Sub-spaces of R^3 are a useful way of generating and picturing examples of 2D manifolds. But these things exist by themselves. As the article mentions with angles of a triangle, you can tell that a 2D manifold is curved from living inside. Likewise you can tell whether a 3D manifold is curved of not, without any mention of a 4th dimension.
Questions of curvature of the universe are a step harder, as we are talking about 3+1-dimensional space-time. But the slice of constant time is a 3D manifold, and as far as we can tell right now, it appears to be flat.