When mutate_all() was soft deprecated, my understanding was that the recommended replacement was to use across(.cols = everything(), ...) inside a mutate() command, e.g.
mutate_all(mtcars, .funs = round)
becomes...
mutate(mtcars, across(.cols = everything(), .fns = round))
With the recent deprecation of everything() being the default value for .cols in across(), I get a sense that maybe that is not expected usage? Is that still the currently recommended replacement, or is there something else?