.1007/s11357-013-9525-z

Take an ordinary Tuesday out of last summer. One of mine, in Lisbon.

Nine in the morning, a call about money. The useful part of it lasted maybe four minutes and I carried the other fifty around for the rest of the day. Then the middle of the day, which is where the small decisions live, six or seven of them, none of them big on its own, every one with a clock attached. My sister and I were building the company and there was nobody to hand any of it to. By six the inbox had gone quiet, which felt like permission, and the session was on the plan anyway. Ninety minutes of threshold work. The kind that’s supposed to hurt in a boring, predictable way.

It didn’t hurt in a boring, predictable way. Someone had moved the intervals up a gear while I wasn’t looking, or that’s how it felt, and I finished it mildly annoyed at my own legs.

Then a short night, because the day had run long and the alarm doesn’t negotiate.

Nothing in that Tuesday was a mistake. The call had to happen, the decisions were mine to make, and the session had been written down weeks earlier by a version of me who thought it was a sensible idea at the time. What that day actually was is four withdrawals from one account. Your body doesn’t keep separate books for work and for training, and I spent about a decade assuming it did.

The physiology behind that is less complicated than it sounds. The system that runs your heart rate and your digestion answers to a single pool of capacity, a difficult conversation draws that pool down in much the way a difficult interval does, and once you can see the balance, most of the arguments you have with yourself about discipline stop being worth having.

This is an educational and strategic perspective, not personal medical advice. The views are the author’s own and not statements by Atlas Cove Lda.

The body keeps one set of books

The autonomic nervous system runs everything you never think about: the heartbeat, the gut, the immune traffic, the repair that gets going once you stop moving. It has two branches and they pull in opposite directions on purpose. The sympathetic branch puts you into output. Heart rate climbs, blood gets pushed toward the muscle that is about to work, and the hormones that make hard effort possible go out into the blood. The parasympathetic branch does the other half of the job, so the heart slows, the gut gets its blood supply back, tissue repairs itself and glycogen refills.

Here’s the part that matters if you train seriously and also work for a living. Both branches answer to overlapping central stress-response networks, so a threshold interval, a difficult negotiation, a short night and a low-grade cold do not get booked to four separate ledgers (Chrousos, 2009). They pull on one shared reserve. The system never asks where a demand came from, only how big it was and how much came back afterwards.

That’s where the intuitive model of stress falls over. Most people I know manage their training load with real care, down to the week and the session, and then treat the other hundred and ten hours as though those were free. The deadline and the broken night are withdrawals in exactly the sense the interval was. I did this for years and I’ll come back to how obvious it looks now.

Subscribe now

The bill arrives as nothing dramatic

In short bursts none of this is a problem, and that’s worth saying before anything else. A sympathetic surge in response to a hard effort or an acute mess is precisely what the machinery exists for, and given a decent night afterwards it resets at no cost. The trouble starts when the demands come often and the recovery between them gets thin.

The idea that describes the long-run cost of that imbalance is allostatic load, which came out of Bruce McEwen’s work on the physiology of stress. It names the wear that the regulating systems take on when they get called repeatedly without enough recovery between calls (McEwen, 1998). Wear is the right word for it. Nothing breaks.

What it looks like in an ordinary week is undramatic to the point of being easy to miss. A resting heart rate a few beats above your own normal. Sleep that no longer puts you back where it used to put you. Body composition that won’t move on a routine that hasn’t changed, and attention that thins out around three in the afternoon. None of that is a verdict on your character and none of it is a diagnosis of anything at all. It’s a system spending less because less is what it has this month, which is a sensible thing for a system to do.

The daily version reads like that Tuesday, repeated. The morning meeting sets off an early hormonal response, the decisions under time pressure keep it up, and the evening session adds a physical demand on top of a cognitive one that never cleared. Then the short night fails to return what the day took out. Each of those is manageable on its own. They arrive in that order for weeks at a time, and then they add up to a load that no single night clears, so the shortfall carries into the week after.

I’ve had a long stretch with this account overdrawn, and I’ve written about that elsewhere, so one line is enough here.

The side of the account almost nobody trains

The restoring side isn’t fixed, which is the encouraging half of all this. It answers to a stimulus much the way your aerobic engine does, so you can build it deliberately instead of waiting around to feel better.

Easy aerobic work is associated with better cardiac parasympathetic regulation (Buchheit and Gindre, 2006). The slow hours build the branch that puts things back, and most people are already doing those hours for a completely different reason. Slow breathing, somewhere near six breaths a minute, moves autonomic balance toward the parasympathetic side within minutes, and a review of work in healthy adults suggests the benefit accumulates with regular practice (Russo, Santarelli and O’Rourke, 2017). Deep sleep is when parasympathetic traffic runs highest. Which is why guarding your bedtime and your evening light pays back more than almost anything you could swallow instead.

Now the limits, because I don’t want that read as stronger than it is. Buchheit and Gindre is an association in people who already train, and the breathing work is a review of small studies in healthy adults, so what I’d claim out of both is the direction. The dose is guesswork. There’s also no trial I know of that takes recovery capacity as its endpoint and runs for a year. I’d like to see one.

Here’s the accidental part, and it took me embarrassingly long to notice. Most of my week was easy aerobic work, done entirely for the engine. Or so I thought. Those same hours were the only thing I was doing for the other side of the account, and I was doing it by accident, which is roughly the level of intent I brought to recovery back then.

The pattern most people fall into is to train the spending side to exhaustion, through hard sessions and sheer volume, to leave the restoring side completely untrained, and then to read the slower recovery that follows as an inconvenience rather than as information. Building the restoring side is one of the few moves available that makes the budget bigger, instead of rationing what’s left of it.

Two kinds of number, and only one answers today’s question

There’s a difference between measuring what your body has produced and measuring what it can currently carry, and the gap between those two matters more than it first looks.

Outcome measures describe where you’ve been over months. Cholesterol, VO2 max, body composition, whatever your last blood panel said. All worth having, and I get mine done. They tell you very little about what you can absorb today. State measures get closer to that question: your morning heart rate, the trend in your heart rate variability, what the last two nights were actually like, how heavy the first ten minutes of easy work feel. Heart rate variability has been studied specifically as an index of how strongly the brain’s regulating regions are shaping autonomic output, which is part of why it tracks with stress and health rather than with fitness considered on its own (Thayer et al., 2012).

The version you can run this week is small. Before the session you’ve already decided on, take four readings together rather than one at a time: your resting heart rate this morning held against your own normal and nobody else’s, whatever recovery trend the watch gives you read across a week instead of as today’s verdict, what the last two nights were genuinely like including the one you’d rather round up, and how heavy the first ten minutes of easy work feel. That last one needs no device at all. It’s the one I’ve come to trust most.

One of the four pointing down is noise. Three or four pointing the same way is a week you won’t win an argument with, and the useful move is to put something back rather than take more out, whatever the plan says. When they’re all fine, spend freely. That’s the whole lens.

Precision is not what any of this offers. A single morning number tells you close to nothing, consumer devices disagree with each other and sometimes with themselves a day later, and the signal lives in the direction across a week. Read the state, then decide. Collecting a good number is a different hobby, and a surprisingly absorbing one.

My spreadsheet had one column in it

I ran a version of that for years, carefully, on exactly one side of the account. Managing a training load is genuinely something I know how to do, and I’ve taken that part seriously for a long time. Every week of the last build was planned properly, the hours and the split between easy and hard. I knew what would come off the plan when a week went wrong, before the week went wrong.

Then look at what else was in those months. We were building a company in a country whose paperwork I was still learning, with money decisions in most weeks, and the occasional four in the morning where I solved something in my head and then had to solve it again properly at nine.

None of that appeared anywhere in the plan. The spreadsheet had a column for training hours and no column for anything else, and I treated everything outside it as free. I’d have laughed at anyone modelling a system that way at work.

The physiology says the two withdrawals are the same withdrawal. The hypothalamic-pituitary-adrenal axis, which runs the hormonal arm of the stress response, doesn’t keep one programme for a hard session and another for a hard week (Chrousos, 2009). Same pathway. Same reserve.

There’s evidence for the interference beyond the mechanism, and I want to be straight about what kind of evidence it is. A systematic review of stress and exercise reported that higher psychological stress goes with poorer recovery, smaller training adaptations and more injuries at a comparable training load (Stults-Kolehmainen and Sinha, 2014). Those things travel together, which is all a design of that kind can establish, and it deserves saying plainly rather than slipping past. Nobody randomises a person into a stressful year. Anyone under strain also sleeps worse and eats differently, so isolating the strain itself is not something these designs can do. What it supports is narrower and still useful: a plan that looks perfectly reasonable on paper can be the thing that tips the total over, because the plan never knew about the rest of the week.

While I’m here, an opinion, flat. Every few months there’s a fresh exposure to be frightened of, the pan, the receipt paper, the bottle, whatever’s in the shampoo. Most of it is noise. The load I’d actually put money on, for most of the people I know, is the strain that never lets up, and that one never makes it into a headline, because there’s no product to put next to it. A pan you can replace on a Saturday morning.

Where this leads

A budget with a balance in it is a model, and every model has an edge where it stops being true, so here is this one’s. Nothing in you is totalling the week. There’s no organ computing a balance, and I find tidy arithmetic reassuring, which is a fact about me rather than about the body. What the model gets right is the pooling, one reserve with many claims on it. What it gets wrong is precision, and once you start pricing your Tuesdays in points you’ve traded one delusion for a better dressed one.

A hard week isn’t damage, either. The system is built for load and gets better at carrying load by being loaded, and what it needs is for the other half of the cycle to actually happen. That half is the one that gets negotiated away first.

There’s a bigger claim I hold and can’t support at the strength I’d like. A meaningful share of what gets treated as purely a head problem is, I think, sitting on top of a body that’s been under-recovered for a long time. That’s a belief. The literature I can point at is associations, and I’d rather say the thing with its label attached than dress it up as settled.

State before load is the order we built the Atlas Cove week around. What a guest can currently carry gets established before anybody decides what to ask of them, and it gets read again at the end to see which way it moved.

Go back to that Tuesday. What I told myself afterwards was that I’d gone soft, which is the nearest explanation to hand and costs nothing to reach for. The legs had carried a call about money and six or seven decisions with clocks attached before anyone asked them for an interval, and the night that was meant to pay for all of it got cut short. Add those up. The evening explains itself. Call it a character flaw instead and there’s nothing left to do with it, which is usually the sign that you’ve picked the wrong explanation.

This is an educational and strategic perspective, not personal medical advice. The views are the author’s own and not statements by Atlas Cove Lda.

Sources

1. McEwen, B. S. (1998). Protective and damaging effects of stress mediators. New England Journal of Medicine, 338(3), 171-179. DOI: 10.1056/NEJM199801153380307

2. Chrousos, G. P. (2009). Stress and disorders of the stress system. Nature Reviews Endocrinology, 5(7), 374-381. DOI: 10.1038/nrendo.2009.106

3. Stults-Kolehmainen, M. A., & Sinha, R. (2014). The effects of stress on physical activity and exercise. Sports Medicine, 44(1), 81-121. DOI: 10.1007/s40279-013-0090-5

4. Buchheit, M., & Gindre, C. (2006). Cardiac parasympathetic regulation: respective associations with cardiorespiratory fitness and training load. American Journal of Physiology-Heart and Circulatory Physiology, 291(1), H451-H458. DOI: 10.1152/ajpheart.00008.2006

5. Russo, M. A., Santarelli, D. M., & O’Rourke, D. (2017). The physiological effects of slow breathing in the healthy human. Breathe, 13(4), 298-309. DOI: 10.1183/20734735.009817

6. Thayer, J. F., Åhs, F., Fredrikson, M., Sollers, J. J., & Wager, T. D. (2012). A meta-analysis of heart rate variability and neuroimaging studies: implications for heart rate variability as a marker of stress and health. Neuroscience & Biobehavioral Reviews, 36(2), 747-756. DOI: 10.1016/j.neubiorev.2011.11.009

Two objects, both of them sugar. A soft drink standing open on a desk at three in the afternoon, most of the way through a day with no movement in it worth the name. And a foil sachet of glucose and fructose that I tore open with uncooperative hands somewhere in the fourth hour at Cascais last October.

Same molecule in both. Six carbons, and no memory of the packet it came out of.

Where it ends up differs completely, and hardly any of that gap sits in the sugar. It sits in the body taking it in, in what that body has already stored and what it is spending right now, and neither label says a word about either. That’s the argument here. It’s more forgiving than the usual talk about sugar, and it takes more work to apply.

This is an educational and strategic perspective, not personal medical advice. The views are the author’s own and not statements by Atlas Cove Lda.

Nothing on the label decides this

Most popular advice about food rests on a quiet belief: that the quality of a carbohydrate belongs to the food, so once something has been sorted into the good pile or the bad pile it stays there whoever eats it, whenever, in whatever state. Handy, because you can print that on a page and follow it without measuring anything. It’s also why two people can follow opposite advice with equal certainty and both report that it worked.

I want to be fair to the eating patterns that do have good data behind them across large groups, the Mediterranean and the Nordic among them. They earn it. What they were never built to do is answer what one carbohydrate does inside one liver on one Tuesday, and they shouldn’t be blamed for missing a target they never aimed at.

The answer to the smaller question sits one level under the food. How full the liver’s glycogen store happens to be. What the muscles are asking for. Energy in has been running above energy out for the past week, or under it. The gut wall can only move so much across before the rest carries on down the tube. You can’t see any of that from outside, and you can’t read it off an ingredient list, which is why rules built on the food alone behave so erratically once they meet a real person.

My own irritation belongs here, so I’ll put it in. Advice built on a single named substance sells well because it turns a complicated situation into one act of shopping, and I’ve followed a version of it myself more than once. Cutting one thing out is a comfortable way to feel like you’ve acted. It also costs nothing, which is roughly why it keeps selling.

Subscribe now

Where the two sugars part company

Glucose and fructose are isomers. Same formula, C6H12O6, and the resemblance stops about there.

They separate at the gut wall. Glucose crosses through SGLT1, a sodium-dependent transporter that hauls it against its gradient, active and high in throughput, while fructose leans almost entirely on GLUT5, which lets it drift down its gradient and saturates at a far lower rate of delivery. That one gap explains a great deal of what people call fructose intolerance: fructose arriving faster than GLUT5 can shift it stays in the lumen, pulls water in behind it, and reaches the bacteria further down. Plumbing, in other words. Nobody’s metabolism is broken.

Whatever does get across meets a second fork. Glucose spreads out and gets used according to who’s asking, taken up by muscle and brain and fat tissue under insulin signalling and local energy status, with the liver holding back only somewhere in the region of twenty to thirty per cent of an oral load and the rest travelling onward (Wolfe, 1998).

Fructose behaves differently, and here’s the thing I had wrong for years. I pictured the gut as a corridor and the liver as the whole story. At modest doses it isn’t a corridor at all. The small intestine turns much of a small fructose load into glucose and organic acids before any of it reaches the portal blood, so a first-pass buffer stands in front of the liver at ordinary intakes (Jang et al., 2018). That buffer saturates. Larger single doses spill past it to the liver, and past a further threshold, on to the colon.

Which sharpens the argument about dose and weakens the one about the molecule. A pear and a litre of syrup do not put the same problem in front of the same tissue.

The enzyme nobody fitted a brake to

The fructose that does reach hepatocytes goes in through fructokinase, also called ketohexokinase, and this is the step that explains why fructose gets discussed on its own at all.

Glucose heading into glycolysis has to get past phosphofructokinase-1, an allosterically regulated enzyme that slows down once the cell already has plenty of energy. A real brake. The liver can decline to commit more glucose carbon downstream, and it does.

Fructokinase has nothing equivalent. Fructose gets phosphorylated to fructose-1-phosphate quickly and largely in proportion to how much turned up, aldolase B splits that into triose phosphates, and the trioses join the pathway below the regulated step (Tappy and Lê, 2010). Nothing toxic happens here. What happens is a loss of throttling, plus a brief local cost in ATP while phosphate gets spent faster than it comes back.

An input line with no brake on it made me sit up the first time I read this properly. I was trained to go looking for the limiter in any system, and here the limiter simply isn’t fitted. Holding that next to the second fact took me longer: this is ordinary metabolism working the way it always has.

Three plausible endings wait for those trioses: storage as liver glycogen, burning for energy right now, or a slow turn into fatty acids. Nothing about the molecule picks the winner. The state of the liver picks it, and that sentence is the whole piece.

When the liver starts making fat

De novo lipogenesis, the route that turns carbohydrate carbon into fatty acids, gets talked about as though it were a fault. It’s a normal, well-regulated pathway that mammals rely on. It becomes a plausible source of harm under one specific setup. Eating fructose is not by itself that setup, and holding the two apart is worth the effort, because it separates a real risk from a general unease.

Keep total energy intake level with what gets spent and the share of fructose carbon ending up as newly made fat stays fairly modest, while piling fructose on top of an already adequate intake for days at a stretch sends hepatic de novo lipogenesis climbing a long way and drops insulin sensitivity measurably. That’s the state where the trouble shows up (Stanhope et al., 2009; Faeh et al., 2005). The honest limit on those trials is small groups, deliberately overfed, built to push the pathway until it shows itself. They tell you what a liver does under load. What an ordinary week does sits further from the data than I’d like it to.

The reason is that the machinery running this pathway reads how much carbohydrate is around and doesn’t care which sugar brought it. When liver glycogen already sits near the top and triose phosphates keep arriving, the signalling that turns lipogenic enzymes up gets switched on, malonyl-CoA builds, fat burning gets suppressed at the exact moment synthesis gets encouraged, and the liver ships the resulting lipid out as triglyceride inside very-low-density lipoprotein. That export is one of the mechanisms behind the raised fasting triglycerides and the extra liver fat seen in the overfeeding studies.

Four things have to line up for that, and they tend to arrive together. Liver glycogen that rarely gets drawn down, because there’s little activity and meals come often. An intake sitting a little over expenditure for weeks rather than hours, a fructose intake that’s high in absolute terms and delivered in large single hits, and not enough physical work to put any real call on the carbohydrate turning up.

Read that back and it’s obvious why cutting the fruit while leaving the surplus and the sitting untouched does so little. Three of the four never moved.

How full the tank is when the sugar lands

Liver glycogen is the variable the public argument about sugar leaves out almost entirely, and together with energy balance it’s one of the two things that most reliably decide where an incoming carbohydrate goes.

An adult liver holds roughly eighty to a hundred and twenty grams of it. That store moves constantly. It gets drawn down overnight and through any fast, spent to hold blood glucose steady between meals, and emptied faster by physical work, at a rate depending heavily on how hard the work is and how well trained the person doing it happens to be (Achten and Jeukendrup, 2004). It’s a buffer. The useful question about a buffer was never what it holds in principle, it’s how full it happens to be at the moment something turns up.

Take two people eating an identical amount of fructose, which is exactly the comparison a food-focused view can’t make. The first trains several times a week and turns liver glycogen over repeatedly, so the store is often part empty, and the trioses coming off fructokinase get pulled toward glycogen synthesis and toward burning, because the liver has both room to store and reason to burn. The second sits for most of the day, keeps the store close to full, and lives in a small daily surplus, so the same trioses arrive somewhere with no room and no demand, and lipogenesis is what’s left. The molecule is the same in both cases and the outcome isn’t, which is about as clean a proof as you’ll get that the outcome was never a property of the molecule.

None of this needs a race entry to work on you. Walk more and stop adding the small unremarked surplus, and you’re moving the same two dials an athlete moves, glycogen turnover and energy balance, in the same direction. The mechanism doesn’t check your training history first.

Two doors into the same gut

That sachet in the fourth hour was one of many. Close to a hundred grams of carbohydrate an hour was going in for most of that day, much of it as a glucose and fructose blend.

The blend isn’t in there for the taste. Fructose crossing the enterocyte through GLUT5 doesn’t compete with glucose crossing through SGLT1, so the two sugars queue at separate doors. Feed both together and total uptake lifts past the ceiling glucose alone imposes, which allows exogenous oxidation rates approaching ninety grams an hour (Currell and Jeukendrup, 2008; Jeukendrup, 2014). Glucose on its own runs into that ceiling long before the legs stop asking.

I still find that satisfying. A transport limit sitting in the gut wall, quietly setting the ceiling on what any endurance athlete can do, and the way past it is a second molecule carrying its own key.

In that setting essentially none of the fructose was free to feed a lipogenic pathway. Liver glycogen was leaving faster than anything could put it back, and whole-body burning was running near its ceiling. The identical amount of the identical sugar, taken on a quiet afternoon into a full liver and a body asking for nothing, travels a different route through the same enzymes.

I’m not making an athlete’s-exemption argument here. Dose and demand are inseparable, that’s all, and the everyday version of it is that the same slice of cake lands differently depending on the day in front of it.

Three questions instead of a rule

What replaces good-or-bad sorting is a short run of questions that read the context first. You can answer all three roughly, from what you already notice, with nothing on your wrist. Roughly is fine.

1. How full is the liver’s store right now? The honest inputs are how recently and how hard you worked, how long it’s been since you last ate, and whether the day had movement in it or a chair. That reading tells you whether the carbohydrate coming in refills a part-empty buffer or presses against a full one.

2. What kind of carbohydrate is it, and does it match what’s actually getting burned? Work near threshold pulls hard on glucose and does better on a mixed feed, a longer and easier effort leans more on fat and cares about the total amount more than about what the carbohydrate is made of, and a resting body has baseline demand and nothing beyond it, so whatever turns up over that gets stored or turned into something else.

3. Where does the timing sit relative to that demand? Carbohydrate taken before or during work is topping up a store that’s actively being emptied. The same carbohydrate arriving into a rested body with full stores and an intake that’s already sufficient is asking the system to do something it has no use for.

Nothing in there requires avoiding fruit or treating any food as forbidden. It asks for an honest read of a metabolic state, which is harder than following a rule and quite a lot more accurate.

Subscribe now

Where this leads

The Atlas Cove week is built around the part of this that no label carries. What a guest’s body is actually doing gets measured before anyone talks about what they should eat, because that state is the thing deciding where the food ends up.

My own rule is short. I eat close to what a human body spent most of its history running on, as unprocessed as I can reasonably get it, meat very much included. The rule isn’t the main thing. It works because I train most days and don’t let a surplus sit for weeks, which keeps the store turning over and leaves the same food arriving somewhere quite different. Better numbers would move me off it, and nobody has produced any.

Fructose landing in an emptied liver after four hours of work and fructose landing in a full liver after a day at a desk are the same molecule doing two different jobs, and no amount of arguing about the molecule will settle which. Context decides outcome, every time I’ve gone looking. The useful work is slower than picking a list. You have to read the state a food is landing in, and no label has ever carried that.

Part two takes the same approach to creatine, which most people still file under bodybuilding and which turns out to be doing something a good deal more interesting.

This is an educational and strategic perspective, not personal medical advice.
The views are the author's own and not statements by Atlas Cove Lda.

Sources

1. Wolfe, R. R. (1998). Metabolic interactions between glucose and fatty acids in humans. American Journal of Clinical Nutrition, 67(3), 519S-526S. DOI: 10.1093/ajcn/67.3.519S

2. Jang, C., et al. (2018). The small intestine converts dietary fructose into glucose and organic acids. Cell Metabolism, 27(2), 351-361. DOI: 10.1016/j.cmet.2017.12.016

3. Tappy, L., & Lê, K. A. (2010). Metabolic effects of fructose and the worldwide increase in obesity. Physiological Reviews, 90(1), 23-46. DOI: 10.1152/physrev.00019.2009

4. Stanhope, K. L., et al. (2009). Consuming fructose-sweetened, not glucose-sweetened, beverages increases visceral adiposity and lipids and decreases insulin sensitivity in overweight/obese humans. Journal of Clinical Investigation, 119(5), 1322-1334. DOI: 10.1172/JCI37385

5. Faeh, D., et al. (2005). Effect of fructose overfeeding and fish oil administration on hepatic de novo lipogenesis and insulin sensitivity in healthy men. Diabetes, 54(7), 1907-1913. DOI: 10.2337/diabetes.54.7.1907

6. Achten, J., & Jeukendrup, A. E. (2004). Optimizing fat oxidation through exercise and diet. Nutrition, 20(7-8), 716-727. DOI: 10.1016/j.nut.2004.04.005

7. Currell, K., & Jeukendrup, A. E. (2008). Superior endurance performance with ingestion of multiple transportable carbohydrates. Medicine & Science in Sports & Exercise, 40(2), 275-281. DOI: 10.1249/MSS.0b013e31815adf19

8. Jeukendrup, A. E. (2014). A step towards personalized sports nutrition: carbohydrate intake during exercise. Sports Medicine, 44(Suppl 1), 25-33. DOI: 10.1007/s40279-014-0148-z

54.7 in April. 52.4 in July. 51.6 in October.

That’s my own VO2 max across one build, three measurements, and the direction is the one nobody wants. On 18 October 2025 I raced a full-distance Ironman at Cascais, 225.97 km, 11 hours and 8 minutes, at 30 years old, and the day went close to how I’d planned it. Both of those things are true. They sat badly together for a while. I’d spent the summer watching the figure I’d been taught to treat as the headline slide, or rather I’d spent the summer half-ignoring it while quietly assuming I would pay for it in October.

I didn’t. That irritated me more than it should have, because it meant I had been reading the wrong instrument all year.

One build in one 30 year old man establishes nothing, and I want that said before anything else follows, because the case here is the literature and my own numbers are illustration and nothing more. Past roughly ninety minutes of continuous work the height of your aerobic ceiling stops explaining much about what happens, and the property that does explain it has a name and a set of parts you can train.

This is an educational and strategic perspective, not personal medical advice. The views are the author’s own and not statements by Atlas Cove Lda.

A ceiling the day never visits

The measurement is honest about itself if you read the protocol. You get put on a treadmill or an ergometer, the load rises until you cannot hold it any longer, usually somewhere in the eight to twelve minute range, and the plateau in oxygen uptake at the top is your number. What sets that plateau is mostly central: how much blood the heart can push per minute, and how much oxygen the working muscle can pull out of it on the way past. Both are real physiological traits. Both respond to a few weeks of hard intervals, which is why the number is popular: a figure that moves inside six weeks feels like proof that something is happening.

It is proof that something is happening. What happened is that your ceiling went up, which is a real adaptation and a genuine gain in a laboratory, and it is also the part of your physiology that the long day is least interested in.

Then you go and race for eleven hours and never visit it. My average heart rate across the whole day at Cascais was 143, which is 72 percent of my maximum, and a maximal test has nothing to say about a body held there for eleven hours, because it was never built to. It answers one question, cleanly: how much oxygen can you use when you’re trying as hard as you can, for ten minutes, fresh, while the distance wants to know how economically you run well below that peak, how long that economy survives once you’re tired, and which fuel you’re drawing on in hour five. Two people can walk to a start line with the same maximal figure and finish forty minutes apart. Nothing mysterious about it. They shared a trait the day wasn’t testing.

Subscribe now

The point where more ceiling stops paying

None of this makes aerobic capacity unimportant, and the case for raising it is strongest exactly where most people are standing, because below a certain level the ceiling really is the binding constraint and lifting it pays back in almost anything lasting longer than a few minutes. If you’re sedentary, or training lightly around a full week, that’s your situation and the standard advice to build aerobic capacity is good advice. I’d take it.

What changes is what sits above that. Among trained endurance athletes, and the level where this sets in seems to sit somewhere in the low fifties in millilitres per kilogram per minute (a working figure picked up from other people’s practice, and I’ll come back to how soft it is), more headroom adds surprisingly little to the pace you can hold for hours, because the question has moved. How high the ceiling reaches stops mattering much. How much of it you can occupy, for how long, with how little decay along the way, starts mattering instead.

Plainly, then: past that point, training weeks spent chasing the number are weeks spent on the wrong thing. That’s an opinion, and it stands until somebody puts long-course results in front of me that track maximal figures better than they track how much each athlete faded.

If your plans contain no start line anywhere, the translation is the same one. Once a reasonable base exists, more maximal capacity is rarely what stands between you and a long coastal walk, or an hour of steady work that doesn’t write off the evening. Those run on sustainable capacity and on how well you hold up under load that has been accumulating, and both of those train separately from the maximum.

Two beats across eleven hours

The literature gave this a name a few years ago, which helped more than a name usually does. Durability has been defined as how well an athlete’s physiological profile resists deterioration over prolonged exercise, and proposed as a determinant of endurance performance sitting alongside maximal capacity, the sustainable threshold and efficiency (Maunder, Seiler, et al., 2021). What I like about it is that it says out loud the thing laboratory testing structurally cannot do. Any rested measurement, the maximal test included, is a portrait of you at the start, taken under conditions kept deliberately clean of fatigue. Durability asks how much of that portrait is still recognisable in hour six, with glycogen drawn down, core temperature up and the same movement repeated tens of thousands of times.

A profile that decays gently and a profile that falls off a cliff look identical when both are measured fresh.

Here is my own day, broken up. Swim, 1 hour 17, average 143. Bike, 5 hours 40, average 142. Run, 3 hours 55, average 144. Two beats of movement across three disciplines and eleven hours of work.

I’d be overclaiming badly if I called that proof. Heart rate on its own can sit dead flat while pace quietly falls apart, a conservative pacing plan flatters anyone’s numbers, and there’s no control condition here and no second version of me who trained some other way. What it does show is that the system I had trained held its shape for the length of the event, which is the property the event was charging me for. The figure that had been sliding all summer described a ceiling the day never asked about.

This one travels out of sport more cleanly than most training ideas do, because very little of adult life happens rested and under laboratory conditions. What you can still do reliably on the second Thursday of a hard fortnight is a durability question wearing ordinary clothes.

The adaptations that take a winter

Several separate things do the work a long effort depends on, and each of them responds to a different stimulus than the one that lifts a maximal test. Start in the mitochondria. Endurance training raises both the volume of mitochondria inside muscle and the activity of the oxidative enzymes in them, which is what Holloszy established when the oxidative capacity of trained muscle roughly doubled in response to a programme of running (Holloszy, 1967). That was 1967. Human work since has added a layer, because the density of the internal cristae membranes changes too, running higher in endurance-trained muscle and tracking with whole-body oxygen uptake, so the same volume of mitochondria can present a considerably larger working surface (Nielsen et al., 2017). Between them these set how much sustained aerobic work a muscle can produce at a moderate intensity it can repeat all day.

Then there’s delivery, at the level of the tissue rather than the pump. Sustained easy volume grows the capillary network around each muscle fibre, and the classic human demonstration found capillary density in the quadriceps up by roughly twenty percent after eight weeks of training (Andersen and Henriksson, 1977). A denser network shortens the distance oxygen has to travel into the fibre and metabolites have to travel out of it, which improves the exchange most at submaximal intensities. Peak power is limited by neural recruitment rather than by local oxygen supply. So this particular adaptation does very little for your maximal test and a great deal for a long steady one.

The parts depend on each other. Skip one and the others stop paying: extra mitochondria with no capillary network to feed them sit there limited by supply, and better fuel handling without the recovery to sustain the training that produced it ends up in the same place. They get built together, across months of accumulated aerobic work, and that is the honest reason they attract less attention than a figure capable of moving inside a fortnight.

Fuel, where the test is blindest

The third adaptation is metabolic, and it decides the most over long durations while sitting furthest outside anything a maximal protocol can observe. Months of consistent aerobic work make the muscle better at oxidising fat and better at clearing lactate, so at any given submaximal intensity a larger share of the demand gets met out of fat stores and a smaller share out of the finite glycogen supply. Comparisons of professional endurance athletes against less-fit individuals show precisely this pattern, with higher rates of fat oxidation and lower blood lactate at matched intensities, which is what metabolic flexibility looks like when measured (San-Millán and Brooks, 2018).

Across many hours that compounds until it is close to determinative. Nothing clever sits underneath that, only the arithmetic of the two tanks: glycogen runs out, and fat, for any practical purpose, does not.

Cover the early hours with a higher share of fat and you reach the closing stages with reserves still there. Cover them leaning harder on carbohydrate and you arrive having spent them. A maximal test, meanwhile, measures you in a state where the body is obliged to run almost entirely on carbohydrate, which is very nearly the opposite of the metabolic state a long day depends on. It’s a good test. It’s aimed elsewhere.

If you’re racing nothing at all, the same flexibility shows up as energy that holds across a long working day, and as a body that isn’t tightly hooked on being fed every two hours. Both are worth having whether or not you ever pin a number to your shirt.

Why so much of it is easy

What well-trained endurance athletes actually do turns out to be remarkably consistent, once you look at their logs instead of at training theory. Examined across sports, the distribution settles near eighty percent of sessions at genuinely low intensity, with the remainder concentrated in properly hard work and comparatively little time spent in the moderate band that sits between the two and feels productive while you’re in it (Seiler, 2010). The middle band is the tempting one. The distribution keeps reappearing because it maps onto how the body answers. Easy volume drives the mitochondrial, capillary and metabolic changes above without generating fatigue that has to be repaid afterwards. A small dose of hard work keeps the top end honest. The middle costs a lot of fatigue in exchange for adaptations it doesn’t specifically target, so it gets avoided by default. A controlled nine-week comparison in well-trained athletes pointed the same way, with a polarised model producing greater improvements in the key endurance variables than a threshold-centred, high-intensity or high-volume alternative (Stöggl and Sperlich, 2014).

Anyone fitting real training around a full working life has a sharper version of this problem, because the fatigue budget is shared with everything else the week wants. Read from that angle, the eighty percent principle is a way of accumulating a large amount of useful stimulus at a price the rest of your life can absorb.

The number fell because I stopped asking for it

Set out in that order, the result stops being a paradox. A block built on easy volume with a modest amount of hard work is not the stimulus that holds a maximal test at its peak, because keeping VO2 max at the top of its range takes regular near-maximal efforts and an endurance block deliberately keeps those scarce. So the figure drifts down across exactly the months in which the systems that carry a long effort are getting stronger. Accumulated fatigue nudges it the same way. So does ordinary day-to-day variation in the test itself.

For a couple of weeks in July I treated my 52.4 as a fault to be found and corrected, went looking for what had gone wrong in the training, and came close to putting intervals back into a plan that had no room for them. Wrong reading, wrong instrument. Had I acted on it, August would have cost me real fatigue in exchange for a number that October never asked to see.

Which brings me back to where this started. Three measurements, one person, one build, no control and no repeat, and the honest weight of that is somewhere near zero. It illustrates a mechanism the literature describes. Take the literature away and what’s left is three numbers off my own body, which would be worth nothing at all to anybody, including me.

What I watch instead

The practical consequence lands on what gets tracked rather than on how hard anyone trains, because a maximal test reports a ceiling a long effort rarely approaches, which makes it one input among several and nothing like the headline. Four signals describe sustained capacity better.

• How stable heart rate stays across a long effort.

• Whether pace at a fixed heart rate improves over a training block.

• How efficiently you oxidise fat at the intensity the day will actually be run at.

• How well output holds up in the later hours, when it’s under pressure.

The last one is the hardest to get at, and it’s the one closest to what the day charges you for. None of the four collapse into a single tidy figure, which is why they get quietly passed over in favour of one that’s easy to publish and easy to compare. My own posture now is to treat VO2 max as a threshold check. Confirm it’s comfortably high enough. Then stop optimising it and put the attention on the adaptations the distance rewards.

By my own standard I’m the awkward case. 51.6 in October. The line I drew a few sections ago sits in the low fifties, so my own figure is on it rather than comfortably above it. The tidy move would be to shift the threshold down until I cleared it, and the accurate one is to admit that the low fifties is a rule of thumb I’ve absorbed over the years and can’t hand you a citation for, which leaves its edges soft enough that my figure is sitting on one of them. What I’d watch from here is the direction of travel, across years rather than one build.

The same preference runs through how an Atlas Cove week is built. Almost any decent week can produce a peak, so the measurements that matter are the ones taken again in month three, when the conditions have stopped being ideal. Part two picks up the other half of holding together: where mobility work earns its place, and where loading the tissue does more.

A figure that slips while your endurance improves is a bill you paid on purpose. What you bought is the capacity the distance rewards and, as far as I can currently tell, the capacity the decade rewards too.

This is an educational and strategic perspective, not personal medical advice. The views are the author’s own and not statements by Atlas Cove Lda.

Sources

• Seiler, S. (2010). What is best practice for training intensity and duration distribution in endurance athletes? International Journal of Sports Physiology and Performance, 5(3), 276-291. DOI: 10.1123/ijspp.5.3.276

• Stöggl, T., & Sperlich, B. (2014). Polarized training has greater impact on key endurance variables than threshold, high intensity, or high volume training. Frontiers in Physiology, 5, 33. DOI: 10.3389/fphys.2014.00033

• Holloszy, J. O. (1967). Biochemical adaptations in muscle. Journal of Biological Chemistry, 242(9), 2278-2282. DOI: 10.1016/S0021-9258(18)96046-1

• Nielsen, J., et al. (2017). Plasticity in mitochondrial cristae density allows metabolic capacity modulation in human skeletal muscle. Journal of Physiology, 595(9), 2839-2847. DOI: 10.1113/JP273040

• San-Millán, I., & Brooks, G. A. (2018). Assessment of metabolic flexibility by means of measuring blood lactate, fat, and carbohydrate oxidation responses to exercise in professional endurance athletes and less-fit individuals. Sports Medicine, 48(2), 467-479. DOI: 10.1007/s40279-017-0751-x

• Andersen, P., & Henriksson, J. (1977). Capillary supply of the quadriceps femoris muscle of man: adaptive response to exercise. Journal of Physiology, 270(3), 677-690. DOI: 10.1113/jphysiol.1977.sp011975

• Maunder, E., Seiler, S., et al. (2021). The importance of ‘durability’ in the physiological profiling of endurance athletes. Sports Medicine, 51(8), 1619-1628. DOI: 10.1007/s40279-021-01459-0

There’s a version of this essay that names a villain. Greedy pharma, cynical insurers, doctors who stopped caring. It reads well, and I’m not going to write it, because if the problem is somebody’s character then the fix is finding better people, and better people won’t touch this.

So, the narrow claim. European health systems are genuinely good at finding and treating illness that has already turned up, and close to structurally blind to the illness that hasn’t arrived yet, and the reason sits down in the plumbing rather than anywhere near anyone’s motives. Keeping a person well produces nothing an institution can bill for, and institutions do what they can bill for.

That’s an accounting problem, and it changes what it makes sense to walk into a doctor’s office and ask for.

This is an educational and strategic perspective, not personal medical advice. The views are the author’s own and not statements by Atlas Cove Lda.

Three percent of the money

I trained as an engineer, so my instinct with any system I don’t understand is to go and look at what it spends. Strategy documents describe intentions. Budgets describe behaviour.

Across OECD countries, spending classified as prevention has sat at around 3 percent of total health expenditure, and it lifted toward 6 percent during the pandemic before settling back near where it started (OECD, 2023). Which leaves well over 90 percent pointed at illness that already arrived. I’d read that as a description of what the apparatus can see, fund and account for, and not as a statement about what anybody inside it believes.

The mechanism sits in how providers get paid, and the two dominant models push the same way for opposite reasons. Fee-for-service pays per service delivered, so billable volume sets the revenue, and anybody responding sensibly to that signal does more of whatever can be coded. Case-based payment was built to stop that. The Diagnosis-Related Group systems now underpinning hospital finance across most of Europe pay a fixed tariff per classified case, set by diagnosis, procedure and complexity (Busse et al., 2011). Germany, France, Spain and Portugal all run on them. Inside the hospital it works, because a fixed tariff turns each admitted case into something to handle efficiently rather than expansively, and restraining the volume incentive was the whole point of bringing it in.

No version of it generates revenue for the case that never arrived.

Porter and Teisberg described the general shape of this twenty years ago, arguing that health systems compete on the volume of services they deliver and on shifting costs onto each other rather than on what a patient actually gets across a whole cycle of care (Porter and Teisberg, 2006). Prevention fits that competition badly. An illness that doesn’t happen produces no volume, no code and no line in any quarterly report anyone is answerable for.

Subscribe now

Fifteen minutes, and the wrong question inside them

The structure reaches all the way down into a single appointment, which is where it stops looking structural and starts feeling like bad luck or a personal failing.

You walk in with a question about direction rather than about a symptom. Where is this going, over ten years, given what you can see? That room is built to produce one output, which is whether a diagnosable illness is present and what’s indicated if it is, and a slow drift across several markers that are each still sitting inside their reference range is an interesting signal to anyone thinking about a decade and a complete non-event for that decision.

Nothing went wrong in the room. The room did its job.

Berwick made the general point back in 2002, writing about the Institute of Medicine’s quality work: what a health system produces follows from how it was designed and from what that design measures and rewards, so expecting different output out of the same structure by exhorting people harder was never a serious plan (Berwick, 2002). Systems count procedures performed and cases coded. They count avoided illness and preserved years of good function very badly, because the first group has identifiers attached to it and the second doesn’t.

The gap turns up even inside what systems fully intend to do. The best known attempt to measure it found that adults in the United States got the recommended standard of care roughly 55 percent of the time across a wide range of conditions, and the authors were explicit that this was systems failing to deliver what was already known rather than clinicians who didn’t know it (McGlynn et al., 2003). European systems are financed differently and would produce different numbers. They share the underlying difficulty, which is doing consistently what they aren’t structured to measure.

I was annoyed at the wrong people

Two winters and two springs, I lost whole days at a time to severe migraine, aching like flu and a tiredness that sleeping didn’t touch. I’d go in, bloods would come back unremarkable, somebody would tell me it all looked fine, and I’d leave with nothing. Four times, I think. Maybe five. Somewhere in there I’d quietly filed the whole thing as a lack of interest.

I had it wrong, and I’d decided who was at fault before I’d checked anything.

Every one of those appointments answered its own question correctly. Is there a diagnosable illness here? No, there isn’t. I was walking in with a different question, about regulation rather than about any one organ, and I kept putting it to the one institution in my life that has no mechanism at all for receiving it. Part of mine does have a name, as it happens: cold urticaria, mast-cell driven, with a family history of autoimmune dysregulation sitting behind it. A name is worth having, and it still isn’t a plan.

What eventually moved things took months and would have looked like nothing on a chart. I stopped drinking, went to bed and got up at hours I stopped negotiating with, and pulled the intensity out of my training for long enough that it stopped really being training. Recovery came back well before performance did, in that order, which I hadn’t expected and didn’t much enjoy. None of it was ever going to be prescribed to me in a fifteen minute slot, because none of it is a thing you can do to a person once.

Bismarck, Beveridge, same direction of travel

A fair objection lands here. Europe contains several genuinely different health systems, so surely one of them handles this better than the others.

The differences are real, and I’ve been a patient in two of them. Germany, where I’m from, runs a Bismarck model, care financed through social insurance contributions and a plurality of sickness funds, while Portugal, where I live and work now, runs a Beveridge model financed out of general taxation with the state as the dominant purchaser, and various hybrids sit in between the two with their own arrangements for cost sharing and waiting times. From the inside they feel nothing alike. What you pay and what you wait for are different, reaching a specialist is a different exercise entirely, and anybody claiming a patient in Berlin and a patient in Lisbon are in the same situation has not tried both.

Neither of them ever asked me a question about the next ten years.

What they do share is which way the money runs, because in each of these arrangements the money follows the diagnosis and treatment of illness that already exists (OECD, 2016). Prevention is also a set of things people do, and things people do can in principle be paid for. The catch is what triggers the payment. Payment is triggered by a diagnostic or therapeutic event attached to an illness that exists, so work whose success consists of an illness never existing has nothing to trigger on.

A hospital paid per case manages case mix and length of stay. A provider inside a fixed global budget contains demand within that budget, and prevention lowers demand at some distance in the future while eating resources now, which returns nothing at all inside the period that provider gets judged on. Mixed systems inherit both logics and arrive in the same place. Moving country doesn’t solve it, and the OECD’s own review of how to pay for care still treats rewarding prevention as an open question rather than a settled one (OECD, 2016).

You can’t send an invoice for an absence

None of this needs anyone acting in bad faith, and it’s worth saying that flatly, because an incentive structure seen from the inside looks a lot like malice when it’s only economics behaving normally.

Treating a person is easy to define and therefore easy to pay for. A procedure has a code, a diagnosis carries a tariff, an admission produces recorded activity and recorded revenue, and a payer can audit every one of those. Prevention fails at all of them. There’s no invoice for the heart attack that didn’t happen, and no billing identifier for a decade of function maintained.

A second economic property sits underneath, and this is the part I find interesting. Medical care is a credence good, meaning the buyer usually can’t judge the quality of what they received and has to trust the seller’s judgement about what was needed, which is most of why medicine ended up regulated and professionalised as heavily as it is. Prevention is the extreme case, because its successful output is an absence, and an absence can’t be inspected or compared or pinned on anyone. A market can’t price the permanently invisible. A public payer can’t easily defend a budget line whose returns arrive in a later electoral cycle, in a different institution’s accounts, under somebody else’s name.

The gap that leaves is wider than it first looks, because clinical care was never the main input to health. Work on the determinants of population health put around 10 percent of premature death down to shortfalls in medical care, with behaviour, social circumstances, genetics and environment carrying most of the rest (McGinnis et al., 2002). Marmot arrived somewhere compatible from the opposite direction, and showed in detail how the conditions people live and work in shape health to a degree clinical care can’t offset (Marmot, 2010).

Set that beside the spending figures and the shape of the problem is obvious. Systems put nearly all of their money and nearly all of their attention on the small share of health that medical care directly determines, while the much larger share, which is also where nearly all the room for prevention sits, lies outside what they fund or count at all. Individual decisions matter enormously in there. They’re being made in a space nobody is paid to occupy.

Subscribe now

Four questions worth sitting with

I’d rather none of this got turned into an argument against clinicians, most of whom understand prevention perfectly well, want to help, and are working inside a structure they didn’t design and can’t move on their own. Seeing the structure is useful for a smaller reason than blame, because once you can see it you stop asking it for the one thing it isn’t organised to supply.

Four questions locate where any given system stands. Aim them at the system, not at whoever happens to be in the room.

• Does this place measure prevented illness, or only treated illness?

• If I stay well and use fewer services over the next decade, does that return anything to me, or to whoever treats me, or does it simply disappear?

• When something gets measured in me, can somebody explain why that one and not another?

• Is prevention funded here as infrastructure with its own budget, or bolted onto the budget for treating people?

For most people inside a European system the honest answers all point the same way, and it took me until my thirties to ask any of them. There’s nothing embarrassing in that. Financing structures stay close to invisible from the inside until someone goes looking for them.

Where this leads

Fixing this at the level of a system is slow collective work, because it means changing how success gets defined, how it gets measured and how resources move across whole institutions and several budget cycles at once. I’d like that to happen. I don’t plan around it.

What one person can do is narrower and considerably more useful. Stop treating a treatment system as though it were a prevention system. Use it well and gratefully for the acute and diagnosable problems it handles with real skill, and assemble the prevention layer deliberately somewhere else instead of waiting for somebody to provide it.

That gap is why my sister and I started Atlas Cove. The week stands in the long stretch of ordinary life where no institution is paid to be present, beginning from a person’s own measurements and ending with decisions that have the reasoning attached, and it replaces nothing, because medical care stays the right answer for genuine illness. I’m confident about the gap and a good deal less sure that one week is the right size of answer to it.

Visit Atlas Cove

Personal health rarely stalls for want of effort, and among people who read essays like this one it almost never does, because it stalls instead at the boundary of a system that measures treated illness precisely and prevented illness hardly at all. Frustration is the wrong answer to that, and so is hunting for a better country to be a patient in. Learn the incentives well enough to use each part of the thing for what it’s actually good at, and build the rest yourself.

This is an educational and strategic perspective, not personal medical advice. The views are the author’s own and not statements by Atlas Cove Lda.

Sources

• OECD (2023). Health at a Glance 2023: OECD Indicators. OECD Publishing, Paris. DOI: 10.1787/7a7afb35-en

• Busse, R., Geissler, A., Quentin, W., & Wiley, M. (Eds.) (2011). Diagnosis-Related Groups in Europe: Moving Towards Transparency, Efficiency and Quality in Hospitals. European Observatory on Health Systems and Policies / Open University Press. ISBN 9780335245574

• Porter, M. E., & Teisberg, E. O. (2006). Redefining Health Care: Creating Value-Based Competition on Results. Harvard Business School Press, Boston. ISBN 9781591397786

• Berwick, D. M. (2002). A user’s manual for the IOM’s ‘Quality Chasm’ report. Health Affairs, 21(3), 80-90. DOI: 10.1377/hlthaff.21.3.80

• McGlynn, E. A., et al. (2003). The quality of health care delivered to adults in the United States. New England Journal of Medicine, 348(26), 2635-2645. DOI: 10.1056/NEJMsa022615

• McGinnis, J. M., Williams-Russo, P., & Knickman, J. R. (2002). The case for more active policy attention to health promotion. Health Affairs, 21(2), 78-93. DOI: 10.1377/hlthaff.21.2.78

• Marmot, M. (2010). Fair Society, Healthy Lives: The Marmot Review. Institute of Health Equity, UCL

• OECD (2016). Better Ways to Pay for Health Care. OECD Health Policy Studies, OECD Publishing, Paris. DOI: 10.1787/9789264258211-en

I apologised. Properly. No “sorry if you felt that way”, but a clear “I was wrong, I see why it hurt, here’s what I’ll do differently.”

She refused to forgive me.

One strike and you’re out. No repair. No messy middle. Just cut.

A few days later I sat in a Catholic church on Christmas Day. The priest preached a sermon of which maybe 20% felt alive and timely. The rest felt frozen in time. When I looked around, I saw almost only grey hair. I have my own little grey hairs (I cover them), but my generation was barely in the room.

That’s when a sentence crystallised in my head:

As a society, we have lost the plot.

Not just about religion. About what a healthy life actually looks like.

This is an educational and strategic perspective, not personal medical advice.

We now treat “healthy” as a performance target

Ask an educated professional what it means to live well, and you’ll usually hear some version of the same list:

Eat vegetables.
Go to the gym.
Don’t smoke.
Do your job well.

It’s a performance protocol in wellness clothing.

If your lab results come back “within normal range”, you get a gold star. If your step counter is respectable and your LinkedIn looks busy, you must be doing fine.

Except you’re not.

I see this in my own circle and in the people I think about when designing the operating system I am building, Atlas Cove Health. High-achieving, constantly on, permanently a bit tired, permanently “catching up”. Good salaries, acceptable lab work, lives designed for eventual breakdown.

We’ve shrunk health down to what can be eaten, counted, lifted, and billed. We’ve almost entirely lost the parts that make a life sustainable: real rest, deep relationships, the ability to forgive and be forgiven, a sense of being held by something larger than our own productivity.

If you trace how we got here, the path runs straight through the decline of the institutions that used to teach those things, including the Church that raised me.

What we quietly removed from the week

My parents weren’t church-every-Sunday Catholics. They’re Catholic in the way many Europeans are: they’ll say they are, and that’s almost the extent of it.

I ended up at a Catholic girls’ boarding school because it was the best education within 90 minutes of our home. It happened to be run by nuns. The regime was strict to the point of comedy compared with my friends’ normal households.

Every Sunday at 9 a.m., no matter how hard we had partied, we were in church. Every school day at 8 a.m., before whatever class came next, we prayed. In French, Latin, English, Greek, German. We prayed more than most people today answer Slack messages.

I’m not nostalgic for the control. When I left, I snapped the elastic. I went to London, then Amsterdam. I did not behave like a convent girl. I’m not arguing we should recreate that system.

But buried inside it was something quietly valuable.

You put the common good above your immediate whims.
You repaired relationships because you had to see those people at breakfast.
You rested on Sunday because the entire system shut down around you.
You were not the centre of the universe.

And there was something else that’s easy to miss if you’ve never really been inside it.

When I think back to Mass, what I miss most isn’t the theology. It’s the operating system for being a human among other humans. In a single hour you touch several things our nervous system is built for. You stand and sit and kneel alongside other bodies. You turn to the people next to you, look them in the eye, and say “peace be with you” while you shake their hand. You sing together, which means you breathe together. You sit close to strangers and are expected to be respectful, quiet, present. You hear the name of someone who is ill or who has just died and, even if only a small fraction of the congregation actually knows them, everyone is asked to hold that person in mind.

It is a weekly reminder that we keep each other alive. That survival is a communal project, not just an individual performance.

None of this means everyone experienced church as sanctuary. For many people, church was a source of real harm, and leaving was an act of self-preservation. I’m not asking them to miss it. I’m saying that when we burned down a damaged house, we didn’t keep the few structural elements that actually held the roof up.

Across Europe, that world has largely been dismantled. Weekly Mass used to be normal. Now, in many places, it’s a niche activity. The Church helped cause this decline through its own scandals, hypocrisy, and refusal to change. Trust was burned. People walked.

This is mostly a Western story. In large parts of the world, churches, mosques and temples are still full. But even there, the same attention economy and work patterns are arriving fast. The tension between communal survival and individual optimisation is becoming global.

None of this excuses the Church’s failures. It just makes their consequences even clearer.

In tearing the institution down, we also tore out one of the last places where ordinary people, across classes, practised limits, rest and repair in their bodies, together, every week.

We did not build anything serious to replace that. That is one way a society quietly loses the plot.

Subscribe now

The new pastor lives in your pocket

Into that vacuum stepped the smartphone.

The typical social media user now spends hours a day on platforms. For younger generations, these aren’t just entertainment channels. They are where you learn what is normal, desirable, morally acceptable.

Health is a good example. Many people in their teens and twenties now say they get most of their health and wellness advice from social media. A non-trivial number end up worse off after following what they see there. It’s not surprising if you look at the incentives.

The incentives are simple and the opposite of pastoral:

Outrage and purity go viral.
Nuance and repair do not.
Performative self-care sells better than boring, real rest.

So the new catechism sounds like this:

If someone hurts you, they’ve “shown you who they are” — cut them off.
If a relationship is hard, it’s “toxic” — leave.
If you feel empty, you’re “burnt out” — book a retreat, buy a supplement stack.

Some of this is genuinely protective. People have escaped situations that previous generations would have been trapped in.

But as a default, it creates a culture with no tolerance for friction and no skill at repair.

My pre-Christmas conflict with my friend is one small example. I apologised. She walked. That is her right. Scaled up, this pattern erodes our capacity to build any relationship that can survive real life. We end up with networks of shallow ties and almost no one who sees us at our worst and stays.

We’ve quietly shifted from a culture organised around communal survival — “how do we keep each other alive and well?” — to one obsessed with individual survival: my boundaries, my optimisation, my brand of self-care. Our bodies haven’t caught up with that shift. They still expect to be part of a village.

The algorithm isn’t evil in a fairy-tale sense. It’s just doing its job inside an ad-driven attention market: keep you there. It optimises for engagement because that’s what we pay it for, not because it hates your cardiovascular system.

An algorithm optimised for engagement does not care whether you reach old age with a stable blood pressure and three friends you can call at 3 a.m. It cares whether you scroll.

We handed that algorithm the teaching role the parish priest used to hold — but without the constraints, without embodied community, and without responsibility.

It’s another way we, as a society, have lost the plot: we replaced a flawed, physical operating system for communal life with a flawless, digital operating system for attention extraction.

A week that looks like rest and isn’t

Now put this together with how we handle time.

Across Europe, many adults see family or friends at most once a month. Loneliness is common and strongly linked to worse health and higher mortality.

The pandemic poured gasoline on this: more remote work, less casual contact, more time online, even more distrust of institutions. We were already drifting towards isolated, always-on lives; COVID just sped the drift up.

On the work side, irregular schedules and weekend work have crept into normality. When companies push into Sundays and evenings, sickness and absenteeism rise. People’s bodies tell the truth that their calendars deny.

It’s also harder to live sanely when your rent eats half your salary, your boss can message you at any hour, and your “Sunday” depends on a shift rota. We didn’t just become more selfish; we built an economy that treats human bodies like infinitely available infrastructure.

I used to be furious as a teenager that supermarkets in Germany were closed on Sundays. If you forgot to buy food on Saturday, that was it. No emergency dash to the shop.

From a health perspective, I now see that as a feature, not a bug. Sunday rest is literally written into German law. It’s inconvenient for shoppers, but it’s also one of the last structural protections for shared rest, especially for lower-paid workers who can’t negotiate their own schedules.

Most of us now live a patterned week that looks like this:

Work hard.
Commute.
Eat something.
Scroll on the phone until you can’t keep your eyes open.
Sleep too little.
Repeat.

From the outside, this looks like work plus rest. From the inside, physiologically, it is work plus low-grade stimulation.

Your limbs are still, but your brain is running on comparison, envy, outrage, and fear. Your nervous system is bathed in light and emotion. Then you ask it to “switch off” on command.

That is not recovery. It is a nervous system that is always “on”, with brief moments of sedation.

When I look at my high-achieving friends, I can tell you roughly how much they sleep. They run from one thing to the next. They are celebrated for being always available, always responsive. Then they “relax” by staring at a small bright rectangle.

My father lived an analogue version of this. Deeply successful, constantly travelling, dinners, late nights, stories of dangling from chandeliers. Later he had a heart attack and several episodes of atrial fibrillation, an irregular heart rhythm.

There is no mystery here. You cannot live in permanent output mode with only cosmetic rest and expect no consequences.

We engineered a world where your body is alone in a room and your nervous system is in a stadium. Then we call that “rest”.

This is what “losing the plot” looks like at the level of a week.

Subscribe now

The “soft” things the data won’t let us ignore

In medicine and diagnostics, we like numbers: cholesterol, HbA1c, eGFR. They graph well.

Forgiveness and rest sound airy by comparison. But they show up clearly in the evidence.

On forgiveness, people who are more willing to forgive tend to report better mental health and more satisfying close relationships. In older adults, forgiveness is associated with higher perceived quality of life. In people living with chronic illness, higher forgiveness and lower rumination correlate with lower stress and better self-reported health and happiness. The causal chains are complex, but the mechanism is simple: unresolved conflict is a chronic stressor.

On rest and rhythms, the picture is just as clear. Sleep deprivation and irregular circadian rhythms are strongly associated with higher risk of cardiovascular disease, metabolic disorders, mood disorders and immune dysfunction. The old “one day in seven” Sabbath pattern turns out to be a crude but effective way to force regular down-shifts of the nervous system. When work and commerce invade every day, sickness and absenteeism rise. The body insists on a rhythm, whether the labour market agrees or not.

So when I say we have lost the plot, I’m not speaking as a nostalgic Catholic pining for incense. I’m speaking as someone who cares about bodies and systems, noticing that we have stripped away the structures that made real rest and repair unavoidable, and replaced them with structures that make stimulation and cutoff unavoidable.

What we can still do, one life at a time

I’m not interested in being the person who says “everything is broken” and then offers nothing. I’m also not naïve enough to think a single article can fix a culture.

But I know, from health and behaviour change, that real change starts with awareness. You cannot shift a pattern you cannot name. If this piece does anything, I hope it puts words to something you already feel:

My labs are fine, my life is not.

From there, a few levers are within reach.

Re-introduce a weekly Sabbath, even if you are secular. Pick one day, or half-day, each week where the rules are clear: no work, no email, no social media, no “catching up” on admin. Do something slow with other humans: walking, cooking, reading, talking. Whenever possible, make this shared. Rest is easier when the people around you are resting too. Real Sabbath is easiest when it’s supported by law and norms; if you don’t have that, you have to fake it at the level of your household or friend group.

Practise repair as a default before cut-off. Before excluding someone from your life, have one hard conversation by default, especially in long-standing relationships. Forgiveness is not the same as accepting abuse. There are times when walking away is the only sane option. But if “one strike and you’re out” becomes your general rule, you end up living in a fragile social world where everything is disposable and nothing is deep.

Choose thick communities over endless content. Most of what we call “community” now is thin: a group chat, a follow, a comment thread. Look for at least one thick community that exists offline, meets regularly, and expects things of you. That might be a parish, a choir, a sports club, a volunteer group, a meditation group, a neighbourhood association. The label matters less than the structure: you are a participant, not just a consumer, and sometimes you are inconvenienced.

Those are the places where you will be forced to rest, to repair, and to accept limits. In other words, the places where you will accidentally practise the parts of healthy living that never show up in your blood work.

How this shapes Atlas Cove Health

I’m not building a church. I’m building an operating system for health: Atlas Cove Health.

I’m not going to prescribe confession or Sunday Mass. I’m going to use diagnostics, clinical judgment and system design. But I am also going to say this, plainly: if we do not rebuild spaces that teach rest, limits and forgiveness, our health systems will keep firefighting damage that never needed to happen.

Your annual check-up will continue to say “within normal range” right up until the moment it doesn’t. The scanners will keep getting better, the drips more expensive, and the basic pattern will stay the same.

In Atlas Cove Health, “base of health” is defined as much by patterns as by numbers:

How many nights a week do you sleep enough?
Do you have three people you could call at 3 a.m. if life fell apart?
Does your week contain any real rest, or just collapse?
Are you living in permanent “one strike and you’re out” mode with everyone around you?

Labs matter. Imaging matters. But they are second-order. If your life is structurally insane, you cannot biohack your way out.

Finding the plot again

At a fundamental level, I do think we have lost the plot as a society.

We’ve convinced ourselves that a healthy life is eating vegetables, going to the gym, not smoking, and performing well at work. That is not a healthy life. It is a performance protocol.

We no longer know what rest really means. Working hard, coming home, and scrolling on your phone is not a yin activity. It is not rest. It is not recovery. Your limbs are still, but your brain is in a negative, agitated spiral. That is not relaxing; it is brain rot.

I am angry with the Church for having damaged its own credibility so badly that it no longer functions as a guide for most people on how to live a sane, healthy life. I am equally wary of a culture that hands that guiding role to algorithms and influencers instead.

The fortunate few are still being taught something deeper at home: what it means to rest, to take responsibility, to forgive and be forgiven. Those parents exist at every income level. This is not about wealth. Some families still carry a living tradition; many do not. There is no judgment in that observation. It is simply the landscape.

Where we go from here is a choice.

We can continue to live inside a story in which health is a checklist and the only real commandment is “be productive”. Or we can start, deliberately, to rebuild the thick, boring, protective structures that any long, sane life requires: shared rest, deep relationships, and the courage to repair instead of constantly starting over.

Your lab report cannot tell you whether you are living a life that makes sense.

For that, we will need better institutions — or we will have to become them ourselves.

I hadn’t.

What I was really attacking was this: in most of healthcare, nobody is structurally paid to ask the only question that matters.

Does this person actually need what we’re selling – and will it have a net positive effect on their health?

Modern medicine is not failing for lack of data or technology.
It’s failing because ethics is optional in the business model.

This is an educational and strategic perspective, not personal medical advice.

What kind of market are we really in?

On the macro chart, health and wellness look like a dream.

• OECD countries spend around 9–10% of GDP on health and the trend is up, not down.

• The global wellness economy has crossed multiple trillions in value and is growing faster than global GDP, roughly twice the pace in recent years. It is now larger than many “sexy” sectors we obsess over, including parts of IT, sports and recreation, and close in scale to the green economy.

• Within that, anti-aging, aesthetic, and “longevity” segments are growing mid to high single digits annually, with IV therapy, hydration and drip clinics hitting high single to low double-digit CAGRs in many markets.

So from a distance, the story is:

• Big market.

• Strong growth.

• Emotionally inelastic demand (everyone ages, everyone fears decline).

If this were any other industry, the rational next question would be: where is growth mispriced?

In longevity, the mispricing is obvious once you look at evidence:

• We have very strong data that controlling blood pressure, improving cardiorespiratory fitness, keeping people moving, and running guideline-level screening programmes reduce cardiovascular events, disability, and premature death. The effect sizes are large and repeatable.

• We have much weaker or absent data that IV vitamin drips, many peptides, and a lot of “optimization” protocols do anything meaningful for otherwise healthy people beyond placebo and transient feelings – while they carry non-trivial risks and costs.

Yet capital flows the other way:

• Billions into aesthetics, drips, aggressive screening, and data-heavy experiences.

• Far less into unglamorous, behaviour-heavy prevention built to guideline evidence.

That’s not because investors are stupid. It’s because health, as a market, has the exact structure in which mispricing thrives:

• It’s a credence good: you can’t easily verify whether what you bought did anything.

• Information is asymmetric: the seller looks like they know more than you.

• Externalities are massive: the real benefits of prevention spill into the public system, employers, and families, not just the person paying.

In that environment, selling anything that sounds plausible is easy.
Selling only what actually works is expensive.

The longevity boom as a misaligned growth story

Compare longevity to other fast-growing sectors.

You’ve seen versions of this pattern before:

• Adtech: optimising clicks and time-on-site instead of long-term welfare.

• Subprime credit: optimising yield today instead of solvency tomorrow.

• Green finance: speculative carbon products growing faster than actual decarbonisation.

On a slide, the verticals look great. Underneath, the growth is partly built on exploiting information gaps and pushing risk into the future.

Longevity is drifting the same way:

• Chain IV clinics, wellness hotels, and “full-body scan” memberships are easy to understand, easy to franchise, and easy to sell.

• Prevention and evidence-led risk architecture are more complex to explain, slower to monetise, and harder to scale in a pitch deck.

From an economist’s point of view, the sector is over-investing in:

• High-margin, low-evidence private goods (drips, vanity diagnostics, annual “executive” packages),

and under-investing in:

• Lower-margin, high-evidence goods with huge externalities (blood pressure control, fitness, guideline prevention, sleep, mental health).

We’re pouring capital into the theatre, not the engine.

Subscribe now

Three structural failures (without the moralising)

I want to challenge the industry without just moralising. So let me put this in system terms.

1. Informed consent is a thin patch over asymmetric information

In theory, informed consent is how we square the information gap: we disclose risks and benefits, the person decides.

In reality, in many “premium” longevity settings, three things go wrong:

• Complexity: full-body MRIs, genomics, “omics” panels, peptide stacks. Even clinicians struggle to keep up with the evidence. Clients don’t stand a chance.

• Narrative demand: people paying €8–20k/year for a membership rarely want to hear, “You mostly need to sleep, walk, lift, and take a statin.” They want a complex story that feels proportional to the spend.

• Sales overlay: the consult is half clinical, half sales process.

We’ve imported DTC playbooks into clinical contexts. The real question in the room is not:

“Is this needed and net positive for you?”

It’s:

“How do we frame this so you feel good about buying it?”

The formalities are respected. The economic logic underneath is not.

2. Conflicts of interest live in the revenue mix

We already know from multiple health-econ studies that:

• When physicians own imaging equipment, imaging rates go up.

• Fee-for-service models tend to increase volume without necessarily improving outcomes.

We accept this as textbook principal–agent behaviour.

Longevity clinics have quietly built the same pattern:

• Margin sits in drips, injectables, niche diagnostics, and high-ticket packages.

• KPIs focus on revenue per square metre, device utilisation, and “conversion”.

• Staff are rewarded more for upgrades than for preventing unnecessary interventions.

I’m not pointing fingers from the outside. In an early Atlas Cove financial model, we had a line that read (in plain language):

“X% of new guests should leave day one with at least one IV or injection booked.”

On a spreadsheet, that looks rational. The fixed cost is high; the margin lives in certain services; guests “expect something”.

Read it as an economist and it’s exactly what you’d criticise in any other sector: a strong financial incentive to oversell a credence good.

Every time a clinician tells a guest, “You don’t need that,” they reduce contribution margin. Ethics becomes negative carry.

3. No one owns the decade

Most health businesses own a tiny slice of someone’s health trajectory:

• Hospitals: 30-day readmissions and acute complications.

• Longevity clinics: monthly revenue, annual membership renewal, NPS.

But the real value of strong-evidence interventions is long-horizon:

• Well-controlled blood pressure, properly managed lipids, sustained physical activity, appropriate cancer screening – these move hard endpoints over 5–20 years, not 5–20 days.

From a system perspective, that’s a horizon mismatch:

• We optimise for quarterly metrics that are easy to measure and sell.

• We neglect ten-year risk that is hard to attribute and hard to monetise.

This is how you end up with:

• Double-digit growth in aggressive screening and “total body” check-ups,

• While epidemiologists warn that a lot of general population screening, outside defined risk groups, leads to overdiagnosis and unnecessary interventions, not improved mortality.

We are de facto paying businesses to create data and anxiety, not to reduce strokes and amputations.

Health is one of the few products you can feel

Here’s the part that doesn’t get discussed enough in investment memos.

Most products can hide behind perception and brand. Health can’t, not indefinitely.

• If your sleep improves, you notice.

• If your blood pressure is controlled and your fitness goes up, you feel it.

• If your baseline anxiety about your body drops because things actually are better, your behaviour changes.

That makes health unusual: it’s a multi-trillion market where the “utility function” is literally embodied.

A business that truly delivers:

• more energy,

• fewer symptoms,

• better function,

• and lower risk,

should have stickier economics than a business that delivers only experiences and dashboards.

In theory, an ethics-first, evidence-first longevity model should behave like a great B2B SaaS product:

• Retention is high because churn feels bad.

• Expansion is organic because satisfied clients bring spouses, friends, and colleagues.

• Acquisition costs trend down as reputation compounds.

In practice, the industry behaves like fashion retail:

• constant new protocols,

• constant new devices,

• constant movement between providers.

Churn is reframed as “people trying different things”.

That’s the opportunity: to build the first layer of longevity that monetises felt, durable health, not novelty.

Atlas Cove: ethics as a design constraint, not a slogan

Atlas Cove exists because I didn’t want to criticise the market while quietly reproducing it in nicer stone.

We started exactly where everyone starts:

• advanced imaging on the whiteboard,

• complex panels,

• an exciting list of frontier tools.

Then we put the stack through three filters:

1. Evidence tier

   • Strong human outcome data?

   • Early but promising human data?

   • Mechanistic/animal only?

   • Speculation?

2. Net benefit by profile

   • For whom, exactly, does this materially change long-term risk or function?

   • At what stage?

3. Sequence

   • Does this belong before or only after we’ve done the boring work: sleep, blood pressure, lipids, weight where appropriate, fitness, mental health, and guideline screening?

A lot of pretty lines died on that whiteboard.

What survived for Atlas Cove is essentially the Atlas Labs stack in real life:

1. Foundations (non-negotiable)
Every member goes through a structured foundations phase:

• movement and strength,

• VO₂max-oriented cardio in practical form,

• sleep and circadian hygiene,

• nutrition and metabolic basics,

• mental load, relationships, and work patterns.

These levers have the largest and clearest impact on morbidity and mortality in the literature. Skipping them to jump into frontier tools makes no sense if you care about outcomes rather than optics.

2. Medical & prevention layer (evidence-aligned)

On top of that:

• basic labs through a quality partner lab: lipids, glucose/HbA1c, organ function, sometimes inflammatory markers;

• vitals and functional metrics: blood pressure, resting HR, strength, simple cardio tests;

• guideline-aligned decisions on medication and screening, with specialist input where needed.

Interpretation is done by prevention-oriented physicians, under clear regulation, with explicit evidence tiers.

3. Frontier layer (tightly gated)

Only after 1 and 2 are mapped and under active management do we even consider:

• advanced imaging,

• targeted pharmacology,

• or more experimental protocols.

Each of those is:

• tagged by evidence level,

• anchored to specific profiles,

• and offered only if it has a plausible net positive effect beyond doing more of the basics.

The operative rule we wrote into the operating model is blunt:

Nobody gets access to the shiny stuff until we’ve done the work on the boring stuff.

This is not about being paternalistic. It’s about aligning the business with what an honest long-horizon risk model would recommend.

It also changes the economics:

• We cannot hit our numbers by pushing drips on day one to exhausted, hypertensive executives.

• We must make prevention design, interpretation, and behaviour architecture themselves into billable, high-value products.

• Retention becomes a function of how people feel and function six months in, not how impressed they are on day one.

Ethics, in this model, isn’t a brand value. It’s a constraint that shapes the P&L.

Join the Waitlist

Ethical debt: the line item nobody models

As an operator or investor, you’re used to thinking about:

• leverage,

• regulatory risk,

• reputational risk.

I think longevity needs one more category: ethical debt.

Ethical debt is the cumulative gap between:

• what you know would be an ethically defensible, net-positive product set and sequence, and

• what you actually sell because it’s profitable and the market is currently willing.

You accumulate ethical debt when you:

• add a “performance drip” primarily because it photographs well,

• keep a rainmaker clinician whose numbers depend on overselling,

• push general health check-ups or scans outside evidence-based indications,

• treat prevention as a thin layer of marketing rather than the core engine.

Like technical debt, ethical debt:

• makes later course correction expensive,

• compounds quietly,

• eventually crystallises as scandal, regulatory action, or talent flight.

Unlike technical debt, it plays out in human bodies.

From a seasoned economist’s perspective, ethical debt is a mispriced liability in a sector where the scarcest assets are trust and long-horizon outcomes. Running high ethical leverage in a credence market is not just morally dubious; it’s financially naive.

A simple filter for serious capital

So here’s the line I now use on myself, and I’d invite you to use it on any longevity asset:

Would this business still be viable if every clinician told every client the whole truth about what they actually need – including “you don’t need most of this”?

If the honest answer is no:

• you don’t own a healthcare company;

• you own an extraction scheme sitting on top of information asymmetry.

That can make money. Plenty of things do.

But in a world where:

• wellness is already a multi-trillion sector,

• prevention has clear, measurable ROI,

• health is a product people literally feel in their bodies,

the compounding opportunity is elsewhere:

In the layer that allocates this growing river of spend into interventions that actually extend healthy, capable years.

Atlas Cove is my attempt to build that layer in one geography.

Not as an ethics performance.
But as the only version of the longevity business that makes sense if I treat my own time and capital the way I treat health: as something to allocate carefully, over decades, with reality – not narrative – as the benchmark.

There would be headlines. Guidelines. Funding.

But when loneliness, social isolation, and relentless mental load sit in a similar risk band, we call it “just life.” And we do something worse than ignore it: we moralise it.

Loneliness is often framed as a personality weakness — as if people are lonely because they’re awkward, needy, “too sensitive,” not trying hard enough, not social enough. That framing is not only cruel. It’s medically wrong.

Loneliness is not a character defect. It’s a biological signal. A risk factor. A load on the system.

And this is where hospital-based western medicine, as it’s practiced across much of Europe today, keeps getting it wrong — not because clinicians are incompetent, but because the system is built for a different job.

Hospitals are optimised for:

• acute care

• disease detection

• crisis management

• protocol-driven intervention

They are not optimised for:

• measuring slow erosion

• identifying negative compounding

• treating the lived drivers that push people toward collapse

So people get the same verdict again and again:

Your bloods are “fine.”
Your scans are “fine.”
And you’re still exhausted, wired, inflamed, and one minor crisis away from falling apart.

Large meta-analyses suggest that loneliness and social isolation are associated with meaningfully increased all-cause mortality risk — on the same order as risk factors we already take seriously, including smoking and physical inactivity. Not because loneliness equals a literal number of cigarettes, but because the effect sizes sit in the same league.

On paper, we are healthier and wealthier than ever. In lived reality, many people are in quiet overdraft.

Their lifespan might look acceptable.
Their healthspan — and what I’d call their joyspan — is collapsing.

The dirty secret is simple: we optimise what we can easily measure, and almost no one is measuring the thing you feel every hour of every day.

Your energy.

Health as an energy bank account

I think about health as a bank account for energy.

You can be low, barely covering expenses.
You can be “comfortable,” getting through the month with nothing left for surprises.
Or you can be flush — with surplus to invest, take risks, stretch without breaking.

Every day, you either deposit into that account or you withdraw from it.

Deposits (things that refill and compound energy):

• Sleep quality + consistency (timing, depth, enough total)

• Deep rest (true downtime, parasympathetic time, low stimulation)

• Real connection (safe relationships, being seen, touch, belonging)

• Movement that builds capacity (strength training, Zone 2, daily walking, occasional intensity)

• Nutrient-dense food (protein, fiber, micronutrients; stable blood sugar; hydration)

• Light + circadian anchors (morning daylight, darkness at night, regular timing)

• Joy signals (laughter, play, music, nature, beauty, sex when it’s regulating)

• Meaning + agency (work you can influence, clear priorities, autonomy)

• Emotional processing (therapy, journaling, breathwork, somatic work)

• Recovery infrastructure (bodywork, restorative weekends, recovery that actually restores)

Withdrawals (things that drain and destabilise energy):

• Loneliness and social disconnection (especially when paired with shame)

• Chronic stress load (high responsibility without recovery; constant urgency)

• Sleep disruption (fragmentation, late nights, alcohol, rumination, screens)

• Ultra-processed food + glycaemic volatility

• Unresolved conflict and relational threat

• Constant stimulation (phone, news, multitasking, dopamine loops)

• Overtraining / under-recovering

• Sedentary stagnation

• Substances used as coping (alcohol to downshift, caffeine to survive)

• Rumination and self-attack

• Environmental drag (noise, clutter, poor air, lack of nature, no privacy)

And then there’s the factor most people miss entirely.

The account has a growth rate

When your internal energy capital is healthy, deposits compound.
When it’s depleted, even “doing everything right” barely moves the needle.

That’s what “wired and tired” is: not laziness, not weakness — a system stuck in negative compounding.

Longevity, in my view, is not about chasing maximal biomarkers. It’s about keeping your energy account in positive growth territory for as many years as possible — protecting and compounding the underlying asset: mitochondrial, psychological, and relational energy.

This is an educational and strategic perspective, not personal medical advice.

Why hospitals miss this (even when they mean well)

Hospital medicine is designed to answer: Do you have a diagnosable disease?
It’s not designed to answer: Are you quietly failing?

If your energy account is deteriorating, it often won’t show up in the places the system looks first:

• standard labs can be “normal” while your growth rate is collapsing

• imaging can be “clear” while your nervous system is in chronic threat mode

• symptoms can be real while tests fail to explain them neatly

So the patient gets minimised. Or bounced between specialties. Or offered another test. Or told to reduce stress as if that’s an instruction, not a treatment plan.

The system doesn’t lack intelligence. It lacks a model for energy erosion — and a method for rebuilding it.

Subscribe now

How my own account nearly went bankrupt

In my mid-twenties, a lot happened at once.

I didn’t just feel “a bit off.” I collapsed into a level of unwellness that made it obvious something wasn’t right. And I had a clear internal sense: whatever this was, it wasn’t going to be solved by chasing one isolated symptom.

So I did two things in parallel.

I went deep into the medical system — GPs, specialists, scans, bloods — trying to understand why I kept getting sick and never stabilised.

At the same time, I started therapy. Not as an add-on, but because I could feel my internal stress system was fundamentally dysregulated. Eventually, that thread had a name: PTSD. I committed to treatment over several years. I won’t go into details here — not because it wasn’t real, but because the point of this article isn’t my biography. It’s what I learned about the full system.

Around the same period, I received more physical diagnoses: Hashimoto’s and PCOS (and at one point suspected endometriosis).

It’s a chicken-and-egg problem, and I’m careful with simple narratives. But looking back, the pattern is hard to ignore. Across childhood, adolescence, and early adulthood, I had been withdrawing more energy than I could replenish for too long.

Hashimoto’s lives in your genes — but activation and flares are shaped by stress biology and immune regulation. PCOS is multifactorial — metabolic, inflammatory, endocrine — and stress is not a footnote.

That’s when the energy bank account stopped being a metaphor and became a practical model.

The hardest moment was a short appointment with a specialist who looked at my imaging and quietly told me my chances of conceiving were very low.

I had always assumed I would be a mother. I walked out of that room feeling like someone had wiped the future.

After a brief free fall, I did what I know how to do: I turned the problem into a research project.

I read everything I could on fertility, endocrinology, inflammation, circadian biology. Traditional Chinese Medicine kept appearing in the footnotes.

That’s how I ended up at an Amsterdam TCM clinic.

After taking my history, the practitioner said something deceptively simple:

“We need to send energy to your female organs.”

I changed my sleep, my training, my food. I stayed in therapy. I cleaned up my relationships. I rebuilt the whole system — mental and physical — at the same time.

Within a year, my scans no longer showed PCOS. I became pregnant on the first try. My son arrived healthy. Today he has the kind of endurance strangers comment on.

By the end of my twenties, the real shift wasn’t “fixing” a diagnosis. It was knowing — viscerally — what surplus energy feels like, and what actually creates it.

That experience is the foundation of everything I’m building now.

Mitochondria: the asset behind the account

Strip away metaphors and you end up with ATP.

Every thought, contraction, repair process, and immune response runs on ATP. Mitochondria generate it — and coordinate signalling around inflammation, gene expression, and cellular ageing.

Think of mitochondria as the asset behind your energy balance.

When mitochondrial function is robust, sleep, movement, and connection have a high return. When it’s impaired, the same inputs produce less usable energy and more noise — unresolved stress signalling, inflammation that doesn’t downshift, fatigue that doesn’t lift.

If you feel “wired and tired,” endlessly optimising habits without changing the slope of your week, that’s what an overdrawn account can look like at the cellular level.

The mind written into biology

This is where loneliness stops being a “soft” topic.

Stress, trauma, meaning, and social connection don’t just affect mood. Over time, they shape physiology — via brain, endocrine, immune, and cellular energy systems.

In plain language:
“I’m under threat.”
“I’m alone.”
“I can’t rest.”

These aren’t only thoughts. Repeated long enough, they become biological instructions.

So withdrawals aren’t only sleep and food. They’re also years of unresolved stress, relational threat, constant vigilance, and disconnection.

Deposits aren’t only training and supplements. They are safety, agency, purpose, and people you can lean on.

This is why I reject the “personality weakness” framing of loneliness. It weaponises shame against the very state that most needs repair.

If we treated energy, mental load, and relationships like cholesterol, many “healthy” people would suddenly look high-risk.

Energy as an allocation problem

Most health advice is vague: do more good things, avoid bad things.

The energy bank lens forces sharper questions:

• What is your balance right now?

• What is your growth rate?

• Where are the biggest leaks?

For many high-responsibility people, more supplements or diagnostics are low-return moves. The account is already overdrawn. The growth rate is negative.

The work is to stabilise the balance, then change the slope.

Longevity becomes an allocation problem, not a gadget hunt.

How this becomes Atlas Cove Health

This is the lens I use for health and longevity. The full picture, for me, is the energy bank account: your balance today, your daily deposits and withdrawals, and — most importantly — your growth rate.

And it’s exactly how we work at Atlas Cove Health.

We start by understanding the full bank account of each guest: what’s fuelling them, what’s draining them, and why their energy is or isn’t compounding. We make the invisible visible — sleep, stress load, metabolic volatility, fitness capacity, relationship strain, sense of safety and belonging — because these are not “soft factors.” They are inputs into biology.

Then we translate that into action:

• identify the active deposits that rebuild energy capital

• cut the active withdrawals that are causing leaks and negative compounding

• shift the growth rate back into positive territory

But we don’t do this in a vacuum. We ask a sharper question up front:

Who do you want to be at 80?

Not “avoid disease.” Actual capability.

Do you want to hike without thinking about it? Travel with ease? Dance late? Lift your grandchild? Keep your mind sharp? Maintain deep relationships?

That future isn’t determined at 80. It’s determined by whether your energy account is compounding or quietly collapsing today.

So the work becomes practical: what has to change now — in workload, sleep, training, nutrition, and relationships — to fix the growth rate, rebuild surplus, and protect the capacity to do the fun things you still want to do when you’re 80.

Atlas Cove is an intimate health retreat in Sesimbra built on exactly this: measure the account, then change the slope.

It’s a corrective to a system that waits for disease — and misses the slow collapse happening in plain sight.

The question I leave you with

Where is your energy balance — low, barely adequate, or flush?
What is your growth rate?
And what are your real deposits and disguised withdrawals this week?

If we stop treating loneliness like weakness and start treating it like risk, the entire map of health changes.

And a lot of people with “fine labs” finally get seen accurately.

No labs. No drips. No devices.
Just a decision to feel more joy in the life you already have.

The other day I came across a Joyspan experiment in San Francisco: a “longevity rave” where participants went substance-free, logged their mood, then danced wired up to wearables while researchers took blood and saliva before and after. They were looking at brain waves, heart rhythms and biomarkers related to inflammation and mitochondrial stress, trying to compress all of that into a “JoyScore” – a data trace of joy and connection in the body.

It sounds like peak Silicon Valley, but the underlying instinct is right: joy is not just a feeling. It has a biological footprint.

Over the last few years, several groups have shown that daily mood and psychosocial experiences are linked to mitochondrial health markers. Higher positive mood has been associated with a higher “mitochondrial health index” in blood cells, and some of the impact of chronic stress on mitochondria appears to be mediated by lower daily positive affect. Newer work suggests that people with more positive psychosocial experiences show greater abundance and efficiency of mitochondrial energy machinery, whereas chronic psychological stress and distress are tied to mitochondrial dysfunction and higher disease risk.

All of this is still relatively early and largely associational. No one can honestly say “increase joy by X and your mitochondria improve by Y.” If every mitochondrial paper vanished tomorrow, the business case for joyspan would still stand on behaviour, connection and retention. But taken together, the data make one thing clear: chronic stress and negative affect damage the energy system; better mood and wellbeing are associated with more robust mitochondrial profiles. Joy is not only “in your head”. It shows up in how your cells handle energy.

If you’re building a longevity business, that matters. Your clients’ mitochondria are where lifespan, healthspan and joyspan all meet.

From that angle, joyspan isn’t a soft idea. It’s one of the most financially rational things you can build into a health business.

Since becoming a parent, this stopped being abstract. There are only so many mornings, school runs, bedtimes. My son won’t copy what I say about health; he’ll absorb the emotional baseline I live in. At the same time, I’m a founder, building a medically led clinic and diagnostics capability in Portugal. Most of my day is protocols, equipment, regulation, unit economics.

Joyspan sits at the junction of those two roles. It is both a personal operating principle and, in my view, a hard commercial lever.

Subscribe now

Joyspan, in operator language

We already design for two spans:

• Lifespan – how long you are alive.

• Healthspan – how long you can function as you wish.

There’s a third span almost no one designs around:

• Joyspan – the share of your functional years that actually feel worth being alive for.

Think of joyspan as the proportion of “healthy” years where you have real reasons to get out of bed - laughter, meaning, people you’re glad to see - instead of just more time to manage admin and anxiety.

We have enough evidence to treat this as more than poetry:

• higher subjective well-being is associated with lower mortality and fewer cardiovascular events

• stronger social connection is associated with longer life and better health

• loneliness and isolation carry a substantially elevated risk of premature death, on a par with major lifestyle risks

On top of that, early human work shows that daily mood and psychosocial experiences are associated with mitochondrial health and cellular bioenergetics.

Most of this is associative, not perfectly causal. That’s fine. If you’re designing a business, it’s irrational to ignore a factor that:

• tracks with hard outcomes

• shapes daily behaviour

• influences the cellular machinery you claim to care about

• heavily affects whether people stay in your product or quietly leave

Joyspan doesn’t replace the hard risk-factor work (sleep, blood pressure, metabolic health, depression treatment). It multiplies the odds that people will actually do that work and keep doing it. If risk-factor work is your principal, joyspan is the interest rate on adherence.

Why joyspan beats “one more biomarker”

Joyspan shows up in three places that directly affect your P&L: behaviour, biology and experience.

1. Behaviour → adherence → outcomes

Clients who feel better and more connected are more likely to:

• move regularly

• avoid heavy smoking and drinking

• stick to treatment plans

• show up to appointments

Over years, that behaviour change drives most of the benefit any clinic can claim. If someone is chronically joy-poor, all the “healthy routines” you prescribe sit on brittle ground. They comply in bursts, then drop off. That turns into:

• disappointing long-term outcomes

• flat NPS

• weak renewal curves

2. Biology → amplifying what you already do

Chronic emotional states show up in:

• stress hormones

• inflammatory markers

• heart-rate variability

• sleep architecture

• and, increasingly, mitochondrial structure and function

A system running on constant rush and low-grade fear stays in threat mode and builds “mitochondrial allostatic load” – wear and tear on the energy system. A system that regularly experiences safety, connection and gratitude spends more time in repair and in efficient energy production.

You do not need a perfect causal model to see that this interacts with every drug, protocol and device you sell.

3. Experience → retention and LTV

Clients do not experience your product as “protocol + labs + app”. They experience their week:

• waking up

• going to work

• parenting

• going to sleep

If their numbers improve but their week feels emotionally unchanged - or more stressful because of added tasks and fear - they leave when a cheaper or more convenient option appears.

If their week feels noticeably better after engaging with you - calmer, more energised, more connected - they think twice before cancelling. That is the commercial core of joyspan: it multiplies adherence, amplifies biological effects and turns your offer into something people don’t want to lose.

What I learnt in places I’d never build

I’ve done my share of retreats—the kind serious people roll their eyes at. I wouldn’t build a company in that container. But I did pick up something useful: how to notice what I actually feel, how to self-regulate instead of outsourcing that job, how to sit with difficult emotions without instantly numbing, and how to feel genuine gratitude for being alive.

Those skills have shifted my day-to-day more than any device, drip or lab panel.

The content—emotional regulation, presence, gratitude—is not the problem. The packaging is. Most of your clients do not want their emotional operating system from a pseudo-spiritual circus.

That’s the opportunity:

• take the parts that work

• strip out the theatre

• deliver them in a clinical, grounded, repeatable way

And do it in group formats that are:

• cheap to deliver

• easy to standardise and improve

• clearly felt as valuable by clients

Operationally and financially, that’s attractive. Anyone can bolt a wellness ritual onto a lab. Designing around joyspan means you reroute intake, plans and follow-up so that a flat joy curve is treated like any other red flag.

Saturdays sell, Tuesdays retain

Back to San Francisco.

In the JoyScore experiment, a small group went clean for several days (no substances, no alcohol, limited caffeine), logged mood and energy, then spent a structured night in a Market Street tower: finger-prick blood draws for stress and inflammation markers, saliva for cortisol and oxytocin, an ear-clip EEG, then three hours of carefully engineered techno tracks while researchers watched their physiology dance.

It was more science experiment than the rave of anyone’s life. But the aim was clear: can you see joy and connection in the data, including biomarkers linked to mitochondrial stress and recovery?

Most of the wellness market optimises peaks like that: retreats, drips, spectacular moments. They’re easy to market and photograph. A drip bar can sell you a Saturday high.

But retention is decided by Tuesdays:

• Did my normal week feel better?

• Am I less frazzled?

• Do I have more real moments of joy?

If your service doesn’t move how Tuesdays feel, your peaks won’t save you from churn. A serious longevity clinic has to make Tuesday mornings better. Only that builds LTV. Joyspan is Tuesday infrastructure.

Joyspan as an operating constraint

You can compress a lot of the mechanism into four rules:

• Presence kills rushing.

• Possibility kills fear.

• Gratitude kills stress.

• Signal kills noise.

These are more than nice lines. They’re design constraints.

If your clinic or programme systematically creates:

• rush (too many touchpoints, too much friction)

• fear (overwhelming risk framing, scary dashboards)

• stress (complex routines, constant alerts)

• noise (too many metrics, no clear priorities)

…you are burning joyspan. Burned joyspan shows up later as churn and low referrals.

We now have early evidence that the stories in your head—your ongoing stress, your sense of safety, your daily mood—are associated with how your mitochondria behave and how your energy system copes with load. Thought patterns and emotional states change how you spend and replenish energy, day by day.

From a business perspective, joy behaves like a broad-spectrum, upstream intervention:

• it changes daily decisions

• shifts stress physiology and inflammation

• improves sleep and social connection

I call it a “multi-target drug” metaphorically: not because it’s a regulated product, but because its footprint across systems is broader than most single interventions. Ignoring that is like ignoring sleep.

Subscribe now

Where joyspan lives in a health business

Joyspan belongs at the bottom of the stack next to sleep and stress, not at the top as a spa extra.

Practically, that means building it into three layers:

1. Assessment

You don’t stop at labs and step count. You add something as simple as:

• “In the last 7 days, how many separate moments felt genuinely joyful?”

• “What were two of them?”

You track that trend. If joyspan is consistently near zero, you treat it as a core risk signal - alongside blood pressure or HbA1c - not a side note.

Internally, you assume:

“If joyspan hasn’t moved in the first 8–12 weeks, long-term retention is at risk.”

That makes joyspan a product KPI, not a decoration. A flat joyspan curve is handled like any other worrying trend: it triggers action.

2. Programme design

Every plan has two parts:

• the risk-factor work (sleep, metabolic, musculoskeletal, mental health)

• at least one recurring element the client genuinely looks forward to

In reviews, you talk about LDL and blood pressure and you ask:

• “Did you have more, fewer or the same number of joyful moments last week?”

• “What changed when that number moved?”

If everything you prescribe feels like work, adherence quietly dies. When adherence dies, outcomes and renewals follow.

3. Follow-up and escalation

If biomarkers improve but:

• joyspan is flat or falling

• the client is more anxious or isolated

• the routines feel like a burden

…you treat that as a failed intervention. You simplify, adjust, or—where appropriate—refer into formal mental health care. You don’t call that success just because the lab is green.

Alongside this, you bake in group-based emotional skills and connection sessions:

• clearly positioned within EU and Portuguese regulatory boundaries (wellbeing / coaching vs medical where appropriate)

• designed or overseen with clinical input

• measured against changes in perceived stress, joyspan and engagement

They’re low marginal cost, emotionally high-impact, and tightly linked to “how my life feels”, which is what clients actually pay to change.

The financial logic

From an investor or operator lens, joyspan is financially sane because it improves three things:

1. Retention

People keep paying for things that:

• clearly improve how their life feels

• fit into their routines

• feel emotionally safe and supportive

If your product moves joyspan in the right direction within weeks, people are far more likely to:

• stay in your membership

• accept price increases

• move up tiers when it makes sense

If it doesn’t, they will rationalise churn: “too busy”, “too expensive”, “I’ll pause for now.”

We plan to test the hypothesis that clients whose joyspan improves in the first quarter have meaningfully better 12-month retention and referral than those whose joyspan is flat. If that proves true, joyspan becomes not just a philosophy but a management tool.

2. Referral

Most genuine word-of-mouth distils to:

“I just feel better on this.”

You can’t fake that with branding. You earn it with a persistent shift in how someone’s week feels. Joyspan is the closest thing to that reality you can track.

Higher joyspan → higher odds they talk about you at dinner, in founder groups, in WhatsApp chats. That lowers effective CAC over time.

3. Return on complexity

Every new device, protocol or feature adds operational complexity. The only good reason to add complexity is if it clearly improves:

• outcomes, or

• the emotional experience of being your client

Joyspan-focused features do both:

• better emotional state → better behaviour → better outcomes

• better emotional state → more enjoyable engagement with your product

That’s a much cleaner justification than “this looks cool in the brochure”.

A quick operator tool: the three-span ledger

One slide, three rows:

• Lifespan – years alive

• Healthspan – years functional

• Joyspan – years that feel worth being alive

Columns:

• “What our product is doing for this now”

• “What we will change in the next 12 months”

Fill it in honestly.

Most health businesses have plenty to say under lifespan and healthspan—tests, scans, programmes—and almost nothing concrete under joyspan.

That blank row is not just an ethical gap. It’s a commercial risk. It’s where churn hides and where LTV flattens.

For me, this isn’t abstract. It’s anchored in two realities: a son who watches how I live every day, and a clinic and lab I’m quietly building in Portugal.

If I extend someone’s lifespan and healthspan but their joyspan stays flat, I’ve given them a longer corridor, not a better life. If I build a business that optimises biomarkers but leaves people emotionally unchanged, I’ve built something fragile and expensive.

Adding joyspan to a health business is not softness. It’s discipline:

• it improves adherence and outcomes

• it strengthens retention and referral

• it makes complexity earn its keep

Joyspan is not a bonus. It’s the yield you are actually alive - and actually in business - to collect.

This is an educational and strategic perspective, not personal medical advice.

That article triggered a lot of input. Clinicians wrote about what actually happens in 10-minute primary-care visits. Operators asked where their projects fit. Founders asked if this “missing middle” was just a nice story, or a category you could actually build.

The same questions kept coming up:

• What exactly do you mean by Tier-2?

• Isn’t concierge already doing this?

• If demand is real, why doesn’t the market already look like this?

This piece answers those three questions and explains why I walked away from building a Tier-1 luxury clinic and chose to build Tier-2 instead.

1. The four tiers: 0, 1, 2, 3

First, a clean map. I think about health for working adults in four layers, always in order: 0, 1, 2, 3.

Tier-0: noise and DIY

Where most health-literate adults actually live.

Podcasts. Twitter threads. YouTube protocols. A supplement drawer that looks like a small pharmacy. Wearables buzzing your wrist. Panic blood panels “to check everything”.

On their own, most of these tools are fine. In theory they could be Tier-2 inputs. In practice, the data lives on apps and anxiety, not in a structured medical plan.

Tier-1: premium access and theatre

Concierge practices. “Membership” primary care. Executive check-ups. Luxury “longevity” clinics.

You pay for smaller patient panels, longer visits, easier access, nicer surroundings. There are excellent clinicians in Tier-1. But structurally, most offers still behave like upgraded GP: same architecture, more time and tests.

The product is access, not prevention as a system.

Tier-2: structured prevention and diagnostics

This is the missing layer.

One-liner: Tier-2 is serious, evidence-aligned prevention and early detection for high-functioning adults, outside hospitals, built as a repeatable programme instead of a one-off event.

Tier-2 is not an inspirational retreat or a PDF with 60 lifestyle tips. It’s an operating system and programme for staying out of Tier-3:

• structured pre-work

• deep in-person diagnostics

• a defined protocol, not a random assortment of tests

• months of follow-through that fit around real work and family life

Think: the layer that turns “I’ll get around to it” into “this is my annual health strategy”.

Tier-3: acute and complex care

Hospitals, emergency departments, oncology, cardiology, ICUs.

The layer that tries to stop you dying or losing major function once something serious is already happening.

If you treat your health like capital, Tier-3 is the only layer that truly has to exist. You don’t want to live in Tier-0 noise. You don’t want to become a full-time Tier-1 tourist. And you’d like to meet Tier-3 as late and as rarely as possible.

Tier-2 is the layer that makes that more likely.

Subscribe now

2. The prevention story vs what actually happens

On paper, every advanced health system is “shifting to prevention”. There are national plans for cancer and cardiovascular disease. The language is always the same: early detection, risk stratification, lifestyle change, health checks.

In real life, almost all of this is dropped on one layer: primary care.

In Portugal, over a million people registered in the national system don’t have a family doctor assigned at all. Those who do often share that doctor with thousands of others. Consultations are short, lists are long, and a lot of time is eaten by forms, renewals, and acute problems.

This isn’t a criticism of clinicians. It’s geometry. You can’t run deep, personalised prevention through ten-minute slots, overflowing lists, and reimbursement designed for acute care.

Meanwhile, if you follow what people do with their own money, you see something else:

• private diagnostics and “health screens” are expanding

• concierge and “direct” care are normalising at the top end

• health and wellness-motivated travel is a standard category, not a curiosity

On the slide deck, prevention is covered. In real life, serious adults are often opting out and improvising. That gap between the prevention story and the behaviour is where Tier-2 belongs.

3. What I actually mean by Tier-2

Here is the working definition I use:

Tier-2 is a structured journey that maps your cardiometabolic risk, functional capacity, sleep, mental load, and relevant cancer risk, and then runs you through a defined programme with enough depth to change risk and enough discipline to be safe.

In practice, that means:

• Mapping what truly matters
  Cardio-metabolic risk, functional capacity, sleep, mental health, and specific cancer risks based on age, sex, and family history – using tests that have real human evidence behind them.

• Running a designed process
  Not “whatever diagnostics we can bill” and not a random buffet of scans. A protocol that fits your constraints: flights, kids, deadlines, not a fantasy retreat version of you.

• Owning follow-through
  Clear hand-offs. Structured coaching and habit change where it actually matters. Knowing who acts on the data, what happens when something concerning shows up, and how escalation works.

Just as important is what Tier-2 is not.

• not a spa with a lab invoice

• not a mini-hospital managing serious disease in a boutique wrapper

• not “GP plus slightly longer slots”

• not “longevity content” plus a supplement upsell

The people I’m designing this for have three simple filters:

1. Don’t waste my time.

2. Don’t talk nonsense.

3. Don’t pretend I live like a monk or a pro athlete.

Tier-2 is being built to clear that bar by design, not by accident.

4. Why I walked away from Tier-1 fantasy

I didn’t start out wanting to build Tier-2. I started obsessed with Tier-1.

I had moodboards of Swiss and Austrian clinics. For a while, my plan was simple: build that in Portugal. Beautiful building, frontier tools, Michelin-level food, the whole script.

Three things changed that.

4.1 The business model didn’t scale the way I wanted

On paper, a single, hyper-premium Tier-1 clinic in Portugal can work.

But when I modelled that same clinic across cities and countries, it looked slow and brittle. You end up with a small fleet of very expensive boxes that only a tiny group of people can afford.

My goal isn’t to own “the nicest clinic in X”. It’s to shift the health trajectory of a meaningful number of working adults. You don’t do that with one marble box and a waitlist.

Once I admitted that, Tier-1 started to look like a brand asset, not system infrastructure.

4.2 The “frontier” tools didn’t have the spine I needed

Longevity’s “wow” elements are seductive: IV drips, exotic compounds, stacks of supplements, experimental protocols.

Some are genuinely useful in specific contexts. Some are promising. A lot is still mechanistic reasoning, small uncontrolled studies, or marketing dressed up as evidence.

I ran a simple filter: strip out anything I couldn’t defend in front of a tough clinician or a serious investor.

What remained looked a lot like Tier-2: diagnostics, structured programmes, behaviour, environment, follow-through. Less magic. More engineering.

That’s a spine I can build on.

4.3 I couldn’t see myself as the repeat customer

I’ve worked hard, built a career, and earned well. I’m also a mother and not in the ultra-high-net-worth group.

I can’t justify dropping €20k every year or two on a reset week as my core health strategy, even if I could stretch to it occasionally. It doesn’t fit how I think about risk, money, or fairness.

I don’t want to build something I wouldn’t buy regularly, with a straight face. I want to build something someone like me can access as a rational line item, not a one-off splurge.

Taken together, these three filters pushed me in one direction:

• a model that can scale across locations and cohorts

• a core product built on decent (or better) evidence

• a price point and format that people like me would use repeatedly

When you apply those filters, you land in Tier-2: scalable, repeatable, less flashy — and exactly what’s missing.

5. Isn’t concierge already Tier-2?

Short answer: it’s close, but it mostly stops halfway.

Concierge and membership models usually offer:

• smaller patient panels

• longer visits

• easier access

Direct primary care removes some billing distortions. Executive programmes bundle tests and consultations into “deep dives”.

These are real improvements on standard Tier-3 and Tier-1. They prove there is willingness to pay for better access and more attention.

But structurally, most of what I see looks like:

• Access and time as the main product, not prevention as the organising principle.

• Opportunistic prevention (“while you’re here, let’s also check…”) instead of prevention as a defined, repeatable programme.

• Test menus that mix guideline-based screening with not-recommended scans and panels, increasing noise and overdiagnosis risk.

• Limited outcome data: better uptake of preventive services and higher satisfaction, but not strong, consistent signals of fewer strokes, fewer late-stage cancers, or more functional years.

So yes, there are Tier-2-shaped fragments in the wild. They validate parts of the demand.

But “expensive GP plus extras” is not the same as a Tier-2 system designed from day one to move risk at scale.

That is the gap I’m targeting.

6. Why the market hasn’t built real Tier-2 yet

If Tier-2 sounds so obvious, why doesn’t it already show up as a clear category?

Because you can’t just decorate Tier-1 with prevention language and expect it to survive contact with reality.

A serious Tier-2 build has to clear four constraints at once:

6.1 Regulatory altitude

Regulators care what you do, not what you call yourself.

Bring certain diagnostics or therapies in-house and, on paper, you start to look like a hospital. That triggers hospital-grade facility, staffing, and oversight requirements.

That’s appropriate for Tier-3. It’s lethal for a small Tier-2 clinic that didn’t design for it.

So Tier-2 has to pick an altitude on purpose: define what never happens under your roof, what only happens via formal partners, and exactly where your responsibility stops.

6.2 Reimbursement

Tier-2’s value is in integration and programmes. That doesn’t map cleanly onto most fee schedules.

Parts may be billable. The whole product usually isn’t.

That’s why I’m treating Tier-2, at least initially, as cash-pay aimed at people who already treat health as a deliberate budget line, not as an afterthought.

If Tier-2 proves its value, other funding models can follow. But you can’t rely on reimbursement to carry the model from day one.

6.3 Clinician time

Physician time is the binding constraint in most systems.

If your model implicitly assumes unlimited doctor hours, it will fail. Doctors have to be reserved for decisions: risk classification, interpretation, escalation, de-escalation.

Everything else — data gathering, education, habit change, monitoring — has to be handled by other humans and systems.

Tier-2 only works if you design around this from the start.

6.4 Liability and pathways

More testing means more findings, and more responsibility.

If you don’t know exactly what you’ll do with the results, you’re lighting a fuse.

Tier-2 has to be conservative about which tests it uses, why, and who owns the next step if something serious appears. That’s less marketable than “full body scanning”, but it’s also where the moat is: a disciplined, defensible testing and escalation strategy is hard to copy quickly and hard to fake.

These constraints aren’t reasons not to build Tier-2.

They’re exactly why most glossy concepts haven’t survived, and exactly where the long-term defensibility lives.

7. Who I’m actually designing Tier-2 for

Once you respect the friction stack, “Tier-2 for everyone” disappears. You’re left with a smaller, sharper, but very real segment.

For me, there are two engines and three groups.

Engine 1: global, mobile professionals

Group 1: health-literate professionals in Tier-1 cities

They already pay for executive check-ups, private imaging, advanced blood work, some form of concierge or direct care. They’ve opted out of pure Tier-3 reliance.

What they lack is a coherent risk story over time and a plan that fits their actual constraints.

Group 2: burnt-out executives and remote workers who already travel for health

They’re already booking weeks in Swiss, Spanish, or Middle Eastern clinics “to reset”. They’re used to spending meaningfully on health.

The blocker isn’t willingness; it’s realism and repeatability. Tier-2, done properly, is a clear upgrade on what they’re already buying: serious, structured, annual, and actually connected back to their normal life.

For both groups, an annual or biennial Tier-2 programme at a mid-four-figure price is a rational business and life decision, not a fantasy spend.

Engine 2: Portugal-based core

Group 3: local professionals as a designed constraint

I’m not imagining the average Portuguese household buying Tier-2.

I care about a narrower group: Lisbon/Porto founders, senior operators, and professionals who already pay for private insurance, gyms, and some diagnostics — and are willing to allocate a rational annual amount to staying functional.

I don’t need most locals. I care whether a few hundred serious people will:

• commit to a clear, structured prevention product

• at a non-Swiss price

• more than once

They force pricing sanity, smooth seasonality, and act as clinical truth-serum. You can fake it for tourists. You can’t fake it for people who live nearby and come back.

If Tier-2 works for these three groups, it’s not a niche. It’s a template.

8. Why Portugal is my testbed, not just my base

Portugal isn’t just “where I happen to be”. It’s part of the strategy.

Economically, Portugal has been growing faster than the euro-area average in recent years, with budget surpluses and public debt back under 100% of GDP. It’s not a perfect system, but it has some headroom and momentum.

On the healthcare side:

• The public system provides broad coverage.

• The private sector delivers good quality at lower price points than many Western European countries.

• There’s already a blend of medical and wellness tourism — especially around Lisbon, Cascais, and the Algarve.

On the travel side:

• direct flights to New York in roughly seven to eight hours

• frequent connections to major European capitals

• you can leave New York in the evening, land in Lisbon in the morning, and start a structured programme the same day

Add to that:

• lower operating costs than London, Zurich, or New York

• a strong tourism brand that isn’t only about cheap breaks

• a growing population of founders and remote workers who’ve already chosen to live here

Put it together and Portugal isn’t a lifestyle choice. It’s a disciplined test environment:

• big enough and growing enough to matter

• under-built in exactly the category I care about

• convenient enough that global clients don’t have to contort their lives to get here

If we make Tier-2 work here under sober assumptions, the playbook is exportable: replicate in similar markets, partner with employers and insurers who need credible prevention, and layer software on top of a proven workflow.

This is not a one-clinic story. It’s an infrastructure pattern that can sit between Tier-1 and Tier-3 in many countries.

9. What I’m actually betting on

The business is still in stealth, so I’m not going to publish the full blueprint.

But I can be clear about the bet.

I’m betting that:

• There is a real structural gap between Tier-0/Tier-1 noise and Tier-3 reality.

• Demand already exists in behaviour, not just surveys or slide decks.

• Existing “Tier-2-ish” models mostly stop at better access, not at programmes that are engineered for outcomes and scale.

• A model that respects regulation, clinician time, and liability can work if you choose your segment and geography carefully.

• Portugal is a rational first place to prove this in practice.

I don’t see this as a vague hypothesis. The behaviour, the incentives, and the current patchwork of half-solutions all point in the same direction: this middle layer will develop.

My job is to turn that inevitability into a disciplined, working template — not a moodboard.

There are plenty of ways to get execution wrong. That’s where I expect to learn. The category itself is not optional.

10. The lens I use to keep myself honest

To keep this grounded, I use a simple internal stack:

• Biology – what actually moves disease risk and function.

• Friction – regulation, reimbursement, workforce, liability.

• Experience – how it feels and whether people come back.

• Capital – who pays, how much, and what story they think they’re buying.

Tier-0 mostly ignores this stack and hopes something sticks.

Tier-1 optimises experience and capital, gestures at biology, and often hand-waves friction.

Tier-3 is deep in biology and friction and has to fight hard for experience and capital.

Tier-2, done properly, is engineered from all four from the start. That’s the work. Not the marble. Not the sauna. Not the supplement stack.

I’m convinced enough of that, and of my three realisations about Tier-1, evidence, and my own role as a customer, to put my time and capital into building Tier-2, quietly, in Portugal.

Executed well, this doesn’t just avoid failure. It becomes the obvious standard for how prevention should look for working adults.

In ten years, I expect “Tier-2” to be a normal phrase in health-system decks — whether or not they use that exact label. My goal is to make sure the underlying template is disciplined, humane, and boringly repeatable.

If you’re building, funding, or practising anywhere near this missing middle, I’d like to compare notes.

Leave a comment

This is an educational and strategic perspective, not personal medical advice.

Most “longevity” today lives in marble lobbies and hotel wings.

A narrow, very wealthy group buys full-body scans, giant blood panels, and a story about outsmarting ageing.

Meanwhile, the biology that actually moves morbidity and mortality is dull, brutal, and very scalable: cardiometabolic risk, a few cancers, muscle, sleep, mental health.

The real commercial opportunity sits between those two realities:

• Industrial-strength prevention and risk management

• Built as infrastructure, not a club

• Priced for a broad professional class, not for the top 0.1%

That middle “Tier 2” is still mostly empty. And that’s the only part of this market I find strategically interesting.

This is an educational and strategic perspective, not personal medical advice.

1. How the industry actually looks from my side of the desk

When I strip away the branding, I still see four layers. The evidence hasn’t changed that.

1. Heritage retreats – where the myth lives

Old sanatoriums that learned Instagram.

Lakes, mountains, thermal baths, attached to five-star hotels.

You buy “programmes” of 7–14 days: controlled diets or fasting, labs, simple imaging, fitness testing, spa, massage, breath-work, maybe light cognitive work.

The price: five figures per week once you add travel and upgrades.

They do two things extremely well:

• Hospitality: everything is choreographed.

• Narrative: “I go there every year to reset and stay young.”

What they do not do is continuous, year-round risk management. They are a ritual, not a system.

2. Concierge / “performance” clinics – medicine as a membership

These sit exactly where you’d expect: New York, London, Dubai, Zurich, Singapore.

• Annual memberships: roughly €10,000 up to $250,000 per year, openly marketed as status plus “the most advanced testing and therapeutics available.”

• Data haul: full-body MRI or CT, cardiac imaging, genome + exome sequencing, polygenic scores, huge blood panels, wearables, “biological age” tests, VO₂max, microbiome.

Clients buy information, access to one or two “name” physicians, and a signal: I have a longevity team.

These are useful prototypes. But structurally:

• Prime locations + thick staffing + tiny patient panels = a model that requires very high fees.

• You can’t simply “optimise” this down to an engineer’s salary. The cost base and narrative both fight you.

3. Performance / wellness centres – the busy middle

Almost every major city has some hybrid of:

• “Executive check-ups”: labs, ECG, ultrasound, sometimes CT/MRI.

• Sports performance, hormone optimisation, IVs, aesthetics, nutrition, generic psychology.

• A subscription layer on top of fee-for-service.

Some are run by serious internists and cardiologists. Others are gyms with lab access.

“Longevity” here is mostly a label on top of conventional check-ups, sports medicine, and wellness upsells.

They are broad. They are rarely deep.

4. The consumer layer – gadgets, blood tests, and content

The outer ring is now huge:

• Wearables, sleep and HRV trackers, CGMs, rings, bands, watches.

• At-home finger-prick blood tests and supplement subscriptions.

• Protocols, podcasts, biohacker playbooks.

This layer is good at awareness. It is bad at integration.

You get dashboard overload and no real clinical plan, because nobody in the chain is structurally responsible for you over ten years.

Subscribe now

2. What the biology and data actually say

If I look at this as an engineer and an allocator, not a mystic, the picture stays stubbornly clear.

Across Europe, cardiovascular disease and cancer remain the main causes of premature death. Depending on the dataset, they account together for roughly two-thirds of premature deaths from non-communicable diseases.

Within that:

• The biggest avoidable losses come from blood pressure, lipids, glucose, smoking, and diet.

• Socioeconomic status still predicts who dies early. Lower-income groups carry more cardiovascular and cancer burden and see lower healthy life expectancy.

• Ultra-processed food intake is now linked to a measurable increase in early death risk; every 10% higher proportion of ultra-processed food in the diet correlates with several percent higher risk of dying before 75.

So the strong-evidence levers are:

• Cardiometabolic: blood pressure, ApoB / LDL, HbA1c, visceral adiposity, smoking.

• Selected cancers: colorectal, breast, cervical, high-risk lung, depending on age and risk.

• Physical function: VO₂max, muscle mass, grip strength, gait speed.

• Mental health and sleep: depression, anxiety, sleep disorders – all feed into CVD and cancer indirectly.

Everything else sits in lower evidence tiers:

• Emerging human data: some “biological age” clocks built on blood biomarkers or DNA methylation; structured resistance training for older adults; specific dietary patterns.

• Mechanistic / animal data: cellular senescence, autophagy interventions, plasma exchange, many “longevity molecules”.

• Speculation: stacks of off-label drugs and devices being marketed as anti-ageing.

The blunt point: the most powerful levers are boring and scale beautifully. The most visible products are theatrical and target people who already outlive everyone else.

That misalignment is an ethical problem. It’s also a margin and scale problem.

3. What the luxury clinics actually get right – and where they’re stuck

I don’t worship the high-end clinics, but I don’t underestimate them either.

What they get right

1. They proved willingness to pay.
   We now have real-world proof that thousands of people are happy to spend $8,000–$250,000 per year on intensive diagnostics, recurring scans, and bespoke prevention.
   That’s not a meme. It’s a demand signal.

2. They made prevention aspirational.
   Architecture, concierge service, exclusivity – these are levers on behaviour. They get people into full-day work-ups they would otherwise avoid.

3. They run live workflow experiments.
   They answer questions by doing:

   • How many diagnostic stations can you fit into an 8-hour day without quality dropping?

   • What is the true no-show rate for annual MRI if people fly in?

   • How many hours of physician time per year does a high-complexity client actually consume?

   That operational learning is real IP, even when the science is messy.

4. They are testbeds for biomarkers and protocols.
   Work like Mironov et al. and the longevity clinic roundtables is starting to codify which biomarkers are actually being used, how often, and with what claimed ranges.

If you view Tier-1 clinics as R&D and operations labs rather than finished products, they become easier to reason about.

Where the model hits a hard ceiling

1. The unit economics are not fixable by “efficiency”.
   Prime real estate, hotel-grade fit-out, nurses standing idle for VIPs, 24/7 concierge, and small client panels are not bugs; they are the product.

   You cannot spreadsheet your way from €100k annual memberships to €5k without destroying the value proposition they actually sell: exclusivity.

2. Evidence discipline is structurally weak.
   The field still lacks standardised, validated biomarkers for “biological age” and harmonised protocols for what a “longevity assessment” even is.

   That doesn’t mean everything is pseudoscience. It means a lot is experimental but sold as established.

   A practical example: whole-body MRI.

   • There is growing interest, amplified by influencers and wellness clinics.

   • It clearly detects some serious disease early in higher-risk groups.

   • But professional bodies like the American College of Radiology explicitly state there is insufficient evidence to recommend total-body MRI screening in asymptomatic, average-risk individuals due to overdiagnosis and cascade testing.

   Many luxury clinics sit ahead of the evidence curve here. That’s fine for R&D. It’s dangerous when painted as “precision medicine”.

3. Data is trapped in silos.
   One Bay Area clinic now collects more than 150GB of health data per client – imaging, genomics, labs, wearables – and charges $8,000–$19,000 per year.
   But this data is largely stored as internal assets, not structured into open cohorts that move the field forward.

4. Reputation risk is shared.
   When a €250k membership bundles robust cardiovascular prevention with speculative interventions and vague promises, critics don’t separate the two. They attack “longevity clinics” as a category.

So I treat Tier-1 clinics as necessary prototypes, not as a scalable blueprint.

4. The alternative frame: longevity as infrastructure

The question I use to force discipline is:

What would this look like if we built it like infrastructure, not like a private club?

Infrastructure has specific properties:

• Tightly defined scope

• Standardised components

• Clear interfaces with public systems

• Designed for volume and repeatability

For longevity, that means a “Tier-2” layer with:

1. Narrow, high-yield clinical scope

Anchor on:

• Cardiometabolic risk

• Selected, high-yield cancer screening

• Strength, function, and falls risk

• Sleep and mental health

On top of that, allow a controlled frontier:

• Advanced imaging, genomics, biological age scores, and novel biomarkers – but only when they change management or are being explicitly studied under protocol, with proper consent as experimental.

2. Diagnostics treated as a utility, not a flex

Instead of trophy panels, design:

• A standard baseline (age- and sex-adjusted)

• Escalation tiers driven by family history, early findings, and risk models

• Strong lab and software integration so that cost per datapoint collapses as volume rises

The value is not in a 140-page report. The value is in the ratio of signal to noise and the ability to act on it.

3. Clinic as operating system, not hotel

The scarce asset is not a star doctor or a view of the marina. It’s the operating system:

• Physicians make key decisions and handle complexity.

• Nurses and allied health professionals run standard protocols and follow-ups.

• Behaviour specialists and coaches work on adherence.

• Everything sits on versioned pathways that update as evidence changes.

A mundane but telling detail: in many EU clinics, a patient still signs three separate paper consent forms (bloods, imaging, data sharing) because the MRI safety checklist is a non-editable PDF printed on an ageing LaserJet in the corner. Staff then re-enter the same demographics into two different systems. That kind of friction is exactly where infrastructure thinking has to start.

4. Geography as a cost and behaviour lever

The location question, for me, is not “where looks nicest in photos” but:

• Is regulation strong and aligned with EU-grade safety?

• Are real-estate and labour costs structurally lower than in wealth hubs?

• Does the environment make behaviour change easier (walkability, food culture, climate, pace)?

Portugal is interesting on those dimensions today, but specifics always depend on current law and must be checked with qualified EU/Portuguese legal and tax counsel.

5. Pricing that a serious professional can repeat

Not mass-market. Not yacht-level.

The target band is: reachable for engineers, founders, senior operators, business owners – and repeatable annually without becoming their defining expense.

If the only people who can afford you are the same families flying to Tier-1 clinics, you haven’t built Tier-2. You’ve just created another Tier-1 brand with different furniture.

5. The market map: where the white space actually sits

If I sketch the market as three tiers, it looks like this:

• Tier 1 – Luxury / R&D
  Retreats, concierge memberships, R&D clinics. High price, low volume, lots of experimentation.

• Tier 3 – Public and insured care
  Primary care, statutory screening, hospitals. Overloaded, indispensable, politically constrained. Prevention budgets are thin and short-term.

• Tier 2 – Longevity infrastructure (the white space)
  Evidence-heavy prevention and risk management. Built like a system, not a programme. Priced for a broad professional class.

Today we have:

• Plenty of Tier-1 cathedrals.

• A noisy consumer layer underneath.

• Under-resourced Tier-3 systems trying to plug gaps.

• Almost no coherent Tier-2 capacity.

Why Tier 2 is still empty

• Luxury incumbents are brand-trapped.
  If they push down into automation, standardisation, and mid-market pricing, they dilute the exclusivity they sell.

• Public systems can’t move fast.
  Their incentives are political cycles and budget caps, not 10-year prevention ROI.

• Consumer health companies avoid clinical weight.
  They stop at data, coaching, and content; the regulatory and liability footprint of running a real clinic is a different game.

That’s why I see Tier 2 as the genuine strategic white space.

Leave a comment

6. The risks I take seriously

Any time I catch myself thinking “this is straightforward,” I go back to this list.

1. Clinical risk: over-testing and cascades

Full-day diagnostic work-ups invite incidental findings and false positives.

The radiology community is openly debating the “perils and promise” of whole-body MRI, especially the risk of downstream anxiety and unnecessary procedures in low-risk people.

This is not a theoretical problem. Over-testing can:

• Harm individual clients

• Destroy trust with referring physicians

• Blow up your own capacity with unnecessary follow-up

A Tier-2 clinic needs explicit thresholds for what it will not do, not just menus of what it can do.

2. Evidence drift

The temptation to add new biomarkers and interventions is constant.

Mironov and others have started outlining frameworks for which biomarkers and interventions belong in a healthy longevity clinic and how to classify them by evidence strength. But standardisation is still weak; the latest roundtable reports explicitly flag heterogeneity and confusion around “longevity biomarkers” and “biological age scores”.

My own rule set would be:

• Strong human outcome data → core offering

• Mechanistic or early human data → clearly labelled frontier; optional

• Animal or in-vitro → stay in trials, not in the clinic brochure

If a clinic can’t show this hierarchy in writing, it’s running on vibes.

3. Behaviour and economics

It is easy to sell one big “day in the scanner and lab”.

It is hard to keep a client aligned on medications, strength training, sleep, and nutrition for three to five years.

Without adherence:

• Clinical outcomes disappoint

• Lifetime value drops

• Your acquisition cost becomes impossible to justify

Any Tier-2 model without a serious adherence and behaviour system is an expensive on-ramp to nowhere.

4. Data and governance

High-frequency labs, imaging, wearables, AI risk scores – all of this creates governance questions:

• Who owns the raw data and the models?

• How is it de-identified and used for learning?

• What happens if employers or insurers want access?

The EU is already moving on cardiovascular and cancer prevention plans and broader NCD data initiatives; the regulatory climate is only going to get stricter.

If a clinic treats data governance as an afterthought, it’s carrying an unpriced liability.

7. Why the timing actually matters now

On the demand side:

• Europe’s gains in life expectancy have stalled; in some countries they’ve gone backwards. Rising obesity, poor diet, and inactivity are key drivers via CVD and cancer.

• Policymakers now talk bluntly about cardiovascular disease and cancer as the dominant causes of premature death and are drafting EU-level cardiovascular health plans to sit alongside Europe’s Beating Cancer Plan.

• Health-literate professionals increasingly want more than a basic check-up but don’t see themselves in a €100k concierge model.

On the supply side:

• The “longevity market” is already a multibillion-dollar segment on track to multiply by 2035, but most capital has flowed into Tier-1 brands and consumer gadgets.

• We have built plenty of cathedrals and dashboards.

• We have not built the plumbing that connects validated prevention to ordinary (but high-responsibility) professionals at a rational price.

That combination – political pressure on NCDs, stalled life expectancy, proven willingness to pay, and immature infrastructure – is exactly when Tier-2 models become possible.

8. How I want investors and operators to read this

If you’re evaluating anything in “longevity”, I’d ignore marble and marketing screenshots for a moment and ask three blunt questions.

1. What exactly are you standardising?

If the answer is “everything is bespoke,” you are buying a craft studio, not infrastructure.

Ask for:

• Versioned care pathways

• Clear inclusion/exclusion criteria for diagnostics

• A documented evidence hierarchy for each intervention

2. How do you learn from every client – and who owns that learning?

If the data model is a graveyard of PDFs and proprietary dashboards, the “learning loop” is storytelling.

You want:

• Structured, queryable data across cohorts

• Clearly defined patient consent pathways for secondary use

• A plan for publishing something – or at least generating internal evidence robust enough to survive external scrutiny

3. In ten years, will this look more like a brand or more like infrastructure?

Both can make money.

• Brands can exit cleanly and leave little behind.

• Infrastructure changes how a population ages and becomes hard to dislodge.

My own bet – and the thing I still argue with myself about at 3 a.m. – is that the real leverage will sit with the teams who quietly build Tier-2 pipes while everyone else keeps polishing Tier-1 lobbies.

9. The forward line

If we do this properly, “longevity” stops being a luxury identity and becomes an unglamorous utility that serious professionals just factor into their annual budget – like tax advice or cloud infrastructure.

The strategic question is simple:

Are you building cathedrals, gadgets, or the plumbing that actually decides who gets extra healthy years?

This article started as a shopping list.

I sat down to plan the diagnostics and “frontier tools” for the longevity project I’m building in Portugal. The list was long: devices, chambers, protocols, drips. On paper, it looked like a greatest-hits compilation of the current longevity scene.

Then I took that list to my medical advisory board.

Item by item, we went through it.
Item by item, I saw the same expression:

• “Interesting mechanism… but for whom?”

• “This is promising… but based on what level of evidence?”

• “Would you really offer this before you’ve fixed blood pressure, sleep, strength?”

I realised something uncomfortable:

I was the founder, the one supposed to design the ethical, evidence-based version of this space – and even I had quietly bought the marketing story before I understood the reality.

If I can fall for the polished narrative while staring at a capex spreadsheet, what chance does a time-poor, health-anxious client have?

That conversation sent me back to first principles:
What actually helps? What belongs where? What does an honest longevity stack look like?

This piece is my attempt to answer that – partly for myself as an operator, and partly for anyone who’s ever been handed a glossy menu of “longevity” options and felt a quiet unease they couldn’t name.

I also keep meeting the same person in different cities.

Mid-40s to late-50s. Successful. Time-poor, health-anxious, “optimisation-curious”. Multiple wearables. Has worn a continuous glucose monitor at least once. A growing folder of lab results and “longevity panels”.

They can quote their resting heart rate, HRV, step count, and “deep sleep minutes” on command. They’ve even done a few “full workups”.

What they don’t have is a serious, joined-up conversation about what all of this should add up to in terms of blood pressure, strength, aerobic capacity, and week-by-week behaviour — or where, in that picture, drips and other frontier tools genuinely belong.

If that’s you — or you’re building or backing a clinic selling to that person — this is who I’m writing for.

Walk through any major city and you’ll see the new pantheon of “longevity”:

• Marble-floored clinics

• Biomarker labs

• Full-body MRI suites

• Peptide bars

• Hyperbaric pods

• “Cellular reset” IV lounges

The implicit promise: if you can afford the frontier tools, you can skip straight to advanced optimisation.

Some of this is genuine progress. Preventive care is finally socially acceptable. Our understanding of metabolism, inflammation, and ageing biology is far beyond “eat less, move more”.

But underneath the chrome sits a simple, uncomfortable reality:

There is a widening gap between what people believe they are buying and what science can actually prove.

To navigate that gap, you don’t need to become a biochemist. You need two mental models:

• the Three-Floor Longevity Stack — how care should be sequenced

• the Longevity Evidence Ladder — how strong the proof really is

Everything else is commentary.

The Three-Floor Longevity Stack

If I had to sketch responsible longevity care on a whiteboard, it would look like a three-floor building.

Floor 1 — Foundations (Behavioural & Physiological)

The boring, powerful fundamentals with the strongest long-term human data:

• Muscle mass and strength

• Cardiorespiratory fitness

• Blood pressure, glucose, and lipids under control

• Sleep and circadian stability

• Low smoking and heavy drinking

• Meaningful, stable relationships and reduced isolation

Floor 2 — Structured Medical Prevention

Conventional medicine used proactively and intelligently:

• Appropriate screening and imaging

• Evidence-based management of hypertension, lipids, and diabetes risk

• Targeted use of established drugs where they change outcomes

• Basic nutritional and psychological support woven into care

Floor 3 — Frontier Tools

The experimental, story-rich layer with more uncertainty and more risk:

• IV “longevity” drips

• NAD⁺ infusions and precursors

• Hyperbaric oxygen therapy (HBOT) for “anti-ageing”

• Grey-zone injectable peptides

• Plasma exchange and “blood cleansing”

In a sane system, people climb in order: Floor 1 → Floor 2 → Floor 3.

In the pitch decks and shopping lists (including my first one), the order is often reversed.

If your clinic doesn’t touch your blood pressure, sleep, and strength before it touches a drip, it’s not doing longevity — it’s doing theatre.

Call that the minimum viable standard for longevity care: blood pressure, sleep, and strength before drips.

That’s the line most of the sector is quietly crossing.

The Longevity Evidence Ladder: Mechanism → Biomarker → Outcome

The second model is the Longevity Evidence Ladder. Whenever you hear a longevity claim, ask: on which rung are we?

1. Mechanism
   Something interesting happens in a cell or pathway — often in animals or a petri dish.

“This molecule activates this pathway in mice.”

1. Biomarker
   A lab value or imaging metric moves in a direction we assume is good.

“This protocol lengthened telomeres in blood cells.”

1. Outcome
   Human beings live longer, stay independent longer, or avoid major disease more often over meaningful time.

“Lower cardiovascular events; better survival; more functional years.”

Most frontier tools live on rung 1 or 2.
Most marketing (and my original equipment list) treated them as if they were safely on rung 3.

If a claim never climbs past mechanisms and biomarkers, it doesn’t get to wear an outcome halo.

Using tools on rung 1 or 2 isn’t inherently unethical. Charging outcome-level prices and making outcome-level promises when you only have mechanism and biomarkers is.

Five Questions to Ask Any Longevity Clinic Before You Buy Anything

These questions simply combine the Stack and the Ladder into something you can use in a consultation.

When someone offers you a “longevity” intervention, ask:

1. “Which rung of the Evidence Ladder are we on — mechanism, biomarker, or outcome?”

2. “Which floor of the Longevity Stack is this — foundation, medical prevention, or frontier?”

3. “If this tool didn’t exist, what would you recommend I do instead to improve my long-term risk?”

4. “What are the main risks or downsides, in plain language, for someone like me?”

5. “How will we know it worked, and when would we stop?”

A credible operator will not be threatened by these questions. If they look vague or defensive, the problem is not your curiosity.

From here, notice how the same pattern repeats: strong story, mechanistic promise, weak outcome proof — and mis-targeting to people whose Floor 1 is not yet in order.

I. IV Drips: Clinical Tool, Lifestyle Illusion

Marketing promises: “Immune boost”, “Detox”, “Cellular repair.”
Evidence tier: Strong outcome-level evidence in hospitals (dehydration, malnutrition); minimal evidence for longevity in otherwise healthy people.

IV therapy is lifesaving in emergency and inpatient medicine. That clearly belongs on Floor 2.

For a generally healthy person walking into an IV lounge “for longevity”, the story is different. There is no robust long-term data showing that routine vitamin drips in people with normal intake and absorption extend lifespan or healthspan.

A specific red flag: IV iron

Iron infusions are essential when someone is truly iron-deficient and cannot correct it orally. But “elective” IV iron in non-deficient people is not a harmless upgrade.

Free iron can drive the Fenton reaction — reacting with hydrogen peroxide to generate highly reactive radicals that damage cells and tissues. You may be paying for a “detox” while quietly increasing oxidative stress in the organs you are trying to protect.

If you have:

• documented deficiency

• appropriate work-up

• a clear indication

…an iron infusion in a medical setting can be appropriate. For the average tired executive with normal labs, it is risk without proven long-term upside.

If someone suggests an IV drip for “longevity”, ask:

• “Which deficiency are you correcting, and can I see that in my labs?”

• “Do you still recommend this if my levels are normal?”

• “Beyond feeling better today, what long-term outcome are you expecting — and based on what evidence?”

Different tool, same pattern: clinical value in Floor 2, promises on Floor 3, sold to people with half-built Floor 1.

II. NAD⁺: The Biomarker That Became a Promise

Marketing promises: “Mitochondrial reset”, “Youth drip”, “Brain reboot.”
Evidence tier: Strong mechanistic and animal data; early and mixed human data; no long-term outcome data in healthy adults.

NAD⁺ is central to energy metabolism and DNA repair. Levels of NAD⁺-related pathways do appear to change with age in specific tissues. That’s the mechanistic story.

Short-term human studies show that infusions or precursors can raise NAD⁺ levels or shift metabolic markers. That’s the biomarker story.

What we do not yet have is convincing, long-term outcome data showing that aggressively boosting NAD⁺ in otherwise healthy adults clearly extends life or reduces major disease events.

The acute reaction no one markets

Fast NAD⁺ infusions commonly cause intense discomfort: chest tightness, abdominal cramping, a sense of “impending doom”. This is likely linked to smooth muscle activation as the infusion hits the system rapidly.

For an already optimised individual who understands that this is an experimental, marginal bet, that may be acceptable. For someone who smokes, sleeps five hours, doesn’t exercise, and is trying to “buy more time”, it is a poor use of risk and capital.

Calling this “reversing ageing” because a metabolite moved on a blood test is not optimism; it’s marketing inflation.

If a clinic proposes NAD⁺, ask:

• “Is the evidence here mainly mechanistic/biomarker-level, or do you have outcome data in people like me?”

• “What proportion of your clients experience acute side-effects, and how do you manage them?”

• “If I skip NAD⁺, what are the top three things you would prioritise instead to change my long-term risk?”

III. Hyperbaric Oxygen (HBOT): Real Medicine, Overstretched Story

Marketing promises: “Reversed biological age”, “Regeneration.”
Evidence tier: Strong outcome-level data for specific conditions; early biomarker-level data for “ageing”; no direct lifespan data.

HBOT is not pseudoscience. It is standard of care for decompression sickness, certain non-healing wounds, and a few other well-defined indications. That is Floor 2 medicine.

The longevity hype leans heavily on early protocols where structured HBOT increased telomere length and reduced senescent immune cells in older adults. These are biomarker shifts — interesting, but not the same as “you are now younger”. Changing telomere length in one blood cell population does not automatically mean you have reversed ageing across an entire human being.

The unglamorous risks

• Pressure-related injuries to ears and sinuses

• Risk of seizures at high exposures

• Fire risk in oxygen-rich environments, especially outside hospital-grade setups

In a tightly controlled medical environment, for a well-selected person who already has strong foundations, HBOT may be a reasonable experiment. In a strip-mall chamber sold as “reverse your age in 10 sessions”, it is being marketed far beyond its proof.

If you are offered HBOT “for longevity”, ask:

• “Is my use case one of the established medical indications, or is this experimental?”

• “What non-trivial risks apply at the pressures and protocol you use here?”

• “What specific changes — beyond a glossy report — would count as meaningful success?”

Again: mechanisms and biomarkers on Floor 3, sold to people who haven’t yet earned Floor 1.

IV. Peptides: Pharmacology Without Guardrails

Marketing promises: “Cellular rejuvenation”, “Fat loss”, “Optimised hormones.”
Evidence tier: Ranges from fully approved drugs with strong trial data to research chemicals with almost no human safety data.

At one end of the spectrum, we have drugs like GLP-1 agonists with large, regulated trials and clear indications.

At the other, we have compounds such as BPC-157, CJC-1295, TB-500 and others, often sold as “research peptides” or “not for human use” while being marketed informally for exactly that.

Because these often sit outside normal pharmaceutical channels, they can escape standard quality controls. Independent testing and regulatory alerts have found:

• Mislabelled vials

• Impurities and incomplete peptide chains

• Microbial contamination and endotoxins

• Vials whose contents do not match the label at all

You end up with a paradox: drug-like effects without drug-level oversight.

In a formal medical program or clinical trial, with real sourcing and monitoring, certain peptides may be worth exploring for a well-optimised person. Buying them from an online reseller because a podcast said “this heals everything” is a different proposition entirely.

Before touching peptides, ask:

• “Is this an approved drug for my indication, or a research-grade product?”

• “Who manufactures this, and under which quality standards?”

• “How will you monitor benefits and side-effects over time — and what would make you stop?”

V. Plasma Exchange and “Blood Cleansing”: Big Procedure, Small Proof

Marketing promises: “Microplastic removal”, “Immune rejuvenation”, “Next frontier.”
Evidence tier: Strong outcome-level data for specific autoimmune and blood disorders; mechanistic speculation for longevity in healthy adults; no proven lifespan or healthspan benefit in that group.

Therapeutic plasma exchange (TPE) is a real hospital therapy. In certain autoimmune and haematologic conditions, it can be life-saving. That is legitimate Floor 2 medicine.

The longevity narrative leans on two ideas:

• Diluting “pro-ageing” factors in the blood

• Physically removing microplastics or other contaminants

Mechanistically, these are plausible hypotheses. They are simply not yet supported by clear human outcome data in otherwise healthy people.

The practical irony

The extracorporeal circuits used for “blood cleansing” are made of plastics that can themselves leach plasticisers. You undergo an invasive, resource-intensive procedure, with vascular and bleeding risks, for a longevity benefit that remains theoretical.

If someone pitches plasma exchange “for longevity”, ask:

• “What specific diagnosis or risk are we treating, in my case?”

• “What human outcome data exists in people like me — not in mice, not in theory?”

• “How do you weigh the procedural risks against that, and who would you advise not to do this?”

Different machine, same pattern: heavy hardware, light outcomes, sold as if it were rung 3.

The Uncomfortable Truth: The Boring Stuff Wins

Why does the industry lean so heavily on expensive, theoretical technologies?

Because the interventions that actually move long-term outcomes are hard to package and hard to monetise.

When you zoom out to large human datasets — decades-long cohorts tracking people over time — a different picture emerges. Simple, unglamorous capacities and behaviours show up again and again as associated with lower mortality and disability. For example, long-running Finnish work on sauna use shows lower cardiovascular and all-cause mortality in people who do something as mundane as sit in a hot room regularly. Not because sauna is magic, but because simple, repeatable behaviours compound.

Across the literature, the most robust drivers of healthy longevity look suspiciously like Floor 1:

• Maintaining muscle mass and strength

• Building and preserving cardiorespiratory fitness

• Keeping blood pressure, glucose, and lipids within healthy ranges

• Protecting sleep and circadian rhythm

• Reducing smoking and heavy drinking

• Staying embedded in real, not performative, relationships

These fundamentals don’t photograph well. They don’t fit neatly into a weekend retreat. Margins are lower. But they are the engine.

For the vast majority of people, frontier longevity tools are not “advanced optimisation” — they are a detour around the only interventions with proven impact.

The small minority who might reasonably consider experimental tools are usually those who have already:

• Lifted consistently for years

• Built meaningful aerobic capacity

• Normalised their metabolic markers

• Stabilised their sleep

• Removed obvious toxins like smoking and heavy nightly drinking

• Put some basic order into their psychological and relational life

They are playing for marginal gains at the edge. Many others are being sold the edge as a starting line.

If that sounds blunt, good. Longevity should be.

If You’re Actually in the Optimised Minority

If you can honestly say you:

• train with resistance and aerobic work most weeks, for years

• have stable, well-controlled metabolic markers

• sleep 7–9 hours on a reasonably consistent schedule

• don’t smoke, and don’t sedate yourself with alcohol every evening

• have at least a basic grip on stress and relationships

…then you are closer to the group for whom frontier tools might be considered as experimental, marginal bets.

If that’s you, a more honest way to think about the frontier is:

1. Treat frontier tools as “risk-budget spending”, not necessities.
   Assume the real compounding still comes from your foundations. Drips, chambers, peptides, and exchange circuits are things you spend a clearly defined risk and money budget on — not the core of your strategy.

2. Demand to know exactly which rung you’re on.
   For every tool: is the evidence mainly mechanistic, biomarker-level, or outcome-level? If a tool never climbs past mechanism and biomarkers, treat it accordingly.

3. Insist on medical-grade process, not vibes.
   Real clinicians. Written protocols. Dosing, monitoring, and specific stop criteria: “We stop if X happens or if we don’t see Y by time Z.”

4. Be explicit about your motivation.
   Curiosity and optimisation are legitimate motivations. Just don’t pretend a mechanistic experiment is a guaranteed longevity dividend. Naming your motive keeps you and your clinician on the ethical side of the line.

Frontier tools are not automatically unethical. Used transparently, on top of a well-built foundation, with a clear risk budget, they can be explored as optional upside.

The ethical failure happens when they are sold as foundations to people who do not yet have one.

If Your Clinic Can’t Do at Least This, It’s Not in the Longevity Business Yet

This is where the conversation needs to go next. Less “are drips good or bad?” and more “what does a defensible model look like?”

A responsible longevity operator would:

• Sequence care by floors.
  Every client starts on Floor 1. Floor 3 tools are not routinely offered until key basics — strength or a fitness proxy, blood pressure, glucose, lipids, sleep — are at least being measured and actively addressed.

• Map every tool onto the Longevity Evidence Ladder in writing.
  For each intervention, there is a short, client-facing summary: what is known (outcomes), what is promising (biomarkers), what is still theoretical (mechanisms).

• Define eligibility and exclusion criteria for frontier tools.
  Not everyone gets everything. Some tools are explicitly reserved for people who have cleared certain basics or meet clear clinical criteria.

• Track outcomes beyond “felt great after the session”.
  Strength, fitness, metabolic markers, sleep measures, functional capacity — followed over years, not weeks, alongside any frontier interventions.

• Disclose risk with the same boldness as promise.
  Oxidative stress from iron, contamination risks in grey-zone peptides, barotrauma and fire risk in HBOT, vascular risks in plasma exchange — on the table, not in the fine print.

If your “longevity clinic” leads with drips and never measures your strength, it is not a longevity clinic. It is a day spa with syringes.

Clinics that get this right will also be the more durable businesses: lower regulatory risk, fewer angry ex-clients, and a compounding data asset on what actually moves outcomes. In jurisdictions like the EU, and countries like Portugal where medical regulation is already strict and likely to tighten further, building as if Floors 1 and 2 are mandatory is not idealism; it’s risk management.

In a mature sector, regulators will eventually demand this kind of sequencing and transparency. The only real question is who chooses to behave as if that future already exists.

The Business Line No One Talks About

Right now, the longevity marketplace thrives on two confusions:

• Blurring mechanisms and biomarkers into implied outcomes

• Treating optimised edge cases and average, under-served clients as if they should be sold the same tools

We don’t need less innovation. We need more honesty about two things:

1. Where each tool really sits on the Longevity Evidence Ladder.

2. Which floor of the Longevity Stack, and which type of person, it is genuinely appropriate for.

We have normalised selling the feeling of safety to people who are actually buying structured uncertainty — when many of them should first be offered sleep, strength, and glucose control.

That meeting with my advisory board – the doubtful looks at my shopping list – was the best kind of humiliation: the kind that forces you to rebuild your model from the ground up.

Innovation will shape the future of preventive medicine. But in longevity, the scarce resource is not another device, molecule, or chamber. It is trust — earned by telling people not just what you offer, but whether they are the kind of person who should be offered it at all.

The venture I am building, and this publication, will live on that line: optimistic about breakthroughs, ruthless about shortcuts, and committed to a simple rule:

If your blood pressure, sleep, and strength haven’t been touched, your drips can wait.

I wrote this first to correct my own thinking. I’m sharing it so you don’t have to pay as much tuition to the marketing machine as I did.

Most health businesses are built backwards.

They start with a logo, a location, and a menu of services - then try to bolt on medical rigor, data, and operational discipline later. The result is what we all see and feel: fragmented care, perverse incentives, and rising costs in systems already consuming close to a tenth of national GDP.

We’re doing the opposite.

We are building a medically-led health and longevity venture in Portugal from the ground up: clinical protocols first, infrastructure second, brand last.

This publication is where we document that build.

Who We Are

We are siblings and co-founders with a shared background in the tech sector.

Our careers were shaped by a world that optimises for speed and scale. Our personal obsession has always been different: the question of how to add more high-quality, functional years to a human life.

At some point, those two tracks converged.

We decided to take the parts of tech that actually transfer - rigorous thinking, data discipline, systems design - and apply them to the most constrained domain we know: physical health.

We’re very clear on what we are not:

• We are not here to “disrupt healthcare” with slogans.

• We are not here to sell longevity fantasies.

• We are not here to pretend that a complex, regulated system can be “fixed” with an app.

Health is a hard business: regulated, capital-intensive, emotionally loaded, and accountable for real human outcomes. We approach it with the fresh eyes of outsiders, and the responsibility of operators who have skin in the game.

The Space We’re Building

This is not a newsletter about “top 10 biohacks” or the latest health trend on social media.

This is a thinking space for people who care about the infrastructure of health.

You might come from:

• Health tech or diagnostics

• Real estate or hospitality

• Investment or capital allocation

• Clinical operations or healthcare strategy

What ties this audience together is not a job title, but a shared question:

How do you build health infrastructure that actually improves long-term outcomes - and still makes economic sense?

Some of the tensions we care about:

• Tech vs. Clinical Reality
  What happens when the software mindset (“ship fast”) collides with clinical risk, medical guidelines, and regulatory friction?

• Hospitality vs. Outcomes
  How do design, service, and environment change adherence, diagnostics quality, and recovery - not just “patient satisfaction”?

• Portugal as a Laboratory
  We are building in Portugal - but the underlying questions about regulation, capital, staffing, and operational discipline apply to anyone operating in a complex healthcare market.

• Innovation vs. Hype
  How do you separate genuinely useful diagnostics and longevity medicine from noise, overclaiming, and unproven interventions?

Our Working Mental Model

The way we think about The Business of Health is a simple three-layer stack:

1. Biology & Diagnostics
   What is actually happening in the body? Which biomarkers, imaging, and protocols meaningfully shift healthspan risk - not just create more data?

2. Infrastructure & Experience
   The physical space, workflows, and hospitality layer that either amplifies or destroys the value of the clinical work.

3. Capital & Incentives
   The business model, regulation, and ownership structure that determine whether good medicine can survive long enough to matter. Chronic disease alone is projected to continue as a major drag on GDP without better prevention and management.

Most failures in health ventures are not due to a missing idea. They come from misalignment between these three layers.

What This Publication Is (and Isn’t)

This is:

• A build log: what we are learning as we design a medically-led longevity and diagnostics venture in Portugal.

• A critical lens: where we dissect broken incentives, mispriced risk, and operational mistakes - including our own.

• A repository of leverage points: regulators, operators, investors, and clinicians all have different levers; we want to map where they actually move the system.

This is not:

• A medical advice column.

• A catalogue of our internal blueprints.

• A hype channel for “the next big thing” in health.

We will get things wrong in public. When we do, we will correct them in public.

Why Portugal, Why Now

Portugal gives us an interesting combination: EU-level regulation, a growing health and wellness tourism market, and a real estate landscape where long-term, medically serious projects can still be built.

We see it as a testbed for a simple question:

Can you build a clinically rigorous, economically sane, hospitality-grade health venture in a system that was not designed for prevention and longevity?

If the answer is yes here, the lessons travel.

What You Can Expect

Over time, you’ll see:

• Deep dives on diagnostics and longevity medicine as a business: where the margins are, where the risk sits, and what actually improves outcomes.

• Field notes from our own build: design decisions, missteps, regulatory puzzles, staffing questions, and real estate trade-offs.

• Maps of incentives: who gets paid, for what, and how that shapes behaviour across tech, clinics, insurers, and hospitality.

The through-line: we care about mechanisms, not slogans.

Welcome to the Build

Substack has a long list of opinions on how to launch a publication. Ours is simple: start with the truth.

The truth is that building in healthcare is slow, expensive, and unforgiving - and that’s exactly why it’s worth doing well.

If you’re an operator, investor, clinician, or just someone who thinks deeply about the future of health infrastructure, we invite you to follow along as we build.

Welcome to The Business of Health.