After the Rupture: Four Turning Points, Four Choices

Karl Jaspers coined a term in The Origin and Goal of History (1949): the Axial Age. Between roughly 800 and 200 BCE, Greece, Israel, India and China — with no knowledge of one another — produced near-simultaneous breakthroughs asking how a human being ought to live. The common soil was an old order coming apart. Athens had lost the Peloponnesian War in 404 BCE and endured the bloody rule of the Thirty Tyrants; it wanted someone to blame. Socrates (469–399 BCE) was the old man who went around exposing everyone's ignorance and had taught a few men who later turned on the city.

In 399 BCE he was charged with impiety and corrupting the young, and a jury of 501 convicted him by roughly 280 to 221. Athenian procedure let the defendant propose an alternative penalty; exile would almost certainly have been accepted. Instead he proposed that the city maintain him at public expense for life — and more jurors voted for death than had voted to convict. In prison his old friend Crito had already bribed the guards and prepared an escape route. He refused: if everyone may defy whichever verdict they dislike, law ceases to exist. At the other end of the same century, Confucius (551–479 BCE) spent fourteen years travelling between states in search of office, failed, returned to Lu in 484 BCE and turned to editing the classics and teaching disciples. Two ways of preserving a body of thought after political failure: one nails it into a city's memory, the other hands it to texts and pupils.

Whistleblowers face the same question: leave, or stay and absorb the punishment so the system's absurdity is on the record. Whether it lands depends on someone documenting and transmitting it — Socrates had Plato.

In October 1347 Genoese ships carried plague from Caffa on the Black Sea into Messina in Sicily; within two years it had swept Europe. It reached England in the summer of 1348, and within a generation the population fell from roughly 4.8 million to about 2.5 million. King Edward III (1312–1377) faced a very concrete problem: the surviving peasants and craftsmen were suddenly scarce, demanded two or three times the old wage, and landlords began poaching each other's labour.

The crown's answer was legislation: the Ordinance of Labourers (1349) and the Statute of Labourers (1351) required every landless person under sixty to accept work, froze wages at pre-plague 1346 levels, and jailed those who demanded more — an administrative decree set squarely against supply and demand. Court rolls are thick with prosecutions, which shows it was enforced for three decades and equally that it never worked. The pressure broke out in 1381 in the rising led by Wat Tyler, whose followers made a point of burning legal records in London. By the fifteenth century English serfdom had effectively dissolved and real wages reached a peak not seen again before the Industrial Revolution.

What automation and AI do to labour is not settled by the technology. The same wave of capability lands one way on organised professionals who can bargain and the opposite way on gig platforms. The shock is exogenous; the distribution is not.

Santa Maria del Fiore was begun in 1296. By 1418 the body of the cathedral stood finished with a hole 43 metres across and 55 metres above the floor at its centre, and for over a century nobody knew how to close it. The standard method was full timber centering to carry the masonry, but Tuscany had neither timber that long nor money on that scale. In 1418 the Wool Guild offered 200 gold florins for a workable proposal. Filippo Brunelleschi (1377–1446) was a goldsmith by training with no record of building cathedrals at all.

His scheme used no centering whatsoever: two shells to cut weight, brick laid in an interlocking herringbone (spinapesce) pattern, four stone and timber chains to contain the hoop stress, the structure carrying itself course by course. Fearing theft of the idea, he refused to disclose the details. The board therefore faced a pure bet under uncertainty: stake public funds and the city's face on a man with no precedent who would not explain himself. Their handling is the interesting part — not a single all-or-nothing wager but a bet broken into stages: he was appointed jointly with his rival Ghiberti, with the work verified as it rose, until Ghiberti was sidelined in practice around 1423. The dome was consecrated on 25 March 1436 and remains the largest masonry dome in the world.

When you commit resources to an unproven technical path, the mature move is not to argue about whether the design is right, but to build milestones that expose failure early — slicing one unevaluable bet into small observable ones.

In 1972–73 Paul Berg, Herbert Boyer and Stanley Cohen produced recombinant DNA. Berg had planned to insert genes from the monkey virus SV40 into E. coli — a bacterium that lives in the human gut. In July 1974, Berg and ten colleagues published a letter in Science calling for a worldwide voluntary moratorium on several classes of high-risk experiment: a deliberate halt in their own most promising field, which is rare in modern science.

From 24 to 27 February 1975, about 140 scientists met for four days at the Asilomar conference centre in California and reached agreement only on the final night: risk-tiered categories with matching physical and biological containment (using enfeebled strains that cannot survive in the human gut), on which basis the moratorium was lifted; the NIH turned it into guidelines in 1976. The price was an agenda deliberately narrowed — eugenics, biological weapons, patents and commercialisation were all kept off the table. Only by confining debate to laboratory safety, where scientists held genuine expertise, could consensus be reached and congressional legislation forestalled. Forty-three years later, on 26 November 2018, He Jiankui announced the birth of CRISPR-edited twins, faced his peers two days later at the human genome editing summit in Hong Kong, and was subsequently sentenced to three years for illegal medical practice.

The 2023 open letter calling for a six-month pause on AI runs the same script: quantifiable safety risks are easy to write into a consensus, while concentration of power, employment and distribution get deferred. The He Jiankui case adds a second rule — the real strength of self-regulation equals the floor set by its least compliant member.