So I asked the integrated AI a question and the answer was just some hallucination. Now that's nothing new. But the AI then proceeded to apply those hallucinated changes to my code base.
Luckily I had no uncommited changes and could use git to revert those multi file changes.
But it could have easily been different. So what did I do wrong? Did I click something? Pressed the wrong key? Don't want to happen this again and possibly destroy work.
