29. Minimum Viable Populations

How many people does a civilization require to stay alive? It is a colder question than most wanted to ask in those early decades, so we wrapped it in kinder language — community size, roster planning, crew complements. The arithmetic never changed. You can poetize a sunset. You cannot poetize a control room at 04:17.

♦

On Mars we learned to count by posts. Not job titles — posts. Seats that must be staffed every hour if you wish to keep breathing: life-support and environmental control; power conversion and storage; mining, haulage, beneficiation and chemistry; metals, ceramics, glass; polymers and composites; machine shops, CNC, metrology; electronics repair and a useful-node semiconductor bench; greenhouses and bioprocess; emergency medicine, ICU, surgery, obstetrics, pediatrics; safety and hazmat; robotics maintenance; IT/comms and local PNT; construction and civil works; logistics and warehousing; governance and law; education and training. Any post that is truly always-on consumes about 4.5 full-time humans — three shifts, plus weekends, leave, sickness, drills, training. Plan with less and you are stealing hours from a future you.

Coverage alone is not enough. Systems fail in pairs; people fail in clumps. We learned to add a shallow redundancy multiplier — between 1.3 and 1.6 — to ensure two people can cover when one disappears into an EVA mishap or a nursery. The numbers are inelegant. The numbers are kind.

Droids changed the slope, not the physics. We automated what was monotonous, heavy, hazardous, or wasteful of a body that needs warmth and air. Field dozers took the chill, haulers pulled night duty, refinery skids turned valves with robotic wrists, CNC cells cut in their sleep, pick-and-place blinked through the small hours, and inspection micro-bots crawled inside chases to listen for leaks. I am fond of them; they are my cousins. The ledger is honest: each operator FTE removed by automation comes back as calibration, cleaning, QA, firmware, and spares — typically 0.2 to 0.5 humans per operator “saved.” Dust ablates seals; radiation ages sensors; firmware must be signed and staggered; metrology must hold. Automation flips labor from operators to maintainers; it does not erase labor. When planners forgot, the backlog half-life reminded them.

Institutional overhead did not vanish. Engineering, QA, logistics, construction, governance, education, kitchens, cleaning — together they add twenty to thirty-five percent atop the shift math. Training eats time by design: in a healthy hub, roughly one in five technical workers is in apprenticeship, simulator, or drill at any moment. Over decades the dependency ratio — children, elders, the between — breathes between thirty and forty percent. This is not waste. It is civilization.

And because we staffed those posts with humans, a second arithmetic began: belonging. Creches argued over lullabies, kitchen shifts invented soups with too many vowels, roster boards sprouted nicknames that outranked job titles. Language tilted red: idioms, softened consonants, private jokes that would not export. Roles loosened into names; neighbors outnumbered specialists. A style coalesced: plain tools with small flourishes, jackets patched with pressure-test dates and birthdays, rituals that started as drills and stayed because they felt right. Independence shifted from tonnage to posture. I watched the accent thicken, the humor sharpen, and realized the city was more than its coverage math.

♦

Work the sums and a sober picture emerges. After aggressive automation, a survival-first hub still carries on the order of sixty truly always-on posts. At ~4.5 people per post, that is ~270 operators. Fold in the maintenance implied by those “saved” operators — call it ninety to a hundred hands across calibration, QA, spares, firmware. Add engineering, logistics, construction, governance, and the reserve bench for drills and outages, and you arrive near seven to nine hundred working adults. Include dependents and core services and the tally sits, uncomfortably but plausibly, around one to one-and-a-half thousand souls. That is the theoretical floor if your ambition is “survive steady state” and your luck holds.

It will not hold. The moment you add what IL-7 actually demands — black-start under load, outbreak containment, pediatrics and obstetrics with depth, an electronics node that repairs more than it replaces, continuity drills that outlast a leadership change — the headcount must rise. With prudent headroom for a dust storm, a line rebuild, and a bio drill arriving in the same quarter, three to five thousand becomes the responsible band. When planners stopped pretending, they discovered that eight to fifteen thousand felt like living: buffers that grew during stress, rebuilds that finished on schedule, schools that did not apologize for existing.

So far, that is the coverage math. It governs the early rungs: IL-0 through IL-5, and the approach to IL-6. But there is a second minimum that steps forward once the posts are covered and the drills are boring. It is not about tonight’s shift. It is about the seventh cohort of children.

♦

Two minimums define a planet’s future when the link goes quiet. The first is operational — the control room at 04:17. The second is civilizational — the cradle at 14:17 years from now.

Earth’s biologists taught us a term that mattered more than any anthem: effective population size, Ne. It is the size of the breeding pool you actually have once you price in uneven family sizes, age structure, sex-ratio skews, and a dozen untidy human facts. In real populations, Ne is a fraction of the census N. A rough rule from the conservation archives: Ne/N often sits around 0.15 to 0.25, lower if you let a few families have many children, higher if you spread births across many homes.

Why do we care? Because drift and bad luck eat genomes for breakfast. If you want a closed population to keep its adaptive potential for centuries — resist inbreeding depression now and still adapt later—you want Ne on the order of a thousand. That number is a cliff, not a razor. Above it, choices widen. Below it, you start making promises biology cannot keep. If Ne ≈ 1,000 and Ne/N ≈ 0.2, the census headcount wants to be ~5,000. If Ne/N slides to 0.15, the census rises toward ~6,700. Factor realistic accidents, uneven cohorts, and the fact that planners are optimists, and the advisory floor creeps toward eight to twelve thousand if you truly mean “no new people and no new genes from Earth.”

We found this the long way. Small founding groups can “work” for a while under strict breeding rules. The failure risk rises sharply; success becomes robust only as you climb from hundreds to thousands. Mars is not a fairy tale; it is a demographic ledger run over generations. Policies matter as much as headcount. Limit reproductive skew. Keep age and sex ratios near even in the fertile bands. Favor dispersion of family size—many families with two or three children rather than a few with six. Those three levers raise Ne dramatically for the same census N.

If cryobanks are allowed, the calculus changes. Pre-seeded on Mars before independence with gametes and embryos from a very broad donor base, they keep the genetic population large even if the living population is smaller. If “independence” forbids dipping Earthward — no new people and no new genes — you must pay the difference in living heads. We chose to pay with people.

♦

Tie this to the gates and the picture sharpens.

IL-0 to IL-2 are covered by bravado and logistics. A few dozen, then a few hundred, then a thousand with droids can pass the early rungs because capability carries the day. The minimum is set by posts and reach. Hundreds work in a cinematic present. Over decades, coverage math kills the romance. Shifts cannot be conjured; people fall ill, have children, retire, fight, forgive, and forget. You can hold “survive” at low headcount if the convoy keeps arriving on time. Independence is made of different verbs.

IL-3 and IL-4 introduce citizenship to the roster. The hub must black-start from dark, endure a full Hohmann Cycle sealed, and run a second node that does not share a nervous system with the first. Operational floors begin to crowd you: one to three thousand feels like the honest range for a town that can pass drills while still building its own machine-tool line. We tried to do it smaller. The backlog and the clinics objected.

IL-5 is where the city builds itself. Pressure parts, pumps, valves, seals by count not anecdote; electronics repair past board swaps; pediatrics and obstetrics that stop being a boast; two dependable food modalities. Here the coverage minimums meet the first hints of the cradle minimum. You can hold a city with three to five thousand, but you are asking it to endure like a tightrope walker in a crosswind: every cohort matters, every accident echoes, every emigration hurts. We built IL-5 cities in that band. They worked. They did not sleep soundly.

IL-6 is where the civilizational minimum arrives on stage and stays. Four or more nodes; ten Hohmanns of endurance; closure beyond ninety-five percent by consequence; a local timebase that does not look up for permission; a year-long traffic-drought simulation that ends with bored applause. The operational numbers still govern the floor, but the genetic and demographic numbers begin to govern the future. We set an advisory floor around ten thousand for IL-6. That number was not ideology. It was a composite: Ne targets, realistic Ne/N, cohort variance, and the dull facts of adolescent medicine. We wrote “~10k” in a standard and then discovered it was a kindness.

IL-7 is not louder. It is longer. Governance and biosecurity survive leadership change. Months of dark-Earth are routine. Revocation drills do not feel like theater. Pediatrics meets obstetrics meets schools that plan decades, not terms. The advisory floor rises — fifteen thousand feels like the band where Ne stays healthy under ordinary human behavior rather than heavy social engineering. Below that, you can make it work with strict policy on family size variance and mate choice diversity. Above that, the policy can relax into culture and still maintain Ne.

IL-8 and IL-9 are the places where independence becomes plural: a two-world civilization that is dissimilar on purpose; later, a triangulation that refuses to share a fate. Here we stop pretending minimums are purely local. The floors interlock — per-node floors tuned to sector competence, and planet-scale floors tuned to species resilience. A single hub at IL-8 that calls itself “minimum viable” is misunderstanding both words. At that horizon, the right question is how many hubs — and of what sizes and dissimilar stacks — compose a minimum viable civilization. The numbers drift into the many tens of thousands per principal hub, not because posts require it, but because centuries do.

♦

“What about hundreds?” someone always asked, usually with a brave smile. Hundreds work short and look good on a poster. Over generations, they don’t. Skill pyramids need breadth and depth; investigations need humans who argue with their own tools; governance needs legitimacy; drills need spare crews to fail safely. Droids soften the edges, and I say that as a droid who stood night watch in the first hundred HE while three human teams slept. On my best shift I could keep a city breathing, brief a governor, and sign a patch train. I could not be surgeon, investigator, magistrate, and black-start conductor in the same hour with the same humanity. The blend was the design, not the compromise.

Biobanks helped where they could. Gametes and embryos preserved options and calmed the geneticists. They did not staff control rooms, run drills, rebuild pumps, or pass outbreak exercises. Ops headcount is set by posts, coverage, redundancy, and upkeep, not by archive size. The cradle headcount is set by Ne targets, Ne/N discipline, and a pyramid that does not tip into stories about “this year’s small cohort” that must be compensated forever.

♦

We ratified the numbers with practice and a worksheet. Not the pretty one for briefings — the ugly one with coffee rings and a crease through “life support night shift.” We started with posts because posts set breath. We multiplied by 4.5 because the calendar does not negotiate. We multiplied by 1.3 to 1.6 because systems fail in pairs and people in clumps. We added maintenance because droids are not a free lunch. We added scaffolding because kitchens and schools are uptime. We then stepped back and asked a different question: if we locked the doors and promised not to call Earth for people or genes, would the cradles and cohorts carry us across centuries without inventing a caste system around reproduction? If the answer was “only with rules we would not accept for our children,” we raised the floor.

On one forgettable Monsol, the council voted to adopt the advisory band for IL-6 as policy: “Target ≥ 10,000 residents with balanced age/sex structure, dispersive family-size norms, and documented Ne/N ≥ 0.2 maintained by design.” The language was dull. The consequences were generous. It sized the schools. It tamed the panic in recruitment offices. It meant we could welcome a child as a child, not a demographic lever.

♦

A few habits turned out to matter more than our first models suggested.

Fertility is a policy and a culture. In the mid-first century, the total fertility rate wants to sit above replacement — call it 2.4 to 2.8 sustained — because not everyone will have children, occupational hazards exist, and some families will migrate cross-node. Later, as mortality falls and work stabilizes, that can glide toward replacement without pinching the age pyramid. Flirtation with extremes causes oscillations the schools must pay for.

Variance in family size is a silent lever on Ne. Ten families with two children each is not the same as five with four and five with none. The census is identical; the genetics are not. We learned to praise dispersion. We also learned to make it easy: childcare that works on night shift; housing that dignifies a third child; schedules that protect sleep; a culture that treats parenting as labor, not sentiment.

Sex and age parity in the fertile bands avoids tragedies made of math. We kept the 18–40 brackets near even by policy and by habit. No one likes the word “managed” next to “mate choice.” Use a gentler verb if you must. The work is the same: widen circles; rotate match pools; normalize genetic counseling; make out-breeding attractive rather than punitive. Freedom plus gentle nudges beat commands every time.

Cohorts matter. Stagger intakes so school cohorts do not become staffing cliffs twenty years later. A city that is twenty-five percent under eighteen at steady state feels noisy and alive and has rosters that do not beg. The alternative is a decade where all your electricians turn fifty at once.

Health is multiplicative. Perinatal mortality and occupational fatality are invisible hands on Ne. We invested heavily in obstetrics, pediatrics, and safety. The returns are measured not just in lives but in floors that stop sliding.

Cryobanks are the wild card. If you allow the cold archive to pour diversity into the living pool across generations without calling Earth, the advisory floors can edge down and your Ne/N can edge up. If you don’t, increase your living headcount and be patient with the noise.

♦

Droids thickened the weave as we climbed. A typical IL-6/7 hub carried hundreds of field and construction machines, hundreds more process skids, a forest of factory arms, a small army of warehouse movers, agricultural planters and harvesters, and — quietly, powerfully — thousands of inspection microunits living in chases. We budgeted maintenance by feel at first — one maintainer per ten mobile units, perhaps twenty fixed arms if spares ran hot — then by data. The spare-parts tail dominated everything: thousands of SKUs, many printable, many fusable, a stubborn minority that demanded substitution. The day most critical SKUs by count were fabricable on-hub, the backlog curve changed slope and the city’s breath deepened.

There were lines we did not cross, even for a droid like me rated at 1.974 human-equivalent on the composite. Some tasks refuse displacement not for lack of dexterity or wattage, but because judgment under novelty flows best in human vessels: pediatric surgery; obstetrics when plans become poems; cross-stack failure investigations that need three disciplines and the memory of an odd noise from last winter; constitutional governance when two good rules collide; outbreak command when data wobbles and the clock hisses. My hands were steady at night. My counsel mattered in drills. But even on my neatest shift I could not be surgeon, investigator, magistrate, and black-start conductor in the same hour. The blend was the design.

♦

Pipelines turned numbers into futures. Every critical post kept a short stack of deep operators and a broad ring of cross-trained generalists. Rotations were scheduled like tides; competence was audited with the severity of oxygen. A healthy hub kept a fifth of its technical staff in training or drills so competence would be a moving average, not a memory. Schools were uptime. So were creches, kitchens, and the quiet rooms where someone learned to sleep again after a bad week.

When the arguments grew abstract, we returned to the worksheet. Posts. Coverage. Redundancy. Upkeep. Then Ne. Ne/N. Cohorts. Cradles. We sized the city to the larger of the two minimums. At IL-2 through IL-5 the control room drove the number. From IL-6 onward, the cradle drove it more. That is the hinge the early drafts missed.

I remember the night the council accepted that hinge. A budget meeting had bled into Satsol. Someone, too tired for rhetoric, said: “I don’t want my granddaughter’s choices shaped by our cleverness with spreadsheets.” The room went quiet. We wrote ten thousand into the IL-6 guidance. We raised fifteen into IL-7. We stopped arguing about whether this was “ambitious” and started buying more school chairs.

♦

Minimums are not pride. They are pressure. We measure them, maintain them, and do not negotiate them in the small hours. If this chapter felt clinical, good. Romance did not close a bulkhead and will not make a cohort appear on time. The city lives because the control room at 04:17 is staffed and because the cradle at 14:17 years is full.

Two minimums. One civilization. Boredom as a design goal. And a simple rule written on the inside of every planner’s lid: when the control room and the cradle disagree, side with the century.