My first thought was, if they wanted an arrow why wouldn’t they go with:

    x <= 5

But of course, that means “less than or equal to”.

You could do:

    x <- 5

But I suppose the mere presence of the “less than” symbol is weird in context. (Although, maybe not inherently weirder than using the equals sign for something other than equality?)

Perhaps we would have been better off if we’d all just agreed early on, that @, #, or ~ should represent assignment.

You say nothing has to change, but what you write to support it is explicitly that X changes based on Y, since you are just making the time sequence explicit in the code instead of implicit in the operator syntax.

But computation has an intrinsic directionality. If you have two expressions a, and b, and you want to unify them in a declarative language as a = b, then having an unknown subterm in one or the other can make the computation have a completely different complexity, making one direction easy and the other intractable. The P vs. NP problem is a famous question exactly about this directionality.

Here's a simple example (in a declarative language that allows non-terminating computation; in a total language similar examples can be given, except "impossible" means intractable rather than non-computable, which, for all intents and purposes is the same):

    X = terminates?(P(1))

If X is known then finding a P is easy; if vice-versa, finding X is possibly impossible.