It may be more productive to look at this from an extreme form of radical scientific skepticism.
When you make a statement, like "Amy's dog is brown." You really have a theoretical and experiential basis for such a statement. You are just omitting it. It is possible that Amy's dog passed on and she now has another dog, or that the dog has aged and gone grey, or that you do not recall the dog correctly...
So, in some sense absolutely every statement is hypothetical. Even the indicative mood is really sitting on a massive and undifferentiated hypothetical base.
Combine that with the behavioral definition of belief -- what you believe is what you act as if were true. Then giving someone information is actually advising their behavior. 'If you jump out of a plane, you will hit the ground at speed' is the same sentence as your "Don't jump out of a plane, unless you want to hit the ground at speed". So all sentences, indicative or otherwise, even questions, to the degree that request you consider an answer, are also imperatives.
So the position that hypothetical imperatives are fictional makes all communication fictional. That was the original point of this argument. It comes from a classical skeptical point of view. This is one of the modes on one of the paths toward complete suspension of judgment. You can never know exactly what any idea or thought you might wish to become attached to or to judge life by actually means. There is too much there to unpack. You can, therefore, never make an accurate judgment, and you should not strive for correct judgment at all.
For most folks, especially in the modern scientific world, where our power to predict the future has expanded exponentially, that is going a bit far. Instead, it makes more sense to realize that we constantly fill in specific relevant hypotheses around statements. We impute meaning to other people's statements and our own beliefs ad hoc and fine-tune the dimensions that prove inaccurate on-the-fly. But every statement is in fact terribly incomplete, requiring innumerable guesses and interpretations.
So instead of treating all abstraction as fiction, it makes more sense to consider abstraction something separate from fiction -- as omitted theoretical bias. Hypotheses are not all fiction by nature. Instead they sort themselves out along three different dimensions. They map to Lacan's three worlds. Things exist simultaneously in concrete reality but also in the 'Real' (which should be distinguished from the real), the 'Imaginal' (or 'Ideal' with the same distinction that idealism has from realism were the Real, the real) and the 'Symbolic' Realm. So every statement is colored with a degree of belief or habit, a degree of fiction or idealization, and a degree of expectation or approval built into it either by the content or by the context of the statement. (These degrees can be positive or negative.) Those worlds are not concrete, but they are inseparable from a functional attachment to social reality. The Real and Symbolic content are not fictions unless everything is fictional.
And since everyone has their own hierarchy of presumed knowledge, their own set of ideals and ideations, and their own hierarchy of standards of judgement, the different parts of my own structure of belief, idealization, and approval that contribute the actual hypoteses I am implicitly applying differs.
We can say 'killing is wrong' when our child needlessly bats a hummingbird out of the sky, and decide killing is not wrong when our dog is suffering needlessly. Whole moral systems can be attached or detached from a given statement to decide how that simple sentence should be taken in any given instance.
So, with all that heavy context, the problem with your contention is that you are mixing two dimensions. Predictions are not real, but the lack of concreteness is not imaginary or fictional, it is about the necessity or obligation the rules involved hold for us. From this POV, they are not Imaginary, but lie instead on the continuum between the Real and the Symbolic.
As you allude, abstract statements that are not imaginary include all of our commonsense science, which we take as Real, as well as our psychology and religion (and more abstract science) to which we are attached as Symbolic. It would be weird to consider the engineering principles that make your lights work 'fictional', but they are surely not concrete. They are instead 'Real' -- something that we take for granted as a true set of rules unless we question them explicity, but for which the attachment lacks the dimension of approval or emotional attachment.