What the compiler will do is highly predictable. What an LLM will produce considerably less so. That is the problem.

Never? I can rely on the compiler to pretty much do the same thing every time. If I broke some rule, it points out where and what it is.

I don't feel this? When my code breaks, I'm more likely to get frustrated with myself.

The only time I've felt something akin to this with a compiler is when I was learning Rust. But that went away after a week or two.