I’ve always felt that AI found its footing in software engineering first because, frankly, the environment is honest. When you’re writing a blog post or a marketing strategy, “correctness” is a vibe. But in our world, it’s binary. Your code either compiles or it doesn’t. Your tests pass, or they fail.
That’s why software is such a perfect laboratory for these models. We already have the infrastructure they crave, like fast failure, clear signals, and the ability to undo/rollback changes. In a domain where the system itself…
