Health

I. Where we start

Today the world is still digesting the pandemic: mRNA proven, trust in medical institutions cracked, obesity and mental health as slow epidemics, public systems overloaded, the United States with the most expensive and unequal arrangement in the West. GLP-1 drugs are already changing weight and the market. AI reads scans in pilots. Japanese and European ageing is the trailer for what is coming to China and, later, to others.

II. The decade in between

Over the decade medicine flips from craft to engineering. The diagnostic layer collapses in price: the phone, the watch, the stool sample, the scan. Then — the decade's real surprise — the treatment layer follows it down. AI-designed proteins and antibodies move from paper to pharmacy in years rather than decades; trials get faster because models pre-screen the failures and biomarkers read the answer early. Cheap molecules flood downward as they always did, and the expensive ones stop stacking upward: gene edits and custom oncology ride the same cost curve every manufactured thing rides once the design becomes software. The critical path is still a nurse, a carer, a bed — and the licence for the model that initiates care, which arrives country by country, because the queue turns out to be worse than the risk.

The retained speculative leap is that partial epigenetic reprogramming generalises across organs over the decade. Today's ER-100 programme is an early human study in optic neuropathies, aimed first at safety and tolerability; it does not establish whole-body rejuvenation. This scenario imagines that later trials demonstrate durable functional gains as well as changes in biomarkers, first locally and then through several organ-specific procedures. Each needs delivery, monitoring, and a licence. That extraordinary extension is the chosen fiction with a scientific on-ramp, not a reported clinical result.

What arrives instead is the course: a supervised programme of injections, imaging, and immune management, repeated on a schedule, for life. There is no molecule to copy into a generic — only a licensed procedure and a proprietary delivery vector. Maintenance is recurring in this scenario, but a lapse does not erase all prior benefit: decline resumes from the attained condition, with uncertain durability. Humanity's first real rejuvenation is therefore also its first subscription to its own youth.

A pandemic will try again at some point; ten years is a short window for guarantees. The difference is the machinery waiting for it: genomic surveillance that reads the sewers weekly, vaccine platforms that go from sequence to needle in weeks, and a first-clinician model in every pocket that notices the cluster before the ministry does. This scenario still budgets for an outbreak. It no longer budgets for a lost year.

III. Ten years from now

The typical package of this scenario:

• The model as first clinician. For most complaints, nearly everywhere, the first licensed actor is software — and this is a triumph, not a downgrade. It reads images, records, voice, and wearables better than the median specialist, writes the protocol, and pages a person for the exception. The general practitioner becomes what they always wanted to be: the human for the hard part, with time to do it properly. Automation error exists and generates lawsuits; the error rate is far below the old queue's, and everyone quietly knows it. Poor countries leap furthest: the phone that names the disease now also orders the generic that treats it, because the molecule got cheap the same decade the diagnosis did.
• Diseases fall by category, not by drug. Weight drugs become ordinary infrastructure on every continent as the generics arrive. A handful of cancers become chronic or cured for most patients, not just the well-insured, as AI-designed therapies compress the pipeline. Gene therapy leaves ultra-rare and enters common conditions with a falling price tag. Alzheimer's gets its first disease-modifying class that visibly works. The drugs that merely slow ageing — senolytics, the metabolic stack — go generic and slightly boring, and hand the median body a handful of decent extra years. This is the scenario's broad medical dividend, and it reaches most people. It is also, this decade, no longer the story.
• Ageing runs backwards — on a payment plan. For perhaps two or three million people worldwide, biological age is a number that goes down. A woman of sixty carries the arteries, marrow, skin, and reaction time of someone in her late thirties, re-verified every quarter by clocks that no longer flatter anyone. The price falls all decade — from the low millions a year to the high hundreds of thousands, a large decline — and changes nothing about who can hold it, because the schedule never ends and the bill arrives forever. No ordinary universal health system covers a course; China reserves a separate selective state programme. Two governments have fallen arguing about it.
• Biological age becomes the number that matters. The epigenetic clock leaves the laboratory and enters the paperwork: insurance, lending, hiring, custody, pensions, dating profiles. Chronological age becomes a courtesy, like a middle name. Two people born the same week can sit forty actuarial years apart, and anyone can read the difference off a cheek swab for the price of a coffee.
• The telemetered body. Watch, ring, continuous glucose, implantables for the sick. The stream's first reader is now your own model, which nags earlier and better than any insurer ever managed — prevention finally has a delivery mechanism. Privacy fights continue and are worth having; the default, though, has flipped to the data serving the body it came from.
• Housing as prevention. Less crowding, dry rooms, tolerable indoor temperatures, and fewer forced moves reduce avoidable illness in adopting cities. Accessible apartments let some older people remain near relatives; clinic staff can live near their shifts. A cheap unit with mould or dangerous heat is not a health gain. Mental-health assistants widen first contact, with human escalation for serious cases. Stable shelter reduces stress without turning a lower rent into a cure for psychiatric disease.
• Elder care splits in two. Night robots and accessible homes delay some institutional care, while human hours remain essential for dementia, bathing, and company. Rejuvenation postpones needs for its wealthy buyers; it does not eliminate disability, accidents, infection, or death. Their political attention drifts away from public care, but the tax base is not excused from funding it. The housing boom helps staff recruitment and home care below the line without making the course affordable.

Systems. For everything that is a molecule, the health curve and the cost curve still bend in opposite directions — and the course sits outside that frame entirely, priced by no insurer and unaffordable to ordinary household and universal-care budgets. The United States becomes the world's clinic: no national system positioned to say no, the deepest capital, the most licences, and inbound medical tourism at a volume that embarrasses its own State Department. Europe legislates an access right it cannot fund, then watches its reset cohort fly to Texas and the Gulf. China builds a state programme and rations it by merit and loyalty rather than by price, producing the decade's only sizeable cohort of rejuvenated people who are not rich — an advertisement it uses relentlessly, and effectively. Brazil's SUS delivers the generics beautifully and the course not at all. Africa gets the cheap miracle drugs a decade early and the course a century late.

Body and class. The last draft of this plate said healthy life expectancy was converging for the first time since anyone measured it. That was true, and below the line it stays true — the generics arrive, the median body gets its extra decent years. But a second curve has detached from the species and is climbing away from the first. The distance between the median body and the reset body is no longer measured in years; it is a difference in kind, and it is legible from across a room — the ones who do not read as any particular age, because they have stopped having one. The oldest human question, how long have I got, becomes for the first time a question about money rather than about biology.

IV. Uncertainties

What does not happen in this scenario: immortality; the end of death; the extinct human doctor; a course in a pill; universal public coverage of the course. What does happen: ageing becomes reversible and priced, the epigenetic clock becomes a class document, and a single species starts keeping two mortality curves.

The tensions are almost all political now. Black-market courses with unlicensed vectors, and the cancers they seed in people who could not afford the licensed kind. States that tax the course, states that quietly subsidise it for their own elites, and the two or three that try a lottery and cannot make the arithmetic decent. An unreset majority who know, numerically and to the quarter, exactly what is being withheld — the most combustible fact anywhere in this atlas. And beneath all of it, the reset cohort's own discovery that a body of thirty-five carrying seventy years of memory is a genuinely new kind of person, and that nobody, including them, knows how to be one.