My main gripe with immutability is that making updated data requires building a full copy of the data again with the changes.

That's not generally true. Many immutable languages are using "persistent" data structures, where "persist" here means that much of the original structure persists in the new one.

For more, see:

- Purely Functional Data Structures by Okasaki: https://www.cs.cmu.edu/~rwh/students/okasaki.pdf - Phil Bagwell's research - e.g., https://infoscience.epfl.ch/record/64398/files/idealhashtree...