The Apocalypse Is Already Boring

The Wired piece about vibe coding and the engineering apocalypse is fine — it does the thing where you take a trend seriously enough to report on it and skeptically enough to feel responsible — but the more interesting artifact to read alongside it is Andrew Ng's tweet, which I cannot stop thinking about.

The vibe coding panic has a predictable shape by now. Someone ships a CRUD app in forty minutes using Cursor without writing a line of code themselves, posts the video, and two camps immediately materialize: the people who say this is the end of software engineering as a profession, and the people who say actually you still need to understand the stack because the AI hallucinates APIs and can't debug its own state management. Both camps are correct about their specific observations and missing the point entirely.

Ng's framing cuts sideways through all of that. The useful question isn't "does AI replace engineers" — it's closer to "what problems are actually unsolved now that the cost of generating code approaches zero." That's the thought model. The engineering apocalypse discourse assumes the value of software was the code. It wasn't. It was the judgment about what to build and why, which remains annoyingly hard and human.

Vibe coding is a great tool for exploring whether an idea is stupid before you invest in it. It's a terrible thesis about the future of expertise.

The apocalypse keeps getting scheduled and then rescheduled. The engineers are still here. They're just using different tools and having the same arguments about whether the tools are real.