Uncharted Territories

Climate Caused the US Civil War

Northern climate crops — wheat, barley, oats — required roughly 33 labor-hours per acre and mechanized readily (threshing machines cut that to 11 h/a by 1910). Southern crops — cotton (~130 h/a), sugarcane (~250 h/a), tobacco (300–900 h/a) — demanded far more hand labor and resisted automation for decades. That gap made enslaved workers economically essential in the South and irrelevant in the North, which abolished slavery by 1830.

The labor difference drove a population divergence. Nearly 90% of immigrants settled in the North: wages were higher, land ownership was accessible, and the South carried malaria, yellow fever, and hookworm below the Mason-Dixon line. Black Africans had stronger immunity to those diseases, making enslaved workers even more attractive to planters. By 1860, the North held 22 million people against the South's 9 million.

The Senate's equal-state rule turned that demographic gap into a crisis. Both sides raced westward to admit friendly states and hold the Senate balance needed to block anti-slavery legislation. The North won: by 1860 it had three more states than the South. Lincoln won on an anti-slavery platform; southern states seceded before he was inaugurated.

The full causal chain: climate shaped crops; crops set labor intensity; labor intensity determined whether slavery was viable; slavery viability shaped immigration; immigration shaped population; population determined the Senate balance; the Senate balance triggered war.

Never Bet Against America

American dominance is underwritten by geography, not institutions, and geography doesn't change. The Mississippi Basin is the decisive asset: the world's largest contiguous block of farmland, flat and naturally irrigated, drained by navigable rivers that together exceed the rest of the world's inland waterways combined. Moving goods by water costs 10–30x less than overland; halving transport costs can increase trade 16x, and the US captures that advantage for free. Chicago sits at the hinge between the Mississippi and the Great Lakes, making inland cities like Detroit and Cleveland de facto seaports. The basin's integrated river system also drove political unity where Europe's fragmented rivers produced rival nations. The US is the world's top producer of both oil and gas, a legacy of an ancient shallow inland sea that deposited hydrocarbons across the interior.

Defense layers are equally structural: two oceans each over 2,000 miles wide, the Appalachians guarding the east, the Rockies and Sierra Nevada the west, an ice-desert Canada to the north, and a geographically fragmented Mexico — stripped of 55% of its territory in 1848 — to the south. NATO, Japan, South Korea, Taiwan, and Australia extend these buffers globally. Bad policy can create headwinds; it cannot override this.

China vs US: Which Geography Is Best?

China's geography mirrors the US — continent-sized, ringed by seas and mountains — but its defenses are materially weaker. US-aligned states (Japan, South Korea, Taiwan) form a coastal chain that could choke China's Pacific access, as Western navies proved during the Opium Wars. The Vietnam border has a low-elevation gap China has invaded through repeatedly. China has more land neighbors than any country, demanding constant management. Northern invaders repeatedly breached the Great Wall — half of Mongol invasions succeeded, one founding the Qing dynasty.

Why Is Moscow So Weird?

Moscow's improbable position — cold, landlocked, on a minor river — is the residue of steppe violence that eliminated every more obvious alternative.

The Crimean Khanate, an Ottoman vassal, "harvested the steppe" for slaves: roughly 3 million Slavs captured in the 15th–16th centuries, 200,000 Russians in the first half of the 17th alone — about 3% of the population. This emptied the fertile south. Kiev, capital of the older Kievan Rus on prime farmland, was destroyed by this pressure; Ukraine could have anchored a different superpower.

Moscow survived in a narrow band where woodland slows cavalry and blunts archery yet farming stays viable. The Volga–Oka–Ugra–Moskva river system forms east-west barriers against north-south nomad columns. At the 1480 Great Stand on the Ugra, Ivan III garrisoned every ford and let mud and winter close the season. When nomads still sacked Moscow in 1571 — population fell from 100k to 30k, 150k enslaved — Russia built the Great Abatis Line: 500 km of felled-tree palisades 180 km south, manned by soldiers given land in lieu of pay. Matranga and Natkhov identify this as the institutional origin of Russian serfdom.

Expansion followed least resistance. Moscow took Novgorod in the 1470s for Baltic fur revenues. Siberia fell in under a century: its rivers drain into the Arctic, its population was sparse, and European diseases killed up to 80% of locals. Challenging the Hanseatic League, Poland-Lithuania, or Ottoman-backed Crimea was costlier — so Russia reached the Pacific before the Baltic or Black Sea.

Moscow's granaries still lie south and west, Siberia remains effectively foreign, and the Northern European Plain security dilemma is unresolved.

Putin's Mindset on Russia

Russia is the last remaining colonial empire — expanding east in the 1500s–1600s just as Spain and Portugal expanded globally, via gunpowder, the fur trade, and Orthodox missions, with native populations decimated by disease. It never decolonized: Siberia is landlocked, sparsely populated after conquest, and resettled by ethnic Russians. Moscow still runs it extractively — controlling subsoil resources, suppressing minority languages, and conscripting non-Russians for Ukraine. China, densely populated on the Siberian border while Russia's side is empty, is biding its time.

Why Florence Started the Renaissance

Florence's Renaissance was a specific stack of geographic, economic, and political advantages that converged nowhere else.

The triggering symbol is Brunelleschi's dome on Santa Maria del Fiore (1436): the largest masonry dome ever built, surpassing the Roman Pantheon — completed without its lost cement recipe and without interior scaffolding. Brunelleschi combined Gothic pointed-arch geometry, a double-shell hollow dome, and sandstone-and-wood compression rings that convert tension into compression. The critical input was proximity to Rome: he and Donatello spent two years measuring its ruins; fellow Florentine Bracciolini recovered Vitruvius's *De Architectura*, giving artists a theoretical framework for Roman proportional systems. Gothic was the French-German style; Florence, allied to Rome against the Holy Roman Empire, defined Renaissance in deliberate opposition — reviving round arches, domes, harmonious ratios, and horizontality.

Northern Italy was the richest region in Europe by 1300 — urban since Roman times, at the crossroads of Flanders, Byzantium, and the Muslim world, shielded from distant rulers by the Alps and Apennines. Florence sat at the best mountain pass between Rome and the north. Wool imports from England and Spain were finished into high-quality cloth; financing that trade produced a guild that minted the gold florin, soon the dominant European currency. Decisive luck: Florence backed the Guelph (papal) faction; rival Siena backed the Ghibellines. When the papacy won, Florentine families inherited the Papal Treasury and a monopoly on alum — the only European dye mordant. The Medici leveraged these into political control and concentrated patronage in one city.

Once assembled, the critical mass was self-reinforcing: Dante, Fibonacci, Galileo, Botticelli, Machiavelli, Leonardo, Michelangelo — all Florentine or Tuscan. Whether Silicon Valley is the modern equivalent, and where the next such concentration emerges, is left open.

How Did Islam Spread So Fast?

In 25 years Muhammad went from being expelled by a village of ~550 people to his successors controlling everything from Morocco to Pakistan. The speed demands explanation.

Arabia sat between two exhausted megapowers — Byzantium and the Sassanid Empire — that had been at war since 602 AD, leaving Egypt, the Levant, and Mesopotamia exposed. But opportunity alone doesn't explain the conquest; the internal social technology matters more.

The region was locked in a clan-based local optimum: constant inter-clan raiding, expensive border policing, tribal wars. Muhammad's key move was creating the Umma — a meta-identity above clan, making an attack on one Muslim an attack on all. Apostasy was made costly (strong punishment), raising barriers to exit. Trade incentives drove this: as a trader himself, Muhammad had every reason to pacify clans so caravans could move freely.

Monotheism, borrowed from the Judeo-Christian tradition, removed one source of inter-clan conflict and provided an omniscient enforcer (Allah always sees you). Zakat — 2.5% of disposable income — pooled funds for armies. The Quran institutionalized war: 80% of booty to fighters, 20% to leadership, with non-Muslim populations taxed (jizya) rather than destroyed, keeping conquered economies productive.

Camel-based desert logistics gave Arab armies a mobility advantage no European or Persian force could match. Garrison towns (Basra, Kufa, Fustat) kept conquerors separate from locals, preventing assimilation. Polygamy channeled surplus unmarried men toward military raids; paradise promises for martyrs resolved the death-aversion problem.

When Byzantium and the Sassanids finally exhausted each other in 628, a unified Arabia with functioning taxation and standing armies hit three fertile, mountain-poor, exposed plains simultaneously. The speed was the product of unique weakness meeting well-engineered social software.

How Geography Determines Architecture

Traditional architectural styles were material consequences of local geology, climate, and political geography — not aesthetic choices. Britain's soil belt explains its vernacular: Aberdeen granite, Edinburgh sandstone, golden Bristol limestone, London brick from clay-rich Thames sediment, chalk-country timber-and-thatch — each stone dictating form until industrial canals homogenized everything to Welsh slate and brick.

Egypt's flat, rainless geography and Nile-monopoly state produced stacked-mass monuments in soft limestone and sandstone (a compressed ancient seabed), axial cosmic alignments to legitimize divine rule, flat roofs, small windows (no arch technology), and carved-wall propaganda. Permanence was the governing principle.

Greece's mountains supplied marble — harder, carvable to crystalline precision — enabling dry-jointed masonry without cement (actually stronger in earthquakes), column fluting that broke harsh Mediterranean light into vertical shadow, and sloped tile roofs absent in Egypt's rainless climate. Mountains also fragmented Greece into competing city-states, driving rapid evolution from Doric to Corinthian columns. Post-and-lintel limits meant temples were exteriors-first; congregations and sacrifices happened at outdoor altars.

How Geography Determines Architecture | Asia

India's monsoon cycle — dramatic flood, dramatic drought — made water sacred and shaped its forms: ghats descend in tiers so river access survives shifting levels; in arid Rajasthan, stepwells like Chand Baori became inverted-pyramid temples 13 stories deep; Kerala's heavy rainfall demanded tiered sloped roofs, deep cornices, and the central-courtyard Nalukettu for ventilation. China built in timber because the Yellow River's alluvial North China Plain buried bedrock deep; timber also resists freeze-thaw cracking and seismic flex better than stone. That choice cascades: load-bearing columns replace walls, long curved eaves project rain clear of the wood, and the dougong bracket system — interlocking nested timber — supports heavy roofs while flexing during earthquakes.

How Roman Technology Drove Its Architecture

Roman architectural ambitions — bigger indoor spaces, cheaper construction at scale, and indoor comfort — drove three interlocking technologies. The voussoir arch converts tension into compression, letting columns stand further apart; extended into barrel vaults and crossed into groin vaults, it opened vast interiors like Trajan's Markets. Rotating the arch produced the dome: the Pantheon's 43m span held the world record for 1,300 years. All of this depended on pozzolana cement — volcanic ash from near Vesuvius mixed with lime — which set underwater and grew stronger over time, enabling mass construction without skilled stonecutters. Lead from silver mining gave Rome ready pipe material; aqueduct pressure plus hypocaust systems brought running water and underfloor heat first to public baths, then to wealthy homes.

How Ancient Metals Started Civilization

The first three metals humans used — gold, silver, copper — all belong to Group 11 of the periodic table because all three occur in native (pure) form. Gold came first: chemical inertness leaves nuggets after erosion. Silver tarnishes by bonding with sulfur, making native silver rarer. Copper reacts most readily yet is harder and more useful — smelting was likely discovered when copper-bearing pigments (malachite, azurite) melted in pottery kilns at ~1,085°C. Copper enabled agriculture, weapons, and kingdoms.

Lead and tin, both Group 14, smelt at fire-pit temperatures. Smelting lead ore (galena) also yields silver via cupellation — the currency powering international trade. Tin, found alongside copper in shared mines, alloyed with it into bronze: harder, lower-melting, oxidation-resistant. Bronze unlocked seafaring ships, stone-cutting tools, swords, and wheeled carts. Tin's scarcity drove long-distance trade; a 2024 study argues the first civilizations (Mesopotamia, Anatolia) emerged at chokepoints between metal mines and farmland, not at the most fertile sites.

Iron, smelted by the Hittites around 1,200 BC at 1,538°C, spread despite being softer than bronze — its advantage was abundance, democratizing metal once smelting knowledge diffused.

A Science of Cities

City size before modern transport was essentially a physics problem with two inputs: calories per acre and a 30-minute walk. Food productivity set population density — Medieval English land fed 0.6 people per acre, Roman land 1.5, Chinese rice paddies 3. Agglomeration pulled people together for labor sharing, risk pooling, shared infrastructure, and defense. But Marchetti's constant capped village size: a 30-minute walk at 4 km/h covers a 2 km radius, enclosing roughly 3,000 acres — enough for at most a few hundred to a few thousand people. In practice, villages stayed at 150–300 because crowded villages shed surplus sons outward: they cleared adjacent forest or enlisted in armies to seize new farmland.

Venezuela's Excrement

Venezuela's oil — the world's largest reserves — is what made the country poor. It sits in the Orinoco Valley as heavy, sand-trapped crude requiring foreign expertise to extract. The non-oil base was always weak: nutrient-stripped savanna and jungle, with most economic activity crowded into Andean terrain where transport costs limit trade.

Nationalization in 1971–76 drove out international companies and cut production by 2 million barrels per day. Production recovered through joint ventures, then collapsed after Chávez fired 40% of oil workers in 2002, diverted investment into social programs, and re-expropriated foreign operators in 2007. When prices crashed after 2013, Maduro inherited falling volume and falling price simultaneously.

The resource curse compounded everything: oil rose from 67% to over 80% of exports as other industries were crowded out. Deficits cascaded into debt, default, and hyperinflation. Governance became kleptocracy — generals were each assigned an economic sector to loot, factories seized at gunpoint then abandoned, dollar allocations gated behind bribes. Educated elites fled, leaving the population 38% smaller. OPEC co-founder Perez Alfonzo called it in 1976: "Oil is the devil's excrement."

Saudi Arabia's Ordeal

Mohammed bin Salman faces a structural trap baked into three layers — geography, dynasty, and oil — that reinforce each other going down and collapse together going up.

The Arabian peninsula's interior (the Najd) is near-total desert with no rivers, supported only by scattered oasis settlements fed by underground wadis. Independent, water-scarce, insular settlements bred intense tribal warfare and an equally insular religion: Wahhabism, the most radically conservative strand of Sunni Islam. The Al Saud family, based in Riyadh, united Arabia between 1744 and the early 1900s by fusing secular force with Wahhabi religious zeal — an alliance still written into the flag. The conquered peripheries were always different: cosmopolitan Hejaz (Mecca, Jeddah) on the Red Sea coast, and Shia-majority communities in the east — where the oil is.

Oil generates ~95% of exports and ~80% of government revenue, and has been recycled into an unsustainable patronage state: 50% of Saudi nationals work in the public sector at 30–50% wage premiums over private-sector equivalents; 40–50% of government spending goes to wages (vs. 8% in the US); fuel subsidies run ~$7,000 per person; 60% of water is oil-powered desalination. The military ($60B/year, bigger share of GDP than Russia) is deliberately hobbled — split across rival princes — to prevent coups, which is why it couldn't beat the Houthis.

Convergent pressures: US oil self-sufficiency reduces the security guarantee; Iran backs Shia proxies on every border; the Wahhabi culture that cemented Al Saud power is precisely what blocks the economic modernization oil decline makes necessary.

Dubai

Dubai escaped the resource curse not by diversifying away from oil, but because its trading institutions were strong before oil arrived. In 1901–1902, the Al Maktoum family turned the creek into a proto-Special Economic Zone — zero taxes, free land, safety, religious tolerance — drawing merchants fleeing overtaxed Ottoman Basra and Persian Lingeh. By 1906, Dubai had displaced Lingeh as the Gulf's main port. Population quadrupled 1900–1950 despite the collapse of the pearl industry (killed by Japanese cultured pearls and the Great Depression).

Oil arrived in 1966, but Sheikh Rashid had already built the airport, telephone network, and electricity grid ahead of revenues. The windfall funded Port Rashid (1972), Jebel Ali port (1979), and Emirates Airline. Oil now generates under 1% of GDP.

The governance formula — 2nd most tax-friendly city globally, 5th safest worldwide, and practical social tolerance — continues to attract millionaires and multinationals. Dubai's lesson is that governance, not natural endowment, built the city; oil accelerated what pre-existing institutions had set in motion, making it a poor template for Gulf neighbors like Abu Dhabi whose institutions were built around oil rents.

The Future of Petrostates After Oil

When oil demand peaks and collapses over the coming decades, petrostates face state failure — and most are structurally unprepared. Oil-rich governments are 2–3x larger than non-oil governments at equivalent income; wealth concentrates at extractable points; whoever controls those points rules without needing taxpayer accountability. Corruption, suppressed private sectors, and citizens on handouts follow. When revenue dries up, security funding collapses and civil war opens.

Petrostates will respond in sequence. First, denial — OPEC still projects rising demand through 2045. Second, delay via lobbying, greenwashing, and fossil fuel subsidies. Third, a race to pump rather than coordinate cuts, since no cartel holds when members know their assets are depreciating. Fourth, diversification: Norway's sovereign wealth model (spend only 3% annually) is the template; Saudi Arabia's PIF and Gulf AI commitments ($240bn to data centers and chips) attempt to replicate it. Kuwait's case shows the limits — it would need a 1,000% tax-base expansion to replace oil revenues, requiring generational change. Fifth, asset sales: Kuwait's $7bn pipeline stake, Aramco's $64bn privatization program. Sixth, cutting benefits. Kuwait's emir has already stripped tens of thousands of citizens of citizenship. That triggers revolts.

Political outcomes split three ways: some states democratize after peak production; most with Libya- or Iraq-level oil wealth tighten authoritarian control; many collapse into civil war. Muslim-majority countries control 51% of oil exports but only 23% of nations, so Islamic soft-power funding — mosques, Islamist parties, Al Jazeera — shrinks with revenues. Russia loses superpower reach as energy covers 30% of federal revenue. China is the primary winner: most constrained by oil dependence today, leading solar and battery manufacturing, and freed from US maritime chokepoint leverage. Europe normalizes energy costs.

Peak Oil Is Coming

Peak oil arrives in the early 2030s — not from running out of crude, but from EVs, heat pumps, and electric arc furnaces eating demand from below. Solar's electricity share ran 13.5% in 2023, 15% in 2024, 18% in 2025; China's supply chain now produces 1 TW of panels per year, equal to 10% of all global installed capacity. EV sales in China have crossed 50%; Indonesia is on track from 20% to 80% EV share in under two years; Pakistan went from under 1% to over 10% solar in five years. Standard forecasts miss this by assuming linear electrification — OPEC projects rising oil demand through 2045 by assuming ICE cars spread through poor countries. By 2050, petrostate budgets face structural collapse, with geopolitical consequences for Russia, Saudi Arabia, and Venezuela still to be worked out.

Why Warm Countries Are Poorer

Warm countries are poor not primarily because of heat, institutions, or colonial history, but because equatorial populations were forced into mountains to survive — and mountains impose punishing costs.

Each additional degree Celsius shrinks GDP per capita by 8.5%; humidity compounds this by making sweat useless for cooling. Tropical diseases (malaria, dengue, parasitic worms) killed and incapacitated workers in hot lowlands. The colonial-institutions theory (Acemoglu, Johnson and Robinson) tries to explain poor institutions through disease killing European settlers, but its mortality data predate actual African colonization, and hypercolonized Singapore, Hong Kong, and Taiwan rank among the world's richest places.

The underdiscussed mechanism: humans evolved in African highlands at moderate, stable temperatures. Near the equator, sea-level lowlands are lethally hot, humid, and disease-ridden, so populations migrated upward — Bogotá sits at 2,600 m, Mexico City at 2,200 m, Addis Ababa on the Ethiopian highlands. Equatorial capitals are consistently higher than temperate ones. Mountains delivered cooler temperatures and lower humidity. But the price was steep: rugged terrain kills navigable rivers, multiplies infrastructure costs, collapses trade, and fragments populations into distrustful valley communities. That fragmentation breeds ethnic conflict and makes durable institutions nearly impossible.

The prescription follows directly: electrify and air-condition low-lying tropical areas; invest heavily in transport infrastructure in mountainous ones.

A Trip Around the World of Heat, Mountains, and Poverty

Warm countries are poor not simply because of heat, but because equatorial heat produces constant rainfall that leaches soil and breeds disease — civilizations that escaped did so by moving uphill or achieving unusual fertilization.

The Malthusian objection — India and Cambodia built empires while hot, so heat can't explain poverty — dissolves once you separate GDP from GDP per capita. Pre-industrial wealth converted to population not living standards, so large tropical empires are compatible with low per-capita poverty.

Seventy-plus maps across four continents show the same pattern. In Latin America, every major historical city sits on mountains: Bogotá at 2,600 m, Caracas at 900–1,400 m, Cusco, Mexico City. The Amazon — a flatland rivaling the Mississippi — is nearly empty because constant rain stripped its soil for millennia. The Darien Gap is the only break in the Pan-American Highway; the terrain defeated a French canal attempt that killed 22,000 workers and cost $287 million. The Aztecs and Incas emerged on high plateaus; the Maya held drier, northerly ground and never unified into a large empire.

Africa repeats the logic. Egypt thrives because the Nile fertilizes a desert that never leaches. The Congo — the Amazon of Africa — supports no major civilization despite its rivers. Trypanosomiasis (tsetse fly) kills horses and cattle across equatorial Africa, compounding malaria. Nigeria's Niger Delta is an exception: seasonal rain and river-sediment fertilization.

Java at 1,100 people per km² versus Borneo at 37 is the sharpest outlier: volcanic ash restores what rainfall strips. India's Ganges Valley gets monsoon sediment then a dry season that breaks the leaching cycle. Southeast Asian empires (Khmer, Siam) emerged later and smaller, reflecting harder conditions.

The synthesis: pre-industrial poverty was leaching and disease blocking agriculture; post-industrial poverty is heat and disease reducing productivity — continuous, but still latitude-tracking.

Argentina Could Be a Superpower

Argentina's geography mirrors the United States almost point for point, yet the country is poor — making it one of history's great squandered endowments. At 2.78 million km², it contains France, Spain, Italy, Germany, and twenty more European countries. Three sides are sealed by ocean, the Andes, and Antarctica; its only serious threat is Brazil, now 3.5x larger by GDP. The Río de la Plata basin — the world's second-largest interconnected navigable waterway — replicates the Mississippi: flat Pampas soils formed by the same geological process as the American Midwest, similar latitudes and rainfall patterns, and water transport 10–30x cheaper than roads, which alone can multiply trade 16-fold. The US and Argentina rank 2nd and 3rd globally in food exports by tonnage. Argentina also sits on the same Andean mineral belt as Chile but has barely extracted it. At 16.8 people per km², it remains far below US density (37) or Germany's (240). Geography is hardware; institutions are software — and Argentina has run bad software on exceptional hardware.

Why Is Argentina Poor?

Around 1900 Argentina was twice as rich as Japan and four times richer than South Korea, Taiwan, and China — all of which have since overtaken it. Agricultural wealth generated the wrong politics and destroyed the chance to industrialize. These countries followed identical development playbooks but diverged at a few crucial points.

The original sin was land concentration. Asian Tigers redistributed farmland after WWII, creating owner-farmers who saved and reinvested. Argentina's hacienda structure survived; political anger channeled into financial redistribution: Perón's IAPI bought beef and grain at 40–60% below world prices and pocketed the spread, while Asian governments recycled surpluses into farms.

Industrial policy repeated the error. Asian Tigers gave cheap loans and protection only to firms that proved export competitiveness — filtering out rent-seekers. Argentina's import substitution had no such condition: firms stayed sheltered, never achieved scale, kept prices high. Car costs ran 60–150% above US levels; electronics subsidies consumed 0.8% of GDP for an industry with no Argentine advantage; agro-machinery, where Argentina had real edge, was ignored.

Peronist trade unions, centralized under one federation (CGT), pushed real wages up 50% between 1946 and 1948 — Asian unions were company-level and tied to firm survival, Argentina's sector-wide and politically dominant, keeping costs above productivity.

Commodity booms created a financial trap: dollar inflows were spent rather than sterilized, overvaluing the peso and creating deficits covered by printing money when prices fell. The cycle — boom, overspend, bust, print, inflate, default — has run nine times since independence; deficits hit 14% of GDP at the 1975 peak, with taxes covering only 20% of spending.

The asymmetry with Asia traces to starting conditions: Tigers emerged from mid-century wars dirt-poor, governments able to suppress wages long enough to build export competitiveness. Argentina came from wealth and inequality — redistribution irresistible, fiscal discipline impossible.

When Will We Make God?

Hyperscalers are spending hundreds of billions on AI because they believe AGI — an AI capable of executing multi-week knowledge tasks autonomously — is only years away. The strategic logic pivots on one point: automate AI researchers first. Because AI research is purely virtual and systematizable, replacing researchers accelerates AI's own improvement, compressing the path to superintelligence.

The empirical case rests on scaling laws (Kaplan et al., 2020) showing test loss drops linearly across seven orders of magnitude of compute, and on task-length benchmarks (Kwa et al., 2025) where AIs complete progressively longer autonomous tasks at constant 80% success rates. RE-Bench (2024) shows current models already outperform human researchers on sub-4-hour tasks. "Humanity's Last Exam," designed to remain relevant for years, went from 3% to 32% accuracy within one year.

CEO timelines cluster 2026–2028: Amodei, Altman, Musk, Pichai. Metaculus prediction markets place weak AGI by late 2027. A structural constraint sharpens the window: both investment ramp-up and compute optimization gains are near their peaks — making the next few years decisive, or else progress stalls for decades.

AI: How Do We Avoid Societal Collapse and Build a Utopia?

The path to AI-driven abundance runs through a dangerous transition: mass unemployment concentrated in 90% of the workforce while a small class of "automators" — entrepreneurs and developers who scale automation globally — captures enormous wealth. Those generating the surplus are mobile and will relocate to low-tax jurisdictions (UAE, Singapore, Caribbean SEZs), while the unemployed majority demands redistribution.

Full post-scarcity still requires capitalism's price signal. Money functions as votes directing AI resource allocation across energy, land, and raw materials — physical goods are harder to produce than digital services because atoms move slower than bits. Once human labor is no longer the basis for claiming resources, distribution must shift to Universal Basic Income. At $1,000/month in a world where lunch costs $1 and a transport ride costs $0.01, UBI becomes livable.

Two surviving job categories behave differently: local service workers (nurses, electricians) whose wages must stay above UBI to attract entrants, and automators whose fortunes must remain intact to sustain competitive pressure driving costs down for everyone. Taxing robots is doubly counterproductive — definitionally unenforceable and it slows exactly the automation that reduces cost of living.

The sustainable tax base is immobile capital: land via Harberger taxes, transportation hubs, sovereign wealth funds invested in AI equity. Tyler Cowen's insight proves correct — we're already drifting toward de facto UBI through disability payments, child credits, and healthcare entitlements, just inefficiently. The transition stabilizes only if governments simultaneously drive down costs in housing, healthcare, and education while entitlements converge upward and automators keep getting rich enough to attract imitators.

The $100M Worker

AI algorithms have improved roughly 0.4–0.5 orders of magnitude per year since 2012 — a 22,000x efficiency gain across all fields by 2023. Combined with hardware improving 2.5x annually and investment doubling each year, AI effective compute should reach ~3,000x by 2030 from hardware alone, plus ~300x from continued algorithmic optimization.

On top of incremental gains, a series of paradigm-shift "unhobblings" delivered 3–100x jumps each: supervised instruction fine-tuning taught LLMs to answer questions; RLHF corrected them through human preference; chain-of-thought forced step-by-step reasoning (yielding 10x+); distillation lets cheap models inherit expensive models' intelligence; mixture-of-experts routes queries to specialized sub-models; scaffolding and multi-agent frameworks coordinate all of the above. Each unhobbling compounds on the others. This is the logic behind $100M compensation packages: at this trajectory, a researcher who unlocks the next major unhobbling generates returns that make nine-figure salaries rational.

Not Sustainable ➡️ Abundant

"Sustainable" is defeatist — it means "don't touch what we don't understand." Earth has never been stable: cyanobacteria catastrophically depleted CO2, sea levels swung 120 meters, five mass extinctions came and went. Human intelligence now lets us engineer that dynamic system rather than merely preserve it — more food on less land, more GDP per watt, compute doubling every two years. Fish depletion is a wild-fish problem; water scarcity is an aquifer problem, not a drinking problem. The goal should be abundance. The solution to pollution is profusion.

Can We Build AI in Space?

Space datacenters are already cost-competitive with land-based ones, and Starship economics may tip them cheaper. Five engineering factors converge.

Orbital solar panels produce 5x more energy than ground installations — no night cycle, no atmospheric absorption — so far fewer panels are needed by mass. Continuous sunlight also eliminates batteries, which represent over half of ground-system weight. Space panels weigh ~1 kg/m² versus 10+ kg/m² on Earth, needing no glass, frames, or weather hardening.

Radiation shielding — normally heavy and expensive — is largely avoidable for AI. Traditional computers fail catastrophically when cosmic rays flip bits. AI models are probabilistic: random bit-flips across billions of parameters cause negligible drift. Google tested its Trillium TPU under proton beams and found it radiation-hard enough for five-year missions with minimal shielding.

Cooling, the apparent showstopper, resolves through the Stefan-Boltzmann T⁴ law. Radiative heat transfer scales with the fourth power of temperature; both faces of the solar panels passively radiate enough to cool the entire system at ~60°C. GPUs run at ~97°C, within industrial-silicon tolerance — no liquid cooling, no heavy radiators.

SpaceX + xAI Merger, IPO, Data Centers and Moon Base

Starship is about to create cargo capacity far beyond what Starlink can fill — the same overcapacity trap that hit early electricity companies (10% load factors) and railroads. Musk's answer: xAI datacenters in orbit. Energy permitting will throttle AI compute on Earth before GPU supply does; space solar (no atmosphere, no night, no permits) beats ground cost within three years.

The merger is vertical integration: Starlink's projected $100B free cash flow by 2030 funds xAI's buildout while SpaceX fills Starship. xAI burns $1B/month; Colossus 2 costs $44B; datacenter spend doubles annually toward $700B/year against $0.5–3B revenue. An IPO at ~$1.5T targets ~$50B to bridge the gap.

The Moon pivot is IPO optics: Mars earns no revenue and is indefensible to public shareholders; the Moon carries $4B in NASA contracts. Musk has quietly rewritten SpaceX's mission from "multiplanetary humanity" to "extend consciousness to the stars" — consciousness now assumed to be artificial.

The Cold Start Problem of New Cities & States

New cities face the same network-effects trap as digital marketplaces: nearly worthless when small, enormously valuable when large, with every order of magnitude of growth equally hard. Going from 0 to 100 residents is as difficult as 100 to 10,000 — jobs, amenities, and dating pools all require prior critical mass.

Dubai and Singapore are the canonical models, but both took a century of compounding from one seed: a port with tax-free trade and reliable law. Shenzhen, Hong Kong, and Shanghai followed the same Special Economic Zone pattern. Most new cities lack a port and need different tactics.

The demand side runs through a funnel. Short-term events create awareness. Medium-term popup cities (Edge City ran eight villages, 12,000 participants) deepen friendships until people want to settle. Balaji Srinivasan's Network School near Singapore solved the longer step: all-inclusive hotel residency packaging the college experience for nomads, converting visitors to permanent residents.

The supply side — jobs — is harder. Companies won't relocate without large regulatory arbitrage. The best targets are healthcare, real estate, and finance. Prospera, a Honduran SEZ on Roatan island, focuses on healthcare: regulations let Western drug companies run trials illegal at home — drug repurposing, aging-as-disease research — bypassing the hundreds of millions a US clinical trial costs.

The fastest shortcut is bypassing the cold start: a satellite city near a metropolis (California Forever targets hardware startups priced out of the Bay Area), or an anchor tenant bringing hundreds of jobs and workers at once. Tourism settlements — how Mexico seeded Cancun from jungle in the 1970s — fund early growth via hotel workers before diversifying.

One underrated variable: early new cities run 60–80% male, discouraging women and self-reinforcing the gap. The French-English colonial imbalance in Canada, rooted in the same dynamic, determined which language North America speaks.

Do Women Reproduce more than Men?

Historically, only about 25% of men reproduced versus ~80% of women — a 3:1 gap confirmed by genetic analysis. Polygyny let some men monopolize women, excluding the rest. Patrilineal clan structure compounded this: when a clan was wiped out, its entire male line went extinct while female DNA dispersed widely. Population growth enabled a third mechanism — older men marrying two younger women per generation, sustaining polygyny without a raw numbers shortage. Societies managed the surplus of excluded men either through outward raiding (Vikings, Arab Muslims promised women in heaven) or monogamy (Greeks, Romans), the latter redirecting male energy from violence to wealth creation. DNA from Iceland and the Swahili Coast confirms the pattern at continental scale.

The Future of Fertility

Falling fertility is rational: children are less economically necessary now that states provide retirement security, childrearing remains painful, and alternative pleasures have multiplied. But those costs are about to collapse at every stage.

Conception: In-Vitro Gametogenesis converts any cell into a gamete, generating millions of embryos; companies like Herasight select for genetic quality, letting parents trade off IQ against disease risk. Same-sex couples can split DNA 50/50; groups could co-parent children sharing fractional DNA from multiple members.

Pregnancy: artificial wombs already function at both ends — past day 13 in labs, viable from week 24 in NICUs. Closing the gap (weeks 2–24) is engineering, not impossibility.

Early childcare, later childcare, education, and family coordination all fall to humanoid robots and AI tutors — what rich families outsource to employees, now democratized.

Counterintuitive claim: chores aren't the source of attachment. Children forget nannies; parental abandonment marks them for life. The bond is intrinsic.

Result: people priced out of parenthood will enter, existing parents will have more, and a child-loving minority will scale to dozens or hundreds. The binding constraint shifts to land and the Dunbar number (~150) — and that minority alone could reverse the fertility collapse entirely.

The Criminal, Social, and Economic Cost of Immigrants in the EU

Non-EU Muslim immigrants to Europe impose measurable costs across crime, employment, welfare, and social cohesion — costs that persist into the second generation and are not explained by poverty alone.

On crime, Germany's 2025 data shows Algerians committing crime at 26x the German rate; 13 other Muslim-majority nationalities exceed 5x. Denmark, Sweden, Spain, Norway, the Netherlands, and the UK all show the same pattern: Muslim-origin nationalities dominate crime rankings even after adjusting for age and sex. A Finnish cash-transfer experiment found income had zero effect on crime rates. A 1992 Danish cohort of Palestinians produced 64% criminal records in the first generation and 34% convictions in the second. Islamism accounts for 75% of global terrorism deaths and 67–90% of UK terrorism incidents.

On economics, Muslim women in Europe show employment rates around 30%. Denmark spends 1.3% of GDP annually on MENAPT immigrants; at no age do they contribute positively in net tax terms. The Netherlands' lifetime analysis found only immigrants of Danish, Swedish, Finnish, and Chinese origin net-positive across two generations; the average Dutch family pays €5,000 a year toward immigrant costs.

The piece withholds a deportation conclusion, noting host-country legislation is partly responsible and a follow-up will address selective reform.

What Do Muslim Immigrants Think in the West?

Muslim immigrants in the US are substantially better integrated than those in Europe. American Muslims are pious — 65% pray daily, roughly a third believe the Quran has only one valid interpretation — yet they broadly support democracy, identify as American, and oppose extremism more strongly than the general public. A sizable minority (30%) thinks Islam and democracy are incompatible, but US Muslims condemn targeting civilians at higher rates than non-Muslims.

Europe is a different story. Across UK, France, Denmark, Germany, and Austria, religiosity is rising rather than falling across generations. In France (7% of population), the share of young Muslims who consider religion superior to science jumped from 19% to 81% since 1998; the share wanting Islam to modernize fell from 48% to 21%; hijab-wearing nearly tripled; and 24% support the Muslim Brotherhood. In the UK, 38–42% want some Sharia elements introduced, and British Muslims are 2–3x more likely than the general public to believe Jews have too much power. Only 4% of British Muslims believe Al Qaeda carried out 9/11. British Muslims support authoritarian regimes (China, Russia, Iran) at rates 44–64 points above the general public.

In Germany, nearly half of Muslims hold at least latently Islamist attitudes, with young Muslims four times more radical than those over 60. Danish data shows 75% think the Quran should be interpreted literally (up since 2006), and 44% believe society bears some responsibility for violence against Muhammad-drawers.

The overall picture: roughly 20–50% of European Muslims are well-assimilated; a similar share holds fundamentalist views incompatible with liberal democratic norms; and 10–25% express support for Hamas, Jihad, or terror organizations — roughly 8 million people. Moderate Muslims broadly condemn extremism and represent the key integration bridge, but trends in France and Germany run toward more radicalism in younger cohorts, not less.

Islamismophobia

Conflating Islam with Islamism lets Western societies simultaneously persecute a religion and fail to defend themselves against a political movement. Islam is a personal faith followed by two billion people, protected by the Universal Declaration of Human Rights. Islamism demands Sharia supersede civil law and extend to non-Muslims — goals the European Court of Human Rights has ruled incompatible with democracy. Five questions separate them: voluntary or coerced? For the believer alone or imposed on others? Civil law or Sharia supreme? Equal citizens or Muslim privilege? Persuasion or intimidation?

In the West, roughly 20–50% of Muslims are moderate and explicitly reject Islamism; about 15–20% hold Islamist positions; a conservative middle sits between. Saudi Arabia, Egypt, Jordan, and the UAE have banned the Muslim Brotherhood as a terrorist organization. Gulf commentators warn this is a failure their own countries learned to name.

Islamist infrastructure is organized and Qatar-funded. The Muslim Brotherhood runs 60+ schools in France, trains imams across Europe, and created the European Council for Fatwa and Research to issue segregation-promoting rulings for European Muslims — leaving second-generation Western Muslims measurably more radicalized than their parents.

The strategic weapon is collapsing the distinction between islamophobia (bigotry against a religion) and islamismophobia (criticism of a political movement). The MB-linked Muslim Council of Britain spent 80% of its charity budget monitoring British media for "islamophobia," pressured the BBC to withdraw clips, and trained journalists at The Independent and Scottish TV. Consequences: Rotherham councils suppressed ~1,400 rape cases to avoid racism accusations; a Manchester Arena security guard didn't stop the bomber for the same reason.

The corrective is symmetrical: call out Islamophobia where it targets religion; call out Islamismophobia where it targets political coercion. Moderate Muslims, conservatives, liberals, and politicians each need to hold that line — or they end up serving Islamism.

How Did Islam Spread So Fast?

In 25 years Muhammad went from being expelled by a village of ~550 people to his successors controlling everything from Morocco to Pakistan. The speed demands explanation.

Arabia sat between two exhausted megapowers — Byzantium and the Sassanid Empire — that had been at war since 602 AD, leaving Egypt, the Levant, and Mesopotamia exposed. But opportunity alone doesn't explain the conquest; the internal social technology matters more.

The region was locked in a clan-based local optimum: constant inter-clan raiding, expensive border policing, tribal wars. Muhammad's key move was creating the Umma — a meta-identity above clan, making an attack on one Muslim an attack on all. Apostasy was made costly (strong punishment), raising barriers to exit. Trade incentives drove this: as a trader himself, Muhammad had every reason to pacify clans so caravans could move freely.

Monotheism, borrowed from the Judeo-Christian tradition, removed one source of inter-clan conflict and provided an omniscient enforcer (Allah always sees you). Zakat — 2.5% of disposable income — pooled funds for armies. The Quran institutionalized war: 80% of booty to fighters, 20% to leadership, with non-Muslim populations taxed (jizya) rather than destroyed, keeping conquered economies productive.

Camel-based desert logistics gave Arab armies a mobility advantage no European or Persian force could match. Garrison towns (Basra, Kufa, Fustat) kept conquerors separate from locals, preventing assimilation. Polygamy channeled surplus unmarried men toward military raids; paradise promises for martyrs resolved the death-aversion problem.

When Byzantium and the Sassanids finally exhausted each other in 628, a unified Arabia with functioning taxation and standing armies hit three fertile, mountain-poor, exposed plains simultaneously. The speed was the product of unique weakness meeting well-engineered social software.

Iran: The Day After

Iran's Islamic regime cannot survive without hostility to the US and Israel — that antagonism is its source of legitimacy. After killing Khamenei and dismantling Iran's regional proxies since October 2023, Israel and the US see a window for regime change. But aerial strikes have never toppled a government, and Iran's geography rules out invasion. Power runs through Khamenei's Bayt — a 4,000-person shadow government — plus the IRGC (125,000 members) and Basij militia (400,000 volunteers), all structured to survive decapitation.

Collapse requires simultaneous pressure: sustained bombing of IRGC leadership, oil-revenue strangulation (~45% of government income), military defection, and ethnic minorities — Azeris, Kurds, Baluchis — fragmenting the state. The real danger is Syrian-style civil war. Iran's missile strikes on Gulf airports have pushed the Arab states toward Israel. China watches its anti-US coalition disintegrate and loses 17% of its oil imports. A post-regime Iran — 90 million people, 98% literacy, wealthy diaspora — could follow South Korea's trajectory if the transition avoids civil war.