The first night after the layoff, I opened my résumé and started editing.
"Led cross-functional initiatives that drove a 30% lift in core metrics." I had written that sentence so many times I no longer knew what it meant. So I handed the whole document to an AI and asked it to optimize.
Thirty seconds later, it gave me a better version. Stronger verbs, sharper numbers, keywords mapped neatly onto the job description.
I suddenly noticed the thing that replaced me was also the best at writing the document meant to prove I was irreplaceable.
If a machine can describe me better than I can, maybe that document was never describing me in the first place.
1. The résumé is a genre written for machines
The résumé was never designed to help someone know you. It was designed to help a stranger judge you in seconds.
To do that, it evolved strict conventions: lead with a verb, end with a number, front-load the keywords, keep it to one page. Today the first reader is often not a person at all but an applicant tracking system. What we call "good résumé writing" is, at bottom, the craft of making ourselves easier for machines to parse.
In Seeing Like a State, the political scientist James C. Scott describes how states govern through legibility: flattening messy, local realities into formats that can be recorded, counted, and compared. The résumé is legibility engineering for the workplace. It is lossy compression. It keeps what can be compared across candidates (titles, years, metrics) and discards what cannot (why you made a choice, what it cost you, how you climbed out after failing).
Legibility and replaceability are two sides of the same coin. The more cleanly a job can be written as input, action, output, the easier it is to automate. For years I made myself clearer and more standardized on paper. In a sense, I was writing my own replacement manual.
So the part of me that AI "optimized" away was precisely the part I had worked hardest to make legible.
2. Optimization needs an objective function
Companies call it "restructuring" or "rightsizing." In some workplaces the euphemism is blunter: yōuhuà, literally "optimization." It is a more honest word than it sounds.
In mathematics, optimization starts with an objective function: what you are maximizing and what you are minimizing. To be optimized out means you were a cost term in someone else's equation.
This year's numbers bear that out. According to Challenger, Gray & Christmas, AI had been the leading stated reason for U.S. job cuts for five straight months as of July 2026. In its June annual filing, Oracle tied roughly 21,000 job cuts, about 13% of its workforce, directly to AI deployment. Yet analysts also note that companies have an incentive to dress up ordinary cost-cutting as "AI transformation," because that story signals efficiency to investors.
In other words, "replaced by AI" often doesn't mean AI can do your job. It means that on a budget spreadsheet, AI spend and your salary ended up in the same column.
That leads to a crucial distinction: being removed from an equation tells you about the equation, not about you.
The most common trap after a layoff is quietly adopting the other side's objective function to judge yourself. Am I not good enough? Are my skills obsolete? Those questions are worth asking. But if your only yardstick is whether you were worth it inside one company's cost structure, you have let someone else's spreadsheet define you.
3. A résumé says what you are, not who you are
In The Human Condition, Hannah Arendt distinguishes between what a person is and who a person is. "What" covers talents, qualities, and shortcomings, things that can be listed and compared. "Who" only shows itself in what a person says and does. She also observed that the moment we try to state directly who someone is, language pulls us back into describing what they are.
The résumé is the perfect container for "what." And "who" is exactly what the format was built to leave out.
The philosopher Michael Polanyi summed up a whole category of knowledge in one line: "we can know more than we can tell." He called it tacit knowledge. And today's AI learns mostly from what humans have already written down.
Consider what never appears on a résumé:
- What you learned the afternoon you sorted out a residence permit, alone, in a country whose language you barely spoke.
- How you rebuilt your confidence, day by day, after completely botching something.
- That colleagues come to you first when things go wrong, before they go to their manager.
- The real reason you left your last job, and everything you weighed before deciding.
- Knowing when to hold on and when to let go, a judgment you paid for.
They are judgment distilled from experience. They are the "who."
4. When generation is free, experience gets expensive
In economics, when the supply of something becomes effectively unlimited, its price falls toward zero, and its scarce complements become more valuable. When photography made realistic depiction cheap, painting turned toward what the camera couldn't do. When search engines made information cheap, judgment became the premium.
AI has made "producing something that sounds right" nearly free. A flawless résumé, a well-organized guide, career advice that sounds reasonable: all of it arrives in seconds.
So what got expensive? Is knowledge backed by cost.
A model can write a convincing guide to job hunting in Berlin. It can't have spent three months doing it. The person who did knows things the guide leaves out, often without realizing they know them.
That changed how I read the list in the previous section. I used to treat those things as background, the personal material behind the real qualifications. Now they look like the only part of my working life that a cheaper substitute can't produce.
So, now what?
This isn't an essay telling you to just be yourself. You still need a résumé. You still need a job.
But if the reasoning above holds, there is more worth doing after being optimized out than polishing the résumé:
- Write an anti-résumé. Write the hardest thing you have been through. A decision you made that will never go on a résumé. Something you know now that you didn't five years ago. Your real inventory is far larger than one page.
- Ask the people around you: "What do you usually come to me for?" The reasons people seek you out reveal your most underrated value, and it rarely has anything to do with your title.
- Give one experience away. Find someone two years behind you on the same path and talk with them once. Explaining forces tacit knowledge into words, and you will find you know far more than you thought. What they get is something no guide can give them.
A résumé is how systems read you. Everything outside it is how people recognize each other. AI may become the best possible reader of the first. The second still takes another person.❤️