Collaborative filtering cannot recommend an item nobody has touched. Content-based filtering can, because it looks at what the item is rather than who engaged with it — and every production system ends up combining the two.
An embedding model turns text into a fixed-length vector of numbers positioned so that semantically similar text ends up nearby in that vector space. That's the entire mechanism retrieval is built on: instead of matching keywords, you compare vectors.
An image and the sentence describing it are, on the surface, completely different kinds of data — a grid of pixels versus a sequence of tokens. CLIP's contribution was to train both an image encoder and a text encoder so that matching pairs land in the same embedding space, close together — and the moment that works, classification becomes a search problem, with no fixed label set required.
One-hot encoding a vocabulary treats every word as equally different from every other word — "cat" is exactly as far from "dog" as it is from "bicycle." Word embeddings replace that with geometry: words that appear in similar contexts end up close together in a continuous vector space, and "close" starts to mean "similar" in a way a model can actually use.