in 1889 thomas edison wrote something that now reads strangely.

people were dying.

wires fell into streets. insulation failed. workers were electrocuted. high voltage could turn an ordinary piece of infrastructure into something lethal.

edison understood the engineering. he had evidence. he had bodies.

and he arrived here:

“my personal desire would be to prohibit entirely the use of alternating currents.”

not regulate it.

not improve insulation.

not invent better breakers.

not establish electrical codes.

prohibit it.

i think we normally tell this story badly.

we tell it as edison versus tesla. dc versus ac. a clever historical fight where one technology happened to win.

but ask the more interesting question.

what would it have cost humanity if edison had won?

not financially.

humanly.

because in 1889 ac was dangerous for almost exactly the reason it was useful.

voltage.

edison’s low-voltage dc system could only economically distribute electricity a short distance from the generating station. the national park service puts the practical range of his pearl street system at roughly half a mile. ac changed the geometry of electricity because transformers allowed voltage to be raised for transmission and lowered again near the person actually using it.

that sounds like an engineering detail.

it wasn’t.

it determined whether electricity was a local curiosity or infrastructure.

with dc, the power station has to come to you.

with ac, the power can come from somewhere else.

twenty-six miles separated niagara falls from buffalo when westinghouse’s ac system began transmitting power there in 1896. by 1902, niagara’s ac plants were producing about one fifth of all electricity generated in the united states.

seven years.

that is the distance between:

“this system should be prohibited”

and

one hydroelectric installation producing a fifth of america’s electricity.

now run the counterfactual.

ban ac in 1889.

cities do not suddenly become permanently dark. humans are annoyingly inventive. dc continues. more generating stations are built. more copper is used. someone probably eventually invents better ways of transmitting dc.

but everything becomes harder.

more expensive.

more local.

more generation has to sit close to consumption.

hydroelectric sites far from cities become substantially less useful. rural electrification becomes harder. large interconnected grids become harder. every additional mile becomes an engineering and economic penalty.

and this is where counting the deaths caused by electricity becomes a very strange way of calculating the danger of electricity.

because you now have to count the other side.

electricity increased manufacturing productivity. historical economic work finds that electrification produced rapid and persistent productivity gains and changed how factories themselves were organized.

high-voltage transmission helped transform employment between 1910 and 1940. one study estimates electrification explains about half of the increase in operative employment and a meaningful part of the movement away from agriculture during that period.

rural electrification increased agricultural output and productivity.

but even these numbers are almost too neat.

because electricity is not one invention.

it is the thing underneath other inventions.

refrigeration.

water pumping.

factories.

hospitals.

ventilation.

communications.

food storage.

sewage treatment.

railways.

elevators.

machine tools.

laboratories.

eventually computers.

you cannot calculate the cost of delaying a general-purpose technology by looking only at the industry selling the technology.

you have to look at everything that does not happen because the infrastructure beneath it arrived ten, twenty or forty years later.

and there is another uncomfortable part.

edison wasn’t stupid.

that matters.

if he were stupid this story would teach us nothing.

his paper is actually quite sophisticated.

he discusses insulation degradation. moisture. arcing. underground conductors. human physiology. different current regimes. municipal regulation. boiler regulation as an analogy. the commercial incentives that encourage companies to use higher voltages.

he even correctly identifies the economic reason people wanted high-voltage ac.

it saves copper and real estate.

he writes this almost as an indictment.

read that sentence again with history behind you.

the thing making the technology economically scalable is presented as evidence against it.

that is fascinating.

because once you know how the twentieth century went, you can see the missing variable.

edison is calculating the danger produced by ac.

he cannot calculate the civilization that ac has not produced yet.

there is no column in the risk assessment for that.

no factory whose productivity never increased.

no village whose pump never arrived.

no vaccine spoiled because refrigeration remained expensive.

no mine ventilated by a motor that was never installed.

no child studying beneath a light that was never connected.

no hydroelectric station whose energy remained stranded twenty-six miles from the people who needed it.

the dead are visible.

the unborn counterfactual is not.

that asymmetry matters enormously when we regulate new technology.

a technological risk can be real.

the people describing it can be intelligent.

their experiments can be correct.

their mechanisms can be correct.

their deaths can be real.

and the policy conclusion can still be catastrophically wrong.

because the cost of a technology is not merely what happens when it exists.

it is also what does not happen when you prevent it from existing.

we know the first number rather quickly.

the second may require fifty years of history before anyone can even see what was being counted.

this is why i keep getting uncomfortable when safety arguments end at capability.

not because powerful technology is harmless.

electricity wasn’t harmless.

aircraft weren’t harmless.

medicine isn’t harmless.

intelligence certainly isn’t harmless.

the question is whether the danger gives us enough information to know the optimal ceiling on the capability.

edison thought it did.

six hundred or seven hundred volts for continuous current, he suggested. for alternating current he struggled even to name a safe pressure.

perfectly understandable.

and completely incompatible with the electrical civilization that followed.

there is a version of history in which regulators listened.

perhaps fewer people were electrocuted in new york in 1891.

that would have been real.

somebody’s son comes home.

some worker survives.

you cannot dismiss that.

but then the ledger keeps running.

at some point the question becomes horrifyingly difficult.

how many immediate deaths would justify delaying cheap, scalable electricity?

five years?

twenty?

fifty?

and how exactly would anyone standing in 1889 have calculated the second side?

they couldn’t.

that is the point.

the future had no statistics yet.

only the danger did.