For me, python is the closest thing to writing pseudocode that functions. Every time I have the instinct to gloss over a thing when writing it down (because it feels obvious in my head), it turns out that python has an intuitive abstraction for it.
Coming from a mathy background I found it incredibly satisfying, although I’ve come around to other languages since.
As a mathy person myself, I find the OOP leaning difficult to think about. Better equational reasoning, better lambdas, fewer side effects to worry about, avoiding mutation so that I can define something and still know what it is later during runtime, those are what help me think clearly. To me OOP is about the furthest paradigm from math that I’ve used.
You’re right, it’s not good for math, but unfortunately we have the legacy of AI / ML being entrenched into that ecosystem, with no clear leaders as alternatives.
Coming from a mathy background I found it incredibly satisfying, although I’ve come around to other languages since.