A Single Bottleneck for the Whole World's Machinery? I Have Seen This Before.

I have read this curious dispatch, purportedly from the year 2026, and I confess it has given me a night of poor sleep and furious thinking. Not because I doubt its plausibility — no, quite the opposite. I recognize the pattern. It is my pattern. They speak of something called a 'semiconductor,' a tiny wafer upon which, it seems, the entire computational future of mankind depends. And they tell me the whole blessed supply of these things hinges on one Dutch firm and a handful of fabrication plants. One firm! A few plants! The entire architecture of what they call 'artificial intelligence' — a phrase I find both thrilling and suspicious — resting on so slender a thread. Well, I know what it means to be that thread. Right now, in 1890, my generating stations and my networks of copper are the thread upon which the electric light of American cities depends. I have built the Pearl Street station. I have strung the wires. I have fought Westinghouse and his alternating current at every turn, because I understand that whoever controls the infrastructure controls the industry. Whoever owns the dynamo owns the future. So when this dispatch tells me that nations quarrel over who manufactures these chips, that a government act must be passed to secure their production on home soil, that the contest between America and China mirrors nothing so much as a war over who holds the generating station — I am not astonished. I am vindicated. But I am also troubled. I have learned, through ten thousand failed filaments and a hundred patent fights, that centralization is both a weapon and a weakness. When I controlled the whole chain — the generator, the wire, the lamp, the meter — I was king. But one fire at a factory, one injunction from a court, one fool in the boiler room, and the whole system trembles. They have evidently built a world empire on the same fragile principle, only at a scale I can scarcely imagine. Theory without application is worthless, I have always said. But application without redundancy is reckless. If I were advising these future men, I would tell them what I tell my own engineers every morning at Menlo Park: build two of everything, test ten thousand alternatives, and never let one supplier hold you by the throat. That is not philosophy. That is arithmetic.