Indeed, but the author works through a deeply idiosyncratic train of thought before getting to that, and then quickly continues with some creative analogizing a

Indeed, but the author works through a deeply idiosyncratic train of thought before getting to that, and then quickly continues with some creative analogizing about how the idea of an invariant might be applied at other scales. For an essay titled "What is an invariant?", it doesn't really focus on answering the question at all and so the GP's failure to spot the traditional answer is pretty reasonable. A more fitting title may have been something more like "Invariants in software system design" -- better framing that it'll be a survey of examples and their rationales rather than an analysis of invariants.