Eh, he's given an interview where he talks about the Swift decision. He and several maintainers tried building some features in Swift, Rust, and C++, spending about two weeks on each one IIRC. And all the maintainers liked the experience of Swift better. That might have ended up wrong, but it's a pretty reasonable way to make a decision.
Two weeks with Rust and you're still fighting with the compiler. I think the LLM pulled a lot of weight selling the language, it can help smooth over the tricky bits.