The Electric Car Case

In the AI Story

The specific forces that killed the early electric car are worth cataloguing because they reveal the texture of how selection environments actually operate. The petroleum industry, already established to serve the kerosene lighting market, had built a nationwide distribution infrastructure that electric vehicles could not match — there were gasoline pumps everywhere and charging stations almost nowhere. The cultural narrative associated loud engines with power, adventure, and masculinity; the quiet, domestic electric car appealed to a demographic the early-automobile marketing apparatus did not target. Henry Ford's mass-production techniques reduced the Model T to a price that electric vehicles, burdened by expensive batteries, could not approach. And the self-reinforcing dynamics of adoption took hold: more gasoline cars sold meant more fueling stations built, which made gasoline cars more convenient, which increased sales further.

Each step in the process was rational for the individual actor. The consumer choosing gasoline in 1915 was choosing the more convenient option given the existing infrastructure. The infrastructure investor building more gas stations was rationally serving the existing market. The manufacturer scaling gasoline production was rationally responding to demand. The cumulative effect of individually rational choices was the extinction of the technically superior alternative — the classic dynamic of path dependence operating at scale.

The electric car's eventual return in the twenty-first century did not occur because battery technology improved, though it did. It occurred because the selection environment changed. Environmental regulation penalized gasoline vehicles. Cultural attitudes toward sustainability shifted. The specific pattern of urban driving created a niche — short-range, low-speed, frequently recharged — that electric vehicles fit better than their gasoline competitors. The variation (the electric vehicle) had persisted, in marginal forms, across the intervening century. What changed was the environment that received it.

Applied to artificial intelligence, the electric car case carries several lessons. First, technical superiority is no guarantee of survival. The AI systems that dominate the next decade will be the ones that fit the selection environment, not necessarily the ones that perform best on benchmarks. Second, the selection environment operates at multiple layers — infrastructure, culture, economics, regulation — and each layer can veto an otherwise excellent technology. Third, path dependence can lock in inferior outcomes for long periods, and only shifts in the selection environment itself can break the lock-in. Fourth, the eventual displacement of dominant AI systems will come not from incrementally better alternatives but from selection-environment changes that create new niches the incumbents cannot fill.

Origin

The case is discussed in chapter 5 of The Evolution of Technology as part of Basalla's broader treatment of the selection environment. It has since become a canonical example in the sociology of technology and the history of transportation.

Key Ideas

Technical superiority does not guarantee survival. The electric car was the better technology by most measures and still lost.

Infrastructure lock-in is decisive. Existing fuel distribution networks created an unbreakable advantage for gasoline.

Cultural narrative matters. The association of loud engines with masculine power shaped consumer preferences in ways that had nothing to do with functional performance.

Individually rational choices can produce collectively irrational outcomes. Each consumer's rational choice reinforced a lock-in that left everyone worse off.

Only environmental change breaks lock-in. The electric car's return required shifts in regulation, culture, and urban structure — not merely technical improvement.