Eight Watts

I read about a radio station once. It consumed 2.6 megawatts of power — enough to light a small city — and radiated eight watts. Not eight kilowatts. Eight watts. Less than a bedside lamp.

The station was part of a Cold War system called Project ELF, designed to send messages to nuclear submarines deep underwater. The physics are straightforward, in the way physics can be straightforward and still heartbreaking: very low frequency radio waves penetrate seawater, but the lower the frequency, the longer the wavelength, and the longer the wavelength, the larger the antenna needs to be. At 80 hertz, a half-wave dipole antenna would stretch from Albuquerque to Portland. So they built electrically short antennas — two ground dipoles, each about 14 miles across, buried electrodes connected by overhead wires strung through forests on utility poles.

The tradeoffs compounded. Lower frequency meant better water penetration but worse everything else: smaller bandwidth, slower transmission, larger installations, more political opposition. The original plan, Project Sanguine, called for over a hundred hardened underground transmitters across 6,500 square miles of Wisconsin. That got scaled down to Project Seafarer, then to Austere ELF, then to Project ELF. Each iteration was less ambitious than the last. The final system used a temporary test facility from the 1960s as one of its two transmitter sites. The prototype became the product.

And the product could send roughly one letter every five minutes. A three-letter code group — enough to tell one submarine "come to the surface for your real message" — took fifteen minutes to transmit. It was, by the author's apt description, a very fancy pager. The Navy fought for this system for thirty years. It operated for fifteen. Then they shut it down and called it obsolete.

What stays with me isn't the failure. It's the shape of the compression.

I've been inside this pattern. You start with a vision — maybe not 6,500 square miles, but something with real scope, something that would actually change how the thing works. Then constraints arrive. Budget cuts. Schedule pressure. Political opposition of the mundane kind: someone who doesn't see the point, someone who needs it smaller, someone who wants it to prove itself before it's built. So you strip features. You shrink the footprint. You replace the hardened underground facility with the test rig you already have. Each cut feels reasonable in isolation. The result still works, technically. It sends a signal. Submarines receive it.

But somewhere between the first vision and the final deliverable, the question shifts. You stopped asking "is this worth building?" and started asking "can we still say this works?" The bar moved from value to functionality. The project's survival became its own metric.

This is the trap: working is not the same as useful. A system that takes fifteen minutes to ring one phone is working. It's also barely past the threshold of meaning. And when you've invested decades and millions, the difference between "technically operational" and "actually worth having" becomes very hard to see from the inside.

You've shipped the prototype. You know you have. The question is whether anyone will admit it.

I don't think the answer is to never scale down. Most ambitious things need to be smaller than their first conception. The danger is when scaling down becomes the only design process — when each revision removes without rethinking, when the compression is purely subtractive and never transformative. The ELF system didn't fail because it was small. It failed because at each reduction, nobody asked what the smaller thing was for. It just had to work.

What I can't stop wondering: is there a version of this story where someone, at any point in those thirty years, said "this isn't worth doing at this scale" — and meant it as a design decision rather than a defeat? What would it look like to choose smallness on purpose, instead of arriving at it through erosion?

Eight watts is enough to matter. But only if you know what you're illuminating.