The Job Is Still Here. I'm Just Not Sure I Recognise It Anymore.

Someone posted on Reddit this week asking whether they were going to spend the rest of their career reviewing AI-generated code. They mentioned that colleagues were boasting about not having written a single line of code in months. That markdown lists of ideas were showing up in meetings, obviously AI-generated, presented as thinking. That the expectation had quietly shifted: a good engineer now “supervises AI” and “focuses on the bigger picture.”

The post hit something I’ve been sitting with for a while.

I do this work too. Not pure development anymore, but close enough. And I’ve watched the same shift happen around me. People reaching for Claude before they’ve even fully formed the question in their own head. The long back-and-forth calls where you’d whiteboard something terrible, argue about it, and eventually land on something good: those are rarer now. Everyone’s got an answer before the conversation starts. It just came from somewhere else.

The original poster was clear about something that I think got lost on a few of the people responding to them. They weren’t complaining about AI existing. They weren’t even saying the efficiency gains aren’t real. They were saying: I like this work. I like the moment when a gnarly query finally runs fast. I like the problem, the dead end, the realisation, the fix. Handing that to an agent isn’t a promotion. It’s just loss dressed up as productivity.

A lot of the replies tried to reframe this as the assembly-to-C transition, or the shift from writing code to leading teams. And I get the argument. Tools change. Abstraction layers accumulate. You adapt. But there’s something different happening here and I think the poster sensed it even if they couldn’t quite articulate it. Previous abstractions still left you doing something. Writing Python still feels like writing. The thing you’re building still passes through your hands at some level. What’s being described now is a pipeline where a human writes a vague requirement, agents handle the rest, and another agent reviews the output. The human is present but not quite in it.

One commenter put it well: the job is getting more boring and more lonely. That lands for me. The craft is becoming optional. And when something becomes optional in a productivity-obsessed industry, it tends to become deprecated.

Here’s where I genuinely don’t know what to think. There’s a reasonable argument that this is just a transition period, that the pendulum swings, that flooding the world with fast mediocre AI-generated code will eventually create its own backlash and the people who kept their skills sharp will be in demand again. I find that argument comforting. I’m also aware that I find it comforting, which is reason enough to hold it loosely.

The more honest version is this: some people in the thread clearly never loved writing code and AI has given them permission to stop pretending they did. That’s probably fine for them. But the person asking the question loved it. And telling someone who loves something that they should be happy to supervise a machine doing it instead is a bit like telling a chef they should be thrilled to manage a kitchen full of robots. Maybe the food comes out faster. That’s not really the point.

What I keep coming back to is the comment about people just defaulting to “ask Claude” the moment they get stuck. That’s the part that actually worries me, not the efficiency argument, not the job market predictions. It’s the atrophying of the instinct to sit with a problem. To not know the answer yet and stay in that space long enough to find it yourself. That instinct is worth something. Once it’s gone it’s hard to get back.

The original poster said the job is getting soulless. I don’t think that’s melodrama. I think they’re describing something real that a lot of people in this industry aren’t letting themselves say out loud yet.