Paul Ford, writing in the New York Times in September 2026, argues that AI did not kill the software developer. It democratized bad software instead.
Ford's core claim is precise and uncomfortable: AI makes it easy to do someone else's job badly. That is why the killer apps never arrived. Every project that handed a non-engineer an LLM and called it a product ran into the same wall. Craft still requires practitioners. Collaboration still requires judgment. The robots are tireless, but tireless is not the same as good.
The piece is worth reading in full because Ford is not celebrating a developer comeback. He is diagnosing a systemic failure in how the industry imagined AI would compound human skill, and instead watched it distribute human error at scale. The line 'now that everyone can code, it's become clearer why many shouldn't' is the sharpest thing written about the AI coding moment so far.
[READ ORIGINAL →]