I think that the idea behind this project is to simplify Emacs' code, e.g., by dropping support for ancient systems (they mention MS-DOS) and using pre-existing crates instead of in-house solutions. A simpler codebase might provide significant advantages for end users as well: (1) there will be less bugs in the code, (2) finding and fixing these bugs will be simpler, and (3) new features and structural improvements will be easier to implement. Of course, users will actually switch to Rustmacs only if they will be 100% compatibile, i.e., if Rustmacs can run every Elisp code in the same way as Emacs does. It's surely a very ambitious project!