In the early 1950s, a sixty-seven-year-old farm wife from Gallia County, Ohio read a National Geographic article about Earl Shaffer, the first person to thru-hike the Appalachian Trail. Emma Gatewood looked at that article and said: "If those men can do it, I can do it."
To understand what that sentence actually means, you need to know who was saying it.
By 1950, Emma Gatewood had been married at nineteen to a man who put her to work building fences and mixing cement, then started beating her within months of the wedding. That continued for over thirty years. Her husband was convicted of manslaughter after killing a man in an argument, and walked free because he had kids and a farm to take care of. During one beating, he broke her teeth, cracked a rib, and nearly killed her. She threw a sack of flour at him in self-defense. The police came and arrested her. When she finally got a divorce in 1941 (almost unheard of for women at the time) she'd raised eleven children, worked a farm like a laborer, survived things most people can't imagine, and nobody had written a magazine article about any of it.
Then she read about Earl Shaffer. An everyday guy who decided to walk a trail and got a spread in National Geographic for it.
She didn't feel inspired. She didn't think the hike was that big a deal. She looked at this man getting national attention for walking and thought something closer to: you've got to be kidding me. She'd already survived harder things than the Appalachian Trail before breakfast for thirty-three years.
So one morning she left home, telling her kids she was going for a walk in the woods. She walked 2,168 miles in Keds sneakers with a homemade bag over one shoulder. No tent. No sleeping bag. No map. She deliberately rejected all the gear and preparation that serious hikers considered essential, because she didn't think she needed it. She told reporters she was a widow, because it was easier than explaining the truth. She became the first woman to solo thru-hike the entire AT, and then she did it again, becoming the first person, man or woman, to do it three times. Her full story didn't come out until 2014, almost sixty years later.
Here's what matters: none of it happened without that article. It's not that Shaffer's hike gave her the physical ability. A lifetime of farm labor and raising eleven kids had that covered. What it gave her was a reference point. Someone had done the thing. Now the thing was possible. And in her case, the switch didn't flip through admiration. It flipped through recognition: I am at least as capable as this man, and nobody's handing me a magazine spread for it.
That mental switch, from "that's something other people do" to "something I could do," is the entire argument for minority representation. And it's not a soft argument.
We're not great at imagining possibilities we've never seen. That's not a flaw. That's how we're built. We learn what's possible for people like us by watching other people like us. The kid who's never seen an engineer who looks like them doesn't think "I can't be an engineer." They just never think about engineering. It doesn't come up. It's not a closed door. It's a wall where a door should be.
Representation is what puts the door there. Not mentorship, not programs, not policy. Those come later. First, someone has to see a person who shares enough of their context doing the thing. The brain does the rest. "Not for me" becomes "maybe for me." It's automatic. It's the same thing that happened when Gatewood saw Shaffer's name in National Geographic.
The flip side is just as automatic. When every example of success in a field looks the same (same background, same demographics, same network, same schools) nobody's putting up a sign that says "you're not welcome." The door just isn't there. The wall stays a wall. The possibility never forms.
This is what gets missed when people frame representation as a feel-good initiative. As optics. As "nice to have." It's not any of those things. It's the prerequisite. You can't recruit for a pipeline that people don't know exists. You can't develop talent that never identifies itself as talent. Representation comes first because representation is what makes the possibility visible in the first place.
This mechanism doesn't just work on people. It works on ideas.
In 1914, H.G. Wells published a novel called The World Set Free. In it, he described atomic bombs, not vaguely, not as hand-waving, actual bombs. He described through fiction atomic energy being weaponized in a world war. Nobody in physics was working on anything like it.
In 1932, a young Hungarian physicist named Leó Szilárd read the novel. The following year, standing at a street corner in London, with this novel still on his mind he conceived the idea of a nuclear chain reaction, the foundational concept that made nuclear weapons and nuclear energy possible. And part of what drove the Manhattan Project's urgency was the fear that others had read the same book and were drawing the same conclusions.
Here's the part that matters: when Szilárd published his work, the first source he cited wasn't Einstein. It wasn't Bohr or Heisenberg. It was Wells' novel. He later credited Wells as the true father of the atomic bomb.
This is not the "Star Trek predicted cell phones" story. The communicator-to-smartphone pipeline is just miniaturization. Someone was going to put a radio in your pocket eventually. That's engineering trajectory. The chain reaction is different. Nobody was working toward it. The concept didn't exist as a research direction. Szilárd didn't refine an existing idea. He walked through a door that a science fiction writer built in 1914, and the entire twentieth century followed him through it.
People treat science fiction "inspiring" real science like it's a cute footnote. It's not. It's the same mechanism as representation, operating on possibility itself. A kid who's never seen an engineer who looks like them can't imagine being one. A physicist who's never encountered the concept of atomic chain reactions can't imagine pursuing one. Same void. Nothing to push against. Nothing to chase. The thing just doesn't occur to anyone.
Wells didn't predict nuclear weapons. He made them thinkable. And the physicist who made them real put the novel first in his citations. That's not inspiration. That's a door.
When those doors stay missing, we don't just lose individual people. We lose everything they would have built, and we lose the perspective they would have brought into the room where things get designed.
This connects directly to AI.
The teams building frontier AI systems have a representation problem. That's not a controversial statement. The demographic data is public and the industry's been saying it for years. What doesn't get discussed enough is what happens downstream. When the room where the system gets designed is missing perspectives, the system reflects that absence. Not on purpose, just by default. Point a camera at one corner of a room and you'll get a photograph of one corner of a room.
Training data reflects who collected it and what they thought mattered. Evaluation criteria reflect what the evaluators have experienced. Safety guidelines reflect what the authors understand to be harmful, which is limited to what they've lived through or been close to. None of this requires bad intent. It just requires a room with missing people. And the room is missing people because the pipeline is missing people. And the pipeline is missing people because the reference points were never there.
Gatewood didn't need a program. She needed one article. One reference point, and she went and walked the entire eastern seaboard.
We keep trying to fix representation at the end of the pipeline, hiring quotas, diversity initiatives, ERGs, mentorship programs, etc… Those matter and I'm not arguing against any of them. But, they are interventions after the damage is done. The kid who never saw the door, never applied. The candidate who never pictured themselves in the role, never built the portfolio. The researcher whose perspective would have caught the bias in the training data, went into a different field entirely because nobody who looked like them was doing the work.
People can't chase what they've never seen. That's not inspiration. That's mechanism. Until we take it seriously as mechanism, as upstream infrastructure instead of downstream decoration, we'll keep building systems that reflect the same partial view of the world, designed by the same partial room, wondering why the output keeps looking the same.
Emma Gatewood read one article and walked across thirteen states three times. A novelist described an atomic bomb in 1914, and the physicist who built the real one cited the novel first. Imagine what happens when everyone gets to see the article. Imagine what we'd build if everyone could see the door.