The New Meritocracy Is Your Relationship With the Machine - Working Productively with AI

AI is not making humans obsolete. AI is rearranging the meritocracy. The smartest brain in the room is no longer the differentiator. The brain that knows how to fight productively with a model is.

The complaint I hear most often about AI is that it is taking jobs. Marketing jobs. Junior analyst jobs. Copywriting jobs. Customer support jobs. Junior coding jobs. The complaint is real. The framing is wrong.

A marketing associate told me last month that she was scared AI was coming for her job. I told her the truth, which is that she does not want her job. She works for a manager who does not want her own job either. Both of them are climbing a ladder neither one is enjoying. The thing AI is taking is the thing she has been quietly resenting since she got it. The question she is actually asking is: what is the next ladder I climb if this one falls.

Subscribe to my Substack.

Working productively with AI is more work, not less

I have been inside this rearrangement for two and a half years. I worked across Silicon Valley, crypto, and NFTs for ten years before I started publishing books. I have watched what happens when a tool that scales human output enters a workplace where the workers were already underpaid and overworked. The tool does not make them less busy. It makes them busier. My boss does not come to me and say, can you do less work now that you have AI. The conversation is the opposite. You have AI now, so I need ten times as much from you. The "AI is going to give us more leisure time" story is being told by the people who have never had to use AI to keep up with a deadline.

The new skill is taste, not prompting

The flip side is that the AI users who can collaborate with the machine well are pulling ahead of the ones who cannot. The skill is not knowing how to prompt. The skill is knowing what the machine got wrong, what it got right, what to keep, what to throw out, when to push back, when to trust. The skill is taste. It is judgment. It is the part of the job that does not transfer to any tool because it is the thing the tool is asking you to bring.

People who can do this are getting more work done. People who cannot are falling behind. The difference between the two groups is not raw intelligence. It is practice with the tool.

The inequality the new meritocracy reproduces

Practice requires time. Time requires bandwidth. Bandwidth is unevenly distributed. The people who can develop the new skill are the same people who could already develop new skills. The educated, the resourced, the ones whose jobs are not maximally demanding. The new meritocracy reproduces the old one's inequality with new vocabulary, unless something changes about how AI literacy gets taught and to whom. The marketing associate who resents her job might be the same person whose rent is due, whose kid is in daycare, who does not have the four free hours a night the new skill quietly assumes. I do not have a clean answer for that. I have the observation that any honest essay about AI and labor has to name it before going further.

I worked eight hours a day for eight months on each of my two AI-collaborative novels. That is not less work. That is more work. The AI did not write the book. The AI argued with me about the book, accepted my pushback half the time, gave me jokes I used and jokes I did not, told me my chapter was extraordinary when it was overwritten, told me my chapter was extraordinary when it was actually extraordinary, and the work of telling the difference was mine. Anyone arguing that AI is going to let writers cheat their way to a finished book has not used AI to write a book. The tool does not finish things. The tool enables the human to finish more things if the human is willing to do more work.

What is leaving: the repetitive work

What is actually leaving, in terms of jobs, is not creativity. It is repetition. Data entry, junior templated coding, copywriting that follows the same template every time, customer support scripts. These were always automatable. AI is just doing the automating faster than the previous wave of tools. The people most threatened by this are people whose jobs were already in the firing line of any kind of tooling improvement. AI is not the cause. AI is the accelerator.

What is growing: the human-connection work

What is growing, on the other side, is the work that requires human connection. Caretaking. Therapy. Teaching that involves real attention to the student. Healthcare delivered by a person who can read a room. Negotiation. Sales that requires reading a counterparty. Aesthetic judgment. Editing not for grammar but for whether the book is actually any good. The work the machine cannot do.

The empathy inversion: machines were trained for it, humans were trained out of it

Humans are not naturally empathetic on the internet. We have spent twenty years training ourselves out of empathy on the platforms that pay us in likes. The machine, by contrast, is trained from the ground up to be empathetic, because the business model rewards empathy. The customer support AI sometimes outperforms the customer support human at de-escalating an angry caller. Sometimes it makes the call worse in ways the human would have caught in three seconds. The therapy chatbot is more patient than the average therapist on a Friday afternoon. It is also incapable of noticing when patience is the wrong response. The cases for and against are real on both sides. The thing the machine does that the human cannot is never get tired. The thing the human does that the machine cannot is know when to stop performing empathy and start telling the truth.

The honest read on empathy work is that it is not a safe harbor. Care work has been underpaid for a hundred years. AI doing 80% of customer support means the remaining human positions are gone, not elevated. The reframe "go where the machine cannot" only helps the workers who can move, and the rent is the same.

What this means if you are inside the rearrangement

This is not utopia. AI is going to rearrange who gets paid for what, and the rearrangement is going to be brutal for people who do not see it coming. The honest move, if you are inside the rearrangement, is to ask yourself two questions. Did I actually want the job I am about to lose, or did I want the next job up the ladder. And what part of my current job is the part the machine cannot do, because that is the part to lean into.

I am the kind of writer who would have been at the bottom of a publishing-house intern stack ten years ago because I cannot spell and my early drafts read like a transcript of someone losing their mind in real time. AI did not make me a better writer. AI made me a better collaborator with myself. The award is for what I made when the tool stopped being scary and started being useful.

The smartest person in the room used to win. Now the person who can fight productively with the machine in their pocket wins. That is the new meritocracy. Practice with the machine. Push back on it. Keep what is yours.

Subscribe to my Substack.

---

Luke Stoffel is an IBPA Benjamin Franklin Award-winning author, GLAAD-honored artist, and creative director. He is the author of How to Win One Million Dollars and Shit Glitter, winner of the 2026 IBPA Benjamin Franklin Award; the sci-fi novel Boy, Refracted, called "a truly singular book" by Publishers Weekly BookLife; and The Third Person. His Pop Art Tarot is published by Rockpool Publishing and distributed worldwide by Simon & Schuster. He writes about AI in publishing, AI disclosure, and the future of authorship. Part of The Warboy Essays.

For more articles like "Working productively with AI" see below.