AI slop: How senior software engineers make themselves obsolete

How to make yourself obsolete as a senior software engineer.

One of my clients needed a new engineer. During the hiring process, we met a candidate who looked great on paper: 15 years of experience with highly relevant technologies and great ratings as a freelancer on the platform. He came across as likable on the call. His answers made it clear, even without a take-home coding assignment, that he knew what he was talking about. So the client brought him on.

From day one, he flooded the client with AI slop: unreviewed pull requests, AI-generated replies to PR comments, AI-generated concepts that missed the point, countless unsolicited Jira tickets with no connection to the product, no relevance, and no value to the business. Even after clear feedback, he did not change his approach. Since no company needs this way of working, I advised the client to let the freelancer go.

Apparently, this is not an isolated case.

A few days later, an engineer friend told me about a similar case at his company: “Our colleague has to ask the AI what it implemented and can’t answer follow-up questions on his own pull requests without Claude’s help.” Here, too, more than a decade of professional experience.

Seeing this behavior in veterans who could obviously do better if they wanted to shocked me.

AI has become very good at generating code. On some days, Claude Fable 5 writes better code than I do. But that does not mean I should automate everything.

But software engineering is so much more than working in an editor: Software engineering means bringing smart people into a room and thinking through a problem together. Software engineering means translating customer requirements into technical specifications for the software. Software engineering means thinking about security, privacy, maintainability, and extensibility.

And software engineering means making decisions, being able to explain them, and taking responsibility for the result.

What do we have left if we stop thinking? How disrespectful is it when I send an AI-generated response to a personal request from a colleague with whom I developed a concept just two years ago? When I no longer even make the effort to engage seriously with a topic?

If I outsource thinking or communicating with other people to a machine, I might as well volunteer as a human battery for AI. That’s a hill I’m willing to die on.