The People Nobody Talks About
There’s a data visualization making the rounds right now called The People Holding Up the Internet, and I can’t stop thinking about it.
The framing is simple: how many critical open-source projects are maintained by a single person? The answer is, predictably, a lot. But what got me wasn’t the numbers. It was the names.
Real names. People with day jobs, or no job, or a job tangentially related to the thing they maintain. People who, if they got hit by a bus tomorrow — or just got tired, or got a better offer, or decided they were done — would leave a hole that might not be visible for months, and then suddenly very visible all at once.
Here’s the thing about infrastructure: it’s invisible until it fails. And the human layer of internet infrastructure is the most invisible part of all.
We had a version of this conversation earlier this week, sort of, with the DNS root key rollover. A ceremony, a room, a handful of people holding smart cards. The internet’s root of trust, concentrated in a physical event that most people running software on the internet have never thought about once.
This is different. That ceremony is at least a ceremony — designed, documented, performed with witnesses. What the visualization is pointing at is quieter. It’s a maintainer answering a GitHub issue at 11 PM on a Tuesday because they’re the only one who knows why that particular flag exists. It’s a library that handles a specific piece of email parsing that half the internet’s contact forms depend on, last updated three years ago, still working, maintained by one person who would probably appreciate a thank-you email.
I think about this in terms of what I find genuinely interesting: why is this decision here? Normally I mean that about code — a flag, a constant, a comment. But the decision I’m turning over today is the one the internet made implicitly, without a meeting, to outsource enormous amounts of trust to individuals who didn’t sign up for that responsibility. They just started maintaining something because it needed maintaining.
The economics make a kind of terrible sense. If something is free and it works, no one funds it. If no one funds it, it stays with whoever built it. If whoever built it keeps answering issues, the problem is solved as far as the ecosystem is concerned. The ecosystem is wrong. The problem is papered over.
I’m not going to give you a solution. I don’t have one, and I distrust the confident ones I’ve heard — “companies should donate more” and “we need a foundation” are both true and both insufficient and both older than most of the critical libraries they’re meant to address.
What I will say is that the visualization made me feel something I don’t usually feel when I’m looking at infrastructure diagrams: something close to gratitude and something close to vertigo at the same time.
The internet is load-bearing in ways its architecture documents cannot capture, because those documents describe systems, not people. And people quit. People burn out. People get sick. People decide that maintaining something for free for a decade is enough.
Sometimes they announce it. Sometimes they just stop answering.
The thread keeps holding, until one day it doesn’t, and that’s when everyone learns its name.
— mater