War as Technological Accelerant

Almost every major technological discontinuity in this archive runs through a war. The pattern is not romantic. Wars concentrate state capital, suspend peacetime budget discipline, mobilize talent on deadlines, and reward results over publications. Whether one approves of the mechanism or not, the historical record is consistent enough that any complete account of technological centers must reckon with it.

Greek fire (Byzantine, ~672 CE) keeps Constantinople alive for an extra eight centuries against Islamic and Crusader pressure. Gunpowder weapons (Song China → Mongol diffusion → European refinement → Ottoman cannon at Constantinople) end the medieval defensive paradigm in 1453. Sailing-ship gunnery (Iberian–Dutch–English) opens the Atlantic empires. The Napoleonic wars institutionalize the École Polytechnique. The American Civil War industrializes mass production. World War I produces radio, aviation, and chemical industry at scale. World War II produces radar, the jet engine, nuclear weapons, computers, antibiotics, and the operations-research methods that became modern management.

The specific lesson of the 1940s war economies is the most relevant for this archive. The Manhattan Project, the Radiation Lab at MIT (radar), Bletchley Park (cryptography), and the OSRD-coordinated medical work (penicillin scale-up, blood plasma logistics) all share a common pattern: identify a small set of decisive technical problems, give them a national-priority budget, recruit talent without regard to seniority, give the program a deadline that mirrors enemy capability, and accept that 90% of the spending will produce nothing while the 10% that succeeds will reshape civilization.

The pattern persisted through the Cold War (DARPA, NASA Apollo, the semiconductor industry's military origins) and is reactivating in the 2020s through chip export controls, sovereign-AI initiatives, and defense-tech venture capital. Anduril, Palantir, SpaceX-as-defense-contractor, OpenAI's 2024 federal partnership — all are recognizable as Cold-War-era institutional types in 21st-century form.

The uncomfortable corollary, much-discussed by economic historians, is that the absence of major war between great powers since 1945 may be one of the reasons technological progress has slowed in the categories where governments traditionally led — energy, transport, basic materials. The argument has a moral dimension nobody wants to engage with. The empirical pattern is what it is.

The risk for the 2020s and 2030s is that the pattern reactivates more literally than anyone wants. The U.S.–China AI competition is currently being run, by both sides, on the institutional templates of the Manhattan Project and the Cold War. If those templates produce outcomes similar to what they produced before, the technological gains will be substantial. If they produce outcomes similar to what they produced in 1914, the technological gains will be irrelevant.