That's up to debate, you have to prove that in x.y.z language the time spend for maintenance would have been avoided if x.y.z had that feature from the start. Especially language like Rust that have a big learning curve / blockers, is it justified on the end? I'm not sure.
In Go I never felt that omg it misses some big features we're screwed, we're going to pay the cost of running that in Prod. All the services I worked on while not being perfect ran like a clock without issues on the long term, and it has to do with Go "simplicity", eventhough you never wrote the code you can actually jump into any code base without issues, there is no weird abstraction layer / complexity that takes you day to understand.
For me the fact that Go is a "simple" language is very powerful because on the long run it's much easier to maintain.
In Go I never felt that omg it misses some big features we're screwed, we're going to pay the cost of running that in Prod. All the services I worked on while not being perfect ran like a clock without issues on the long term, and it has to do with Go "simplicity", eventhough you never wrote the code you can actually jump into any code base without issues, there is no weird abstraction layer / complexity that takes you day to understand.
For me the fact that Go is a "simple" language is very powerful because on the long run it's much easier to maintain.