Ok, so could we agree that then doubling down on even more breaking changes is likely to create more codebases stuck at these "evolutionary bottlenecks"?

Whenever I hear someone say - "Gee, the only path ahead is a breaking change, sorry charlie!", what I really hear is - "I gave up thinking up a solution on how to do it an evolutionary way and I am lazy and just want to declare a revolution!".

There is always a pathway ahead on the evolutionary path. That's the premise at least.