This phenomenon has been going on for quite a while, and more and more behaviors and remarks have appeared about completely abandoning IDEs and fully embracing CLI Agent development.
I wouldn’t dare put it too absolutely, but most of them are greenfield projects with no legacy baggage.
Using an Agent to quickly slap together a project in a short time, enjoying the thrill of taking a project from 0 to 1, then turning around and mocking anyone still using an IDE as an old fossil.
CLI Agents use AI—so does that mean what’s used in IDEs isn’t AI? People can even form a hierarchy of contempt over a tool; their brains must have short-circuited.
Of those who preach this logic every day, how many have actually built commercial systems? When a single update changes a thousand lines of code, and the AI lazily writes an if (!token) return false, it instantly becomes a toy project where authentication is effectively nonexistent.
The funniest part is that after making the changes, they themselves have no idea at all. Seeing dozens of changed files, they can’t be bothered to look and just commit directly—as long as the app runs, that’s enough.
The above is not aimed at any “experienced engineers”; please don’t take it personally.