256 | Kelly and Zach Weinersmith on Building Cities on the Moon and Mars
What this covers
Kelly and Zach Weinersmith discuss their book A City on Mars with host Sean Carroll, examining whether permanent human settlement beyond Earth is feasible, desirable, or wise. Rather than cheerleading for space colonization, they methodically catalogue the scientific unknowns, legal constraints, and ethical hazards that make rushed settlement dangerous. The conversation moves through biology, international law, economics, and geopolitics—grounding the romantic vision of Mars cities in the stubborn realities of gravity, radiation, closed ecosystems, and the absence of crucial research.
The Weinersmiths argue that core sciences remain essentially undeveloped. No mammal has yet reproduced and had surviving offspring in space, leaving the effects of partial gravity and radiation on human gestation unknown. Biosphere 2 demonstrated how fragile closed-loop ecology proves—the crew lost 10–18% of body mass, oxygen had to be piped in, and internal factions grew hostile. China's Lunar Palace experiment had to swap crew members by body size just to balance carbon dioxide. They examine whether astronaut self-reporting is trustworthy given incentive structures that punish honesty about health problems. On the practical side, they assess settlement locations: the moon's scarcity of water and carbon makes it less viable than Mars, though lunar lava tubes offer radiation shielding. They dismantle the helium-3 mining fantasy as technically implausible and economically absurd.
The legal and strategic picture proves equally thorny. The Outer Space Treaty genuinely forbids land appropriation, a principle sustained through repeated crises since the 1960s—the seabed and Antarctica followed the same commons logic. Yet ambiguity persists: is resource extraction permitted? Who authorizes terraforming? The Weinersmiths argue that going multi-planetary may increase existential risk rather than hedge it—interplanetary warfare poses nightmare scenarios, and a near-term scramble between superpowers for the moon's scarce premium real estate resembles colonial land-grab. They contend humanity should slow down, master the unknown biology, and negotiate international agreements for resource-sharing and property rights before people arrive.
The Weinersmiths argue that permanent, self-sustaining space settlement is far harder, more ethically fraught, and more legally constrained than advocates admit, so humanity should proceed slowly—mastering biology, reproduction, closed ecosystems, and international law before settling—rather than rushing to put people on Mars.
- Core sciences like off-world human reproduction and closed-loop ecology are essentially undeveloped, making rushed settlement unethical
- International law (the Outer Space Treaty regime) genuinely constrains land appropriation and is more durable than space advocates assume
- Settling space may increase rather than decrease existential risk via interplanetary warfare and great-power scrambles for scarce lunar real estate
We lack critical knowledge about human reproduction and development in partial gravity and space radiation before attempting settlement.
- A baby developing under one-sixth (moon) or 40% (Mars) gravity may not develop bones strong enough to survive Earth's atmospheric entry and walking on Earth, and it is unknown whether human hips could even handle childbirth after a lifetime at reduced gravity—since in microgravity bone mineral density in the hips drops about 1.5% per month, though partial gravity might mitigate this in unknown amounts.
“You lose 1.5% of aerial bone mineral... Density... in your hips is lost per month.”
- Human reproduction off Earth is essentially undeveloped science: no mammal has yet had sex, given birth, and had surviving offspring that themselves reproduced in space, so we do not know how partial gravity (one-sixth on the moon, 40% on Mars) or space radiation affect gestation and development—making it unethical to attempt human settlement that depends on reproduction without first running animal experiments.
“We don't even have rodents that have had sex, had babies, those babies survived, and they went on to have babies. None of that has happened in space so far.”
- Closed-loop life support is so finely balanced that China's Lunar Palace experiment had to swap two large male crew members for two smaller women to stop carbon dioxide from building up too much—illustrating that we have not yet worked out the calculations needed to scale life support before sending people to Mars.
“they're breathing gases where it was like too much carbon dioxide was building up. So they switched out two big guys for two smaller women, and then it worked.”
- Astronaut and cosmonaut self-reports of mental and physical health are unreliable because the incentive structure punishes honesty: admitting depression, panic, or chest pain means losing future flights, so they routinely conceal symptoms—evidenced by astronauts who hid suspected heart attacks, memorized vision tests, removed pages from medical records, and a cosmonaut culture in which lying is expected and breaking the lie is the offense.
“they almost all lie about their mental and physical health. And it's the incentives, if they say, oh, you know what? I'm actually, I'm depressed... You don't get to go up for another flight.”
Settlement requires international legal frameworks governing resource extraction, property rights, and dispute resolution under the Outer Space Treaty.
- Terraforming Mars—via hurling comets or detonating fusion bombs—is not just a millennium-scale technical challenge but raises an unaddressed geopolitical-legal question of who is authorized to do it; such actions are arguably legal but obviously terrifying and environmentally fraught, and using nuclear weapons would tank space settlement for 20–30 generations.
“what geopolitical legal framework countenance is this? Like, who's doing it?”
- The most important provision of the Outer Space Treaty is that you cannot appropriate land—so a private actor on Mars cannot declare territory as their own country—though it remains legally ambiguous whether extracting and selling resources counts as appropriation; the US interprets resource extraction as permitted while some developing nations view the moon and celestial bodies as a global commons that should not be exploited unilaterally.
“the most important part is it says you're not allowed to appropriate land. So if Elon Musk gets to Mars and he says, this is Moscow and this is my country... that is not allowed by international law.”
- Under modern international law a people may gain a right to 'remedial secession'—taking part of a nation—if they are deeply oppressed (as in Bangladesh), a high bar that Quebec failed in 1998 because Québécois can vote and hold office; a future oppressed moon settlement could conceivably win recognition to secede, but doing so would be unprecedented because it would carve property out of a global commons rather than from an existing nation.
“if a people are under deep oppression by some government, they may have a right to what's called remedial secession.”
- Claims that private actors like Elon Musk can do whatever they want in space because they have the rockets are bogus, because under international law activities on Mars are the responsibility of the launching state, and any actor antagonizing every government on Earth faces consequences—resupply ships can be grounded, and a billionaire with terrestrial business interests has too much to lose to genuinely defy Earth.
“the things that happen on Mars are the responsibility of the United States... So the FAA should be thinking about the ethical implications of this before they allow it to happen.”
- The Outer Space Treaty's commons approach was not a 1960s fluke: every time in the 20th century humanity was called to regulate a newly accessible region—Antarctica, the seabed, outer space—it chose a commons, and these have been sustained; the Antarctic moratorium (CRAMRA) even banned looking for exploitable minerals until 2048, showing the planet has previously said no to profitable resource extraction over fears of conflict and environmental outcry.
“every time in the 20th century, humanity was called upon to regulate a newly accessible region of the universe, we picked a commons. And that has been sustained.”
- Space law arose not mainly because nations were good at cooperating in the 1960s but because high-atmosphere and in-orbit nuclear weapons tests in 1962–63 so thoroughly terrified everyone that they came to the table—a dynamic potentially relevant to current AI governance, where being scared into action by rational actors may matter more than baseline cooperative spirit.
“there were high atmospheric and in-orbit tests of nuclear weapons, which so thoroughly terrified everyone that people were willing to come to the table.”
A self-sustaining settlement needs enormous population and infrastructure complexity that existing closed-ecosystem experiments show we cannot yet manage.
- A truly self-sustaining (autarkic) off-world settlement would require an enormous population—plausibly a million or more people—because survival depends on advanced technology like computer chips (currently made only in water-intensive, highly specialized fabs) and thousands of medical and engineering specialties; the lowest published estimate, by a pro-settlement author, is 100,000 people and even that assumes extraordinary advances in AI and robotics.
“the number of people that you would need... we're probably talking about like a million or more people.”
- Biosphere 2 demonstrated how difficult closed ecosystems are: across 3.14 acres the crew had to pipe in oxygen because curing cement absorbed it, suffered pest stowaways, faced crop and livestock failures, lost 10–18% of their body mass while working all day just to survive, and split into two hateful factions—teaching that you should start with many small experimental systems rather than one giant biome.
“they had to pipe in oxygen because the cement was messing with their atmosphere and they all were getting like headaches and stumbling around”
- NASA was created as an Apollo-era political instrument to beat the Soviets, and after that goal was achieved it entered what historian John Logsdon calls a forty-to-fifty-year identity crisis; because it was never set up as a settle-the-moon agency, space-station research is a forty-year 'grab bag' of uncoordinated experiments lacking the systematic program needed to answer settlement questions.
“There's a historian named John Logsdon... he referred to as what happened afterward as a forty-year identity crisis. I think he might say fifty years now.”
Lunar lava tubes offer the best near-term settlement sites because they provide radiation shielding and temperature regulation while reducing habitation complexity.
- Lunar lava tubes—caves potentially up to 100 times larger than Earth's—could be the best places to set up a base because you can seal off the walls to create a habitat bubble that also shields against radiation, temperature swings, and regolith, likely producing a small set of distributed settlements rather than one metropolis given how few good sites exist.
“these absolutely extraordinary caverns, which would be a great place to set up a base because instead of bringing a habitat or building it on site, you just seal off the walls.”
- Before attempting Mars, humanity should build out a substantial base on the moon to learn unknown best practices—an 'employee handbook' of how to convert dead soil into a sustaining ecosystem and how to manage reproduction—because the moon is only two days away (versus six months to Mars), so people can be sent home quickly if things go wrong.
“It's almost less about science than something like an employee handbook. There's just a lot of stuff you don't know until you get there.”
- Humanity should not settle space as fast as possible but should first run more biology experiments to understand the risks and reduce ethical problems, and use the extra time to work out an international system for resource extraction, profit-sharing, and property rights to forestall future revolts—making the process take decades by design.
“We would like to see this take decades, but it could be rushed.”
- The popular idea of mining helium-3 on the moon for fusion is unrealistic because we lack any fusion device—let alone the exotic aneutronic reaction it requires—the aneutronic benefit is overstated, and helium-3's low concentration means strip-mining vast lunar surface with equipment built for an airless, low-gravity, two-week-day/night environment; as an ESA scientist put it, even bars of gold sitting on the moon would not be worth retrieving.
“if there were bars of gold sitting on the moon, it would not be worth it to go get them.”
- Although backing up the biosphere is a real argument for settlement, going multi-planetary may actually increase existential risk: interplanetary warfare has nightmare potential because weapons like bioagents thrown down a gravity well are hard to retaliate with given six-month transit times, and a near-term scramble between nuclear superpowers (US and China) for the moon's tiny premium real estate resembles a colonial land-grab.
“it is a kind of open question whether settling Mars increases or decreases over the long term existential risk. Because... interplanetary warfare has like nightmare potential.”
Mars is the most viable settlement target because it has abundant water and a carbon-dioxide atmosphere, unlike the water-poor, carbon-scarce moon.
- Of the moon, Mars, and rotating space stations, Mars is the most likely place for a self-sustaining human settlement because it has abundant water nearly everywhere and a carbon-dioxide atmosphere that supplies carbon for life and plants, whereas the moon has little water (only at hard-to-reach poles) and very little carbon.
“The three places that we focus on the most are the moon, Mars and rotating space stations. And out of all of those, we think that Mars is probably the most likely.”
- Martian soil is about 1% perchlorates, a chemical that binds to thyroid hormones regulating blood pressure and heart rate and that studies show is taken up by plants, so settlers would have to remove perchlorates from the soil before farming—on top of dust storms that would make solar power difficult.
“it's about 1% perchlorates, which is a chemical that binds to the hormones made by your thyroid... first, you're gonna have to pull the perchlorates out of the soil before you can even start your farm.”
Settlement timelines and methods require careful design: small diverse protein sources for early diets, gravity's practical benefits for manufacturing, and multiple small experimental systems rather than one large biome.
- Earth is so hospitable that it would be far easier to build a self-sustaining city on the bottom of the sea than on the moon or Mars; as a thought experiment, the only way to make Earth as bad as Mars would be to slam it with enormous asteroids and superheat its surface—so unless you fear Earth literally ending, there is little benefit to an underground moon/Mars city over an underground Earth city.
“it would be far easier to build a city on the bottom of the sea... We actually as an intellectual exercise ask yourself what you could do to Earth to make it as bad as Mars... You'd basically have to superheat the surface to get anything like Mars. Earth is good.”
- For an energy-constrained settlement, the bigger the animal the less efficiently it converts feed energy into edible calories, so early Martian diets would favor very small protein sources—mealworms, crickets, cultured cellular meat, or flavorable fungus—and likely be largely vegetarian rather than featuring beef or chicken.
“the bigger the animal, the less efficient you're going to get at converting energy to calories... you'll arrive at something like Crickets”
- Gravity, though annoying, is genuinely useful: nearly every zero-G vehicle has had problems with waste and feces floating through the air (nicknamed 'brown trout' on the shuttle), and the difficulty of matching floating objects suggests manufacturing large spacecraft may be easier on the moon than in orbit.
“almost every single vehicle that has been in like Zero-G has had problems with escape of waste and faeces floating through the air. On the shuttle, they were called brown trout.”
- A space settlement, as distinct from an outpost, is defined as somewhere people can have children—implying at least a few thousand people—and the most ambitious version requires permanent self-sustainability, meaning it could carry on across generations even if Earth were destroyed or lost contact.
“we would say settlement is somewhere where people can have kids. And so you're probably at least talking like a few thousand people.”
- Space debris is less of a problem than commonly feared because space is very large and organized into distinct altitude shells, with coordination systems for moving objects; debris is a concern for low Earth orbit but not for the moon, Mars, or L5, and even the roughly 37 nuclear reactors currently in space at 700–1000 miles altitude are not widely worried about.
“I wouldn't want to be living in a rotating space station in low Earth orbit with all that debris, but I wouldn't worry about it if I was on the moon or Mars or L5”
- Whether humans have had sex in space is a probabilistic question with no confirmation; Zach estimates a 1–5% chance using actual historical facts (early all-male crews, careerist astronauts, constant modern monitoring), while Kelly argues humans will inevitably find a way, with both agreeing any occurrence most likely happened on a late, less-monitored Russian space station.
“what value do you assign to the likelihood that it has happened? And I hit somewhere like one to 5% chance that it's happened.”
Cheaper reusable rockets and AI advances have made previously implausible settlement timelines conceivable, creating genuine uncertainty about future feasibility.
- Apollo deliberately avoided in-space rendezvous—launching a single skyscraper-sized rocket whose pieces fell away until a tiny lander reached the moon—because rendezvous is genuinely hard (as the Gemini-Agena docking problems showed), but that single-launch approach is unworkable for building a settlement, which is why scaling up many reusable launches matters.
“part of the weird thing about Apollo is there is no rendezvous. You just launch a skyscraper and little pieces of it break off until finally a teeny weeny dinghy goes to the moon”
- Recent rapid advances in AI and robotics make low population estimates for self-sustaining settlements—once dismissed as fantasy—more plausible, since in 20 years a world with working fusion and robots that set everything up in advance could substantially change the picture, which is why predicting settlement timelines is uncomfortable.
“I kind of laughed about it two years ago. And that's changed... you wonder about if in 20 years, there's a world where we have fusion figured out and you can just send robots to set everything up in advance.”
- The cost of putting mass into orbit dropped by as much as 99% between roughly 1957 and 1970, stagnated for decades, then fell by around three-quarters since 2015; cheaper, reusable rockets like Starship matter beyond per-kilogram price because they allow off-the-shelf parts rather than mass custom-fit to a single fairing—though expensive modules, not just launch, dominate station costs.
“between 1957, say, and 1970, the cost of going to space... dropped by maybe as much as 99%. It hasn't happened since, but since 2015... it's dropped by like three quarters.”