Member-only story
AI Isn’t Replacing Developers. It’s Doing Something Worse.
The replacement narrative is loud. The quiet one is already happening.
A few weeks ago I was reviewing a PR I’d put together with Claude Code. Standard refactor, splitting a chunky service class into something cleaner. The diff was maybe four hundred lines. I scanned it. Tests green. I approved it.
Two days later I was back in the same area for an unrelated reason and noticed something off. One of the methods I’d “refactored” had quietly lost a null check. Not catastrophically, the path it gated wasn’t hot. But it was gone. I’d reviewed the diff. The tests had passed. I’d shipped it.
Twenty-five years of writing code, and the version of me from 2020 would have caught it in twelve seconds.
I sat with that longer than I want to admit. Not because of the bug. Bugs are bugs. I sat with it because of how the bug got through.
That’s the piece I want to write about. Not the bug. The how.
The Wrong Worry
There’s a conversation happening in our industry right now that I think is missing the point. The conversation is about replacement. Will AI replace developers? When? How many? Half the headlines you see are some flavor of that question, and…
