C-like languages are much simpler for newcomers to pick up. There are people out there who don't have extensive CS experience and can't easily pick up complex concepts like mutability, ownership, lifetime subtyping, and so on.
There are some anecdotes of people picking up Rust as their first language, but they usually have something else going for them, such as a lot of time, a very patient and involved teacher, or some sort of extra external motivation.
We also often hear how easy Rust is to pick up, but from actually talking to people who have tried, it's clear that this is mostly survivorship bias; people don't tend to broadcast their failure to learn a language. For every N people that successfully learned Rust, there are M people who couldn't push past the learning curve and had to go with something else.
If the entire industry is going to jump straight to Rust and require it for all new projects, I hope it doesn't become a moat of gatekeeping, where only the smartest and most tenacious programmers are allowed to do low-level programming and everyone else is harassed out of the field. That's how a field becomes an isolated echo chamber and stagnates.
I'm not saying we should stop investing in Rust. We should definitely keep at it. But we should also invest in languages that prioritize simplicity and a more merciful learning curve, and make it clear which situations in which they are acceptable/encouraged.
(This might sound biased from me, but the language I'm involved with is a high-level language, so I have no dog in this fight.)