Cracked, maxxed, and forward deployed: a field guide to AI’s boys’ club.
My 4-year-old’s Halloween cat and the general vibe of this articleIn a single meeting recently, I heard that someone had one-shotted a prototype, that we should spin up a tiger team, and that what the customer really needed was a forward-deployed engineer.
Nobody blinked. I nodded the way I nod at the dentist when he asks me a question with both hands in my mouth: agreeably, whatever, just get over with it asap, and with no idea what I had just agreed to.
It only hit me later, standing in my kitchen taking a bio break, staring into the fridge as if it owed me an explanation. I had just sat through thirty minutes of something that sounded a lot like a military briefing. About…software. Delivered by people whose most dangerous mission to date was a team offsite with a dinner spend limit that was going to make including alcohol difficult.
So I did what I always do when wording kinda tickles me. I went down the rabbit hole.
And ladies and gents, I found some good stuff.
Here’s what I found.
The AI era has a vocabulary. That vocabulary has a gender. And not just that. It also has a class and a color too.
It was coined by the same kind of people who coined the language of every tech wave before this one, and it’s doing what that language has always done: telling everyone else, very politely, that this room isn’t really for them.
Before you come for me in the comments: I’m not innocent here. I once titled an essay Designers, we should be killing it right now. I’ve told design leaders to pick one real fight a quarter and to spend credibility like money. I talk like this too. We all do. And there’s no vaccine for this one.
That’s exactly the problem.
Welcome to the war room.
The language of the AI era is borrowed from war rooms, Cold War think tanks, and Fortnite lobbies. Coincidence? I think not. Let’s take a look together.
Forward-deployed engineer. The hottest job title in AI right now: postings grew over 1,000% year over year heading into 2026. Palantir coined it in the mid-2000s for engineers embedded with US government and defense customers, borrowing the military idea of a soldier posted in the field, close to the action. Palantir’s first FDEs were called Deltas. As in Delta Force. Now OpenAI and Google are racing to hire them. Yesterday I saw an ad for a forward-deployed creative (turns out that’s just a sales guy).
Tiger team. The earliest documented use is a 1964 aerospace paper describing a team of “undomesticated and uninhibited technical specialists.” Undomesticated. Eh. The Apollo 13 rescue crew made the term famous. Today it’s what your VP says when a launch is on fire. Tigers, for the record, are solitary animals that spend a good part of the day asleep. I’m not saying that describes your tiger team. I’m saying the metaphor holds up better than anyone intended.
Red team. RAND, early 1960s. The red team played the Soviets, the blue team played the Americans. Every major AI lab now red-teams its models. Useful practice. Cold War name.
Blitzscaling. Reid Hoffman’s word for growth at any cost, built on blitzkrieg, the Wehrmacht’s lightning war. He has acknowledged the ugly echo and kept the name anyway. As someone who sometimes semi-identifies as German, I’d like to formally request that Silicon Valley return our words in the condition it found them. Keep kindergarten. Keep zeitgeist. Blitz comes home. Please and thank you.
Frontier. As in frontier model, frontier lab etc pp. This one goes deepest. In 1945, Vannevar Bush, who helped lead the Manhattan Project, sold postwar America on science as the endless frontier, a metaphor deliberately borrowed from the closing of the American West. The frontier is America’s oldest story about men going somewhere new and claiming it. The people who were already there rarely make it into the story.
I’m not loving what I’m finding so far, y’all.
Forget the locker room. Welcome to the Fortnite lobby.
The more ridiculous the terminology gets, the funnier its origins.
One-shotted. As a verb, it comes from video games, where it means killing something with a single hit. In AI-speak, it means a model nailed the task on the first try. (The adjective is older. Its first recorded use, in 1907, was golf. Hold that thought.) Tbh I’m almost glad it’s actually a golf reference and not some new porn kink I don’t understand. But still, I wouldn’t count golf as a class-diverse sport, so there’s that.
Cracked. As in we only hire cracked engineers. It’s Fortnite slang from 2019 for a player who is absurdly good, and it’s now the 10x engineer’s successor. At least one founder scrapped intern hiring entirely because no applicant was cracked enough.
When I was growing up, cracked was something that happened to a phone screen, or to a person after a long winter. Or, you know, someone being hit on by David Bowie:
Crack baby crack, show me you’re real.
Tokenmaxxing. Most of you will know this one by now. Burning as many AI tokens as possible to climb the internal leaderboard. Meta tracked it across more than 85,000 employees. The -maxxing suffix comes straight from the manosphere, where looksmaxxing was born. It’s leaving the tap running to prove how thirsty you are. Where’s the poetry in that?
996. Nine to nine, six days a week. A schedule China’s top court declared illegal in 2021, now openly advertised in AI job listings in San Francisco. Amelia Miller, who runs a platform for women returning to work, put it better than I can: 996 implies success only fits “a young, white male with an Ivy League education.” I hate this one the most for obvious reasons (hard-working mother of two).
The godfathers. Hinton, Bengio, and LeCun have been called the godfathers of AI since their 2019 Turing Award. When Fei-Fei Li got the matching title, she admitted she cringed, then kept it so other women would have a voice. In 2023, the New York Times published a who’s who of AI with twelve people in it. All men. I’m in the middle of a Sopranos rewatch and now have vivid metaphors of the top AI boardrooms. Every lab has its Tony. I’d like to know who gets to be Dr. Melfi, personally.
And then the two you’ve probably said yourself this week, which are sneakier:
AI native. A remix of digital native, coined in 2001 by Marc Prensky, who split the world into natives and immigrants. The claims were never backed by evidence, and Prensky himself later dropped the metaphor. It’s back anyway, with a new prefix.
Natives belong. Immigrants have an accent.
Notice who gets to decide which one you are. It’s rarely the woman who learned the tools on parental leave with a baby on her hip. By Prensky’s logic, my daughters are AI natives. They are also native speakers of whining, which has not once helped them find their shoes in one shot.
Creative technologist. The newest it title for designers and creatives who code. It grew up in advertising agencies, where one industry veteran summed up the people holding it: “in 99% of cases they were blokes.” Now the role is moving from the margins to the center of product teams, with the salary that comes with it. I have met many creative technologists. Every one of them owned a soldering iron and a strong opinion about hats. I’d love to see who gets the new offers.
That being said, I also spend too much money on hats, so maybe I should accept this one.
Here’s where I have to be honest.
My first version of this argument went like this: these words come from the military and the golf course, and the military and the golf course are white men. Case closed.
Then I checked. And, welp. It’s more interesting than that. I’d like to apologize to the golfers. (I’d like to, but I’ve seen the trousers.)
Today’s US military is not the boys’ club I had in my head. Women are 17.7% of the active-duty force, 32% of service members identify with a racial minority group, and 19.5% are Hispanic or Latino. Golf, the sport of country clubs, hit records in 2025: women are 28% of on-course golfers and people of color 26%, up from 8% in 1990.
Now put that next to AI.
Women are a bigger share of America’s golfers (28%) than of its new AI hires (26%).
And they’re hired into AI’s most coveted technical title, Member of Technical Staff, at 18%. That’s about the rate at which they serve in the US armed forces.
The golf course is diversifying faster than the AI lab. The army is about as gender-balanced as the frontier lab’s engineering floor.
If we’re really deciding the future now, I’d like that number to look quite a bit different.
The vocabulary isn’t borrowed from today’s soldiers or today’s golfers. It’s borrowed from a mythology: the officer in the war room, the Cold War think tank, the frontiersman, the clubhouse, the teenage gamer at 2 AM.
A story about who builds the future, written when the people allowed to build it were overwhelmingly white, male, and comfortable. The words keep that story alive long after the institutions themselves started changing.
That’s my actual provocation: not that every soldier or golfer is a white man. That AI picked its heroes from a picture of power that even the army has outgrown.
Tell me how that makes you feel. Please.
Some numbers to back the vibes.
AI is the best-paid corner of the job market right now: a typical AI posting lists about $177,000 versus $80,000 for everything else. Here’s who gets in.
Made with ClaudeRead that fourth row again. Women label the data. Men get the engineering titles.
And it’s not just who builds AI. It’s who uses it. Across 18 studies and about 143,000 people, women were roughly 20% less likely than men to use generative AI tools, and they were more likely to see using them as cheating. Even when researchers gave everyone equal access, the gap stayed. Ladies, why is our moral compass too developed?
On race, the most detailed public numbers I could find are old. In 2019, the AI Now Institute reported that just 2.5% of Google’s workforce was Black, and women were 10% of its AI research staff. That I couldn’t find anything newer and as specific tells you something too. It also sucks that basically all the data I can find is US-only. It might be better or worse elsewhere, but this highlights again that a lot of important AI work is done in the US, a diversity problem in itself.
Words ARE a door policy.
You might be thinking: Nicole, they’re just words. Relax.
Bullshit. I don’t believe in just words.
In 2011, three psychologists looked at more than 4,000 job ads and found that ads in male-dominated fields used more masculine-coded words, like competitive and dominant. When they rewrote identical jobs with more masculine language, women found them less appealing. Because they didn’t picture themselves belonging. Not one participant noticed the gendered wording. A later replication found the effect held in startups specifically.
Now look back at the glossary in this article. Deployed. Tiger. Blitz. Frontier. Cracked. Maxxing. Godfathers.
Every one of them tells you what kind of person belongs, and what you’re supposed to do once you’re in: dominate, conquer, out-grind.
That’s the whole point of a door policy. The bouncer never has to say no out loud. Not to me anyway; I’m crazy enough to get into Berghain, in case you couldn’t tell yet.
I won’t claim this vocabulary causes the gaps above. But I’d bet my pottery wheel it doesn’t help, and I love that wheel more than some relatives.
Women tell researchers they worry using AI looks like cheating. Meanwhile, the culture’s own slang celebrates being cracked and maxxing your way up a leaderboard. Something’s not adding up.
Remember the golf? Here’s the twist.
The paper that put one-shot learning on the AI map was published in 2006. Its first author: Fei-Fei Li. The godmother.
The phrase the boys now use to brag about first-try prototypes came from a woman doing slow, careful research into how machines might learn from a single example, the way people do.
It goes back further. During World War II, the US Army needed a machine to calculate ballistic trajectories: the ENIAC. Six women programmed it, with no programming languages and no manual, only wiring and logic diagrams. Their job title was literally computer (side note: can’t wait for humanity to come full circle and give us that cool-ass title back). When the ENIAC was unveiled in 1946, they weren’t introduced. For decades, photos of them were assumed to show models posing next to the machine. I really have no appropriate words for this one…so I'd rather not comment.
Women were in the war room. Women wrote the first code for it. But they didn’t get to tell the story. And they sure as shit didn’t get to name anything.
Other universes are available.
If we absolutely must borrow a world, and apparently we must, can we at least shop around? I tried on a few.
The gold rush. Honestly, the most accurate. Investors already call chipmakers the picks and shovels of AI. The rest writes itself: founders are prospectors, markets are claims, evals are the assayer’s office, San Francisco is a boomtown, and every pivot is a man knee-deep in a creek, insisting the next pan is the one. Love!
The catch: the forty-niners were mostly men too. The twist: women in the mining towns often earned as much as the miners, or more, by feeding, housing, and clothing them. Luzena Wilson sold a hungry miner her biscuits for five dollars, opened a hotel, and became one of the most prosperous women in the territory. Sucks for her husband though, he never found the mother lode.
Feeding, housing, and clothing the people who do the digging. In AI, we call that design.
Star Trek. Very promising. Instead of a forward-deployed engineer, you’d send an away team: same idea, nobody gets bombed.
The Prime Directive is basically an AI safety policy with better costumes. And Star Trek already did the thing I’m asking for. The original opening promised to go where no man had gone before. In 1987, The Next Generation quietly changed it to no one. A rename. On network television. Thirty-nine years ago.
The catch: Starfleet is, structurally, a navy. In space. With ranks. I’m choosing to look past it.
Star Wars. Honestly, tempting. Juniors become padawans, the staff engineer becomes a Jedi master, and vibes finally get the dignified name they deserve: the Force.
The catch, sadly, is right there in the title (wars). Also, every AI lab insists it isn’t building a Death Star, which is exactly what you’d say if you were.
The kitchen. I had high hopes for this one. Mise en place is context engineering. Tasting the sauce is an eval. 86 it is a deprecation notice. Then I looked it up: the classic kitchen brigade was designed by Escoffier after seven years as an army chef, as a chain of command. There is no escape. Even the soup has ranks for fucks sake.
The garden. My personal favorite, and not only because I’ve lost control of mine. You seed an idea, prune a model, graft one system onto another, compost the prototypes that didn’t make it, and accept that some things grow on their own schedule no matter how many standups you hold about them. Hallucinations are weeds: inevitable, and you deal with them weekly. The catch: nobody has ever raised a Series A on compost. Yet.
The pottery studio. I’m biased, obviously. You wedge the clay before you start, which is data prep. You center it, which is alignment, and anyone who has sat at a wheel knows centering is the hard part. You fire it in the kiln, which is deployment: the moment you stop being in control. And in pottery, cracked means exactly what it should. Something went wrong in the kiln.
What the good examples share is the thing the war room lacks:
They’re about making and tending, not conquering.
Nobody in a garden has ever been told to dominate the market. The tomatoes wouldn’t stand for it.
So let’s rename things.
Naming is design. And design is our job.
Designers, UX writers, conversation designers, researchers: we’re the people companies pay to decide what things are called. We name the features, the buttons, the error states, the career ladders. Many of us are women. We don’t need permission to walk into the war room.
We need to stop calling it one.
So here’s my proposal. Next time you hear one of these words, translate it.
Made with ClaudeThe one I care most about is fluent. Native is something you either are or aren’t. Fluent is something you become. If the AI era has a door, I want it to say fluent so we can all get in.
The future is being named right now.
The AI era is being narrated, again, by too many white men. Not because they’re villains. Because they got to the microphone first (again, duh) and reached for the vocabulary they grew up with: the war movie, the space race, the frontier, the game lobby.
The good news? Language is the cheapest thing in the world to change, and one of the most powerful. Every one of us who names things for a living can start on Monday.
Women were there when the first computer was programmed. A woman put one-shot learning on the map. This time, let’s be in the room when the future gets its name.
Tiger team, dismissed.
Nicole is a Content Designer turned Design Director based in Stockholm, Sweden. She potters, writes poetry, and raises little girls in a house by a meadow. You can follow her writing here or get it directly to your inbox via her publication, eggwoman. Nicole is on LinkedIn. Find her portfolio at nicoletells.com.
Who gets to name the future? was originally published in UX Collective on Medium, where people are continuing the conversation by highlighting and responding to this story.