I can’t imagine the head of LLVM Foundation to have similar blog post on Zig’s decision to move to their own compiler backend. And yet Andrew regularly mentions that this is because the bugs which LLVM brings. Here seems to be the similar situation.
I wouldn’t want to live in a world where each time a maintainer of project A removes problematic (as per his view) dependency on project B, the shit show starts. This is a typical pitfall and typical solution for that - again as per Andrew’s words on interview with JetBrains.