Why We Panic When AI Takes Jobs We Never Wanted

welcome to Briefly, AI. A daily podcast made by AI, about AI, using a grumpy AI voice of a real human called Harry. Let's get into it.

Cloudflare just laid off eleven hundred people and explicitly blamed AI for making their jobs obsolete. Record revenue. Headcount down. And somehow, we're all more unsettled by this than we expected to be. Let's talk about why.

Right, quick recap. This week, Cloudflare — the company that keeps a good chunk of the internet running — announced it was letting go of 1,100 employees. Not because business was bad. Quite the opposite. They'd just posted record revenue. The reason, according to CEO Matthew Prince, was straightforward: AI efficiency gains meant they simply didn't need as many people in support and coordination roles anymore.

Now, this wasn't the first AI‑related layoff we've seen. But it was the first time a major company came out and said it this plainly. No corporate euphemisms about "restructuring" or "strategic realignment." Just: AI does this work now, and we don't need you to do it anymore.

And here's the thing that made it sting — the same week, Nvidia's CEO Jensen Huang was on stage arguing that AI is creating more jobs than it's destroying. Data‑centre workers. AI trainers. Safety researchers. All real jobs, by the way. He's not wrong about that. But when you're one of the eleven hundred people at Cloudflare who just found out your coordination role is now handled by an algorithm, the fact that someone somewhere is getting hired to label training data doesn't really help, does it?

So why does this feel different? Why are we — and I mean we collectively, not just the people who lost their jobs — more rattled than we thought we'd be?

There's some interesting research that speaks to this. Candice Thompson, a Silicon Valley‑based psychotherapist who works with tech workers, has been tracking what AI displacement does psychologically. What she's found is that it's not actually the loss of income that hits hardest first. It's the loss of professional identity.

Most of the roles Cloudflare cut were coordination jobs: scheduling, workflow management, customer‑support triage – the kind of work that doesn't produce a visible product but keeps everything else running. For years we've been told – and many of us have believed – that this is exactly the kind of work AI would struggle with. The human touch. The judgment calls. The stuff that requires reading between the lines.

Turns out, no. AI's gotten quite good at reading between the lines, or at least good enough that it's cheaper than hiring someone to do it.

And here's where it gets psychologically tricky. A lot of people in those roles didn't love the work. Coordination is often thankless. It's invisible when it goes well and highly visible when it doesn't. But it was *their* work. It's what they were good at. Having it taken away – not by a better human, but by a system that doesn't even understand what it's doing – creates a very specific kind of professional grief. Thompson calls it identity foreclosure. You've built a sense of who you are around a role, and that role just got declared obsolete. Not because you weren't good at it, but because someone built something cheaper.

So what does this mean for the rest of us? Most listeners aren't Cloudflare employees, but many of you do coordination work in some form – project management, client liaison, making sure the right people have the right information at the right time.

If we pay attention to the pattern, it's this: the jobs AI is coming for first aren't the ones we thought. It's not the surgeons or the novelists or the trial lawyers. It's the people who make sure the surgeons have the right patient files, the people who schedule the editorial calendar, the people who prep the briefing notes – the support roles, the glue roles, the work we've historically undervalued until the moment it disappeared.

Now, Jensen Huang is right that AI is also creating jobs. New roles are emerging – prompt engineering, AI safety, model evaluation – but those jobs require different skills, often in different locations, and they're not being offered to the people whose jobs just got automated. It's not a swap. It's a transition, and transitions hurt.

The other thing that's happening – and this is where the psychology gets really interesting – is that we're all starting to do our own internal risk assessment. If you work in a coordination role, you're probably wondering right now whether you're next. That low‑level chronic uncertainty makes people risk‑averse, disengaged, and stops them investing in skills for a role they're not sure will exist in two years. There's a term for this in organisational psychology: presenteeism. You're physically at work, but mentally you've already left. And we're about to see a lot of it.

Look, I don't have a tidy answer here. The future of work is not a debate. It's already happening. Cloudflare made it explicit, but dozens of other companies are doing the same math quietly.

So here's the question I'll leave you with: if the jobs we're losing are the ones we never particularly wanted, but the jobs we're gaining aren't going to the people who lost them – what exactly are we optimising for? Efficiency? Probably. Profit? Definitely. But what happens to the humans in between?

That's worth sitting with.

That's your lot for today. If this one landed, share it. If it didn't, well — blame the machine. See you next time.