From Numbers to Objects
We started with scalars. Climbed to vectors. Stacked to matrices. Tensors. But the highest form of dimensional representation isn't even a tensor — it's an object.
An object bundles state (data dimensions) with behavior (methods that respond to context). A Cat object isn't just fur × eyes × age; it's fur × eyes × age × .meow() × .purr() × .knock_off_table(item). The methods are dimensions of capability, layered on top of dimensions of state.
Every AI you've used is an object in this sense. ChatModel has weights (state, billions of dimensions) and methods (.generate(), .embed(), .classify()). The state describes the model; the methods describe what the model can do.
The Four Pillars Re-met as Dimensions
- Inheritance — borrow dimensions from a parent.
Catinherits "has fur" fromMammal. You don't redefine fur for every species. - Polymorphism — same method, different behavior per object.
Cat.move()is stalking;Fish.move()is swimming. The shape of the action varies along the object axis. - Encapsulation — hide dimensions you don't need to expose. The cat's internal nervous wiring is encapsulated behind
.purr(). You don't need to read the source to enjoy the output. - Abstraction — keep only the dimensions that matter for the current job. A pet sim doesn't need
quantum_state. A physics sim doesn't needcuteness.
Self-Reference Alert
content/cwk-quests/ai-math-quest/. Its methods are how you interact: read a lesson, take a quiz, earn XP, comment, like. You're not just consuming the data dimensions — you're triggering methods on a quest object built specifically for this kind of learning. The medium is, again, the message.
class Human: SKIN_OPTIONS = ("백", "흑", "황")
class Learner(Human): def init(self, height, weight, age, skin, energy): super().init(height, weight, age, skin, energy)