You’re wrong — AI will take your job. Eventually.
Everyone insists AI will take someone else’s job, never their own. I think we’re all a little wrong — and that’s mostly fine.
I’ll be honest: this one annoys me. Every single day there’s an article or a podcast where some expert explains, calmly, that their job is the one AI can’t touch. I just heard an economist do exactly this — financial analysts and traders, gone; economists, never, because they’re so special and have skills no computer could ever match. Yeah. Right.
The other favourite move is to “test” ChatGPT by handing it something deliberately impossible, watching it stumble, and declaring the whole thing overhyped. Of course it stumbles — it’s a generalist, not your specialist. Yet. And every time you make it fail, you’ve just handed OpenAI a tidy example of something to fix. Thanks for the training data.no, really — thank you
It’s moving fast — uncomfortably fast
We can argue all day that the model got this or that wrong. It probably did. But it’s getting things wrong less and less. In January 2023 we had a GPT-3 that invented facts and fake citations. A couple of months later GPT-4 was writing prose coherent enough to send teachers everywhere into a panic about cheating. That’s months, not decades.
And it’s everywhere now. Google has its model, the UAE has one, China has several, even Finland has its own large language model. On top of that: realistic voice cloning (the original of this very post was read aloud in a clone of my own voice), plus image and video generation. Most of that arrived inside a single year.
We’re not losing the work — just the boring parts
Here’s the part the panic misses. The AI is good, and yes, it will take our jobs. But not really our work — just the current shape of it. And honestly, that shape could use the change.
A doctor doesn’t want to spend the day typing records like a secretary. A professor doesn’t dream of filling spreadsheets and writing project reports like a middle manager. An electrician doesn’t enjoy wading through a tedious manual to find which wire goes where. And as a programmer, I do not want to build the same login screen and stand up the same web server for the hundredth time. That work is repetitive; it rarely teaches us anything or challenges us in the way we actually wanted to be challenged.
Instead, the work gets propped up by tools. Some draft the report; some dig out the one fact you needed; some proofread, or wrangle a meeting across five calendars, or make the graphics, or even ring customer service and argue an insurance claim on your behalf. Probably all of that, and plenty we haven’t thought of yet.
So the next time someone tells you “AI won’t take my job,” take it with a pinch of salt. Not today, sure. But in five years? I genuinely think AI will take my job as a programmer — and I honestly hope it does. Then I get to spend my time on the interesting things. Writing. Teaching. Research.
…oh. It does those too already, doesn’t it.