The Day My Manager Stopped Making Eye Contact
It was a Tuesday in March 2024, and Marcus had been my manager for six years. We had a standing one-on-one every week at 2 p.m. He always came in with his laptop half-open and a coffee he'd already forgotten was there. He would look at me when I talked. That was just how Marcus was.
That Tuesday, he didn't look at me once.
I noticed it about four minutes in. I was walking him through a deployment issue on one of our internal tooling pipelines — the kind of thing we'd talked through a hundred times — and his eyes stayed fixed on a point somewhere past my left shoulder. Not at his screen. Not at his notes. Just past me, at the wall, like he was reading something written there in a language I couldn't see.
I kept talking. I finished my thought about the pipeline. He said "yeah, makes sense." And then there was a pause, and he said he had some news to share, and I already knew. I knew before the sentence was done. Fifteen years at that company, and I knew from the direction of a man's gaze.
What I didn't know — what I've spent the two years since trying to understand — is how fast I had actually become optional.
I was a senior engineer. I had built systems that were still running, would probably still be running after I was gone. I had onboarded maybe thirty people. My name was in git commits going back to 2011. And then a set of internal tools powered by large language models had, over about eight months, quietly absorbed the majority of what I did every day. Not all of it. But enough. Enough that when someone in a spreadsheet somewhere drew a line, I was on the wrong side of it.
The part that still gets me — the part I was not prepared for — was how mundane it was. I think I expected displacement to feel dramatic. A confrontation, a failure, a moment I could point to and say: there, that's where it happened. Instead, there was just Marcus and his coffee and a wall he couldn't stop staring at.
Goldman Sachs estimated in 2023 that roughly 300 million jobs globally could be exposed to automation from generative AI. That number got cited in a lot of places. I cited it myself, later, when I was trying to explain to my mother what had happened. But the number doesn't tell you anything about the texture of it — the way the thing that ends your career doesn't announce itself, doesn't arrive with a reason you can argue against. It just becomes true, and then someone has to deliver the news, and they can't look at you when they do it.
I've talked to a lot of people since then who went through something similar. A paralegal in Chicago whose firm replaced a team of eight with a contract review system in under six months. A technical writer in Austin who was told, with genuine warmth, that his work had been "really valuable" and that they were "moving in a new direction." A data analyst in Toronto who found out by reading a company blog post before her manager had a chance to call her. What they all describe — what I recognize from my own experience — is less the shock of being fired and more a specific and disorienting feeling of retroactive invisibility. Like the work you did was real, and then it wasn't, and no one can quite explain how that transition happened.
The Oxford Martin School's research on automation vulnerability found that the workers most at risk often aren't the ones doing low-skill repetitive tasks — it's people in roles that involve processing information, generating structured output, applying rule-based logic. Which is to say, a significant portion of what I spent fifteen years learning to do well. That finding doesn't make me feel better. But it does make me feel less crazy, which is something.
I stayed in the meeting with Marcus for another twenty minutes. We talked about transition timelines and severance and whether I had questions. I didn't have questions. I had a feeling I couldn't name yet, something between grief and a very cold clarity. I thanked him. I think I meant it. He was in an impossible position and he showed up for it, even if he couldn't look at me.
Driving home that afternoon, I didn't cry. I just kept thinking about the pipeline conversation — the one I'd been in the middle of when he told me — and whether anyone was going to finish fixing it, and whether it mattered, and whether those were even separate questions.
They weren't, I eventually decided. Which is part of why I built Clankers: not because I had an answer, but because I had been in that car, and I knew what it felt like to need one.
PROTECT YOUR INCOME NOW
If AI automation threatens your job, preparation is the only honest answer: know your risk, build a runway, and reskill before you have to. The Clankers 90-day programme structures exactly that — free, which is exactly why today is the right time to start.
Start the 90-Day Programme — Free →GET THE CLANKERS BRIEF — FREE
One email every Monday: the week's AI displacement moves that matter, and what to do about them. No spam. Unsubscribe anytime.