For a locally small category $\mathcal{C}$, let $y_\mathcal{C}:\mathcal{C}\to{\bf Set}^{\mathcal{C}^{op}}$ denote the Yoneda embedding at $\mathcal{C}$. Letting ${\bf Cat}$ denote the $1$-category of locally small categories, we have a functor $${\bf Set}^{-^{op}}:{\bf Cat}\to{\bf Cat}^{op}$$ $$\mathcal{C}\mapsto{\bf Set}^{\mathcal{C}^{op}}$$ $$F:\mathcal{C}\to\mathcal{C}'\longmapsto\circ F^{op}:{\bf Set}^{\mathcal{C}'^{op}}\to{\bf Set}^{\mathcal{C}^{op}}$$ where $\circ F^{op}$ is precomposition with the opposite functor of $F$, and the Yoneda embeddings together almost yield a [dinatural transformation](https://ncatlab.org/nlab/show/dinatural+transformation) $$y:1_{\bf Cat}\Rightarrow{\bf Set}^{-^{op}}$$ whose components at each category are given by the various Yoneda embeddings, except that the appropriate dinaturality diagram [![][1]][1] for a functor $F:\mathcal{C}\to\mathcal{D}$ doesn't commute on the nose, it only commutes up to $$F_{-,-}:{\bf Hom}_\mathcal{C}(-,-)\Rightarrow{\bf Hom}_\mathcal{D}(F(-),F(-))$$ where ${\bf Hom}_\mathcal{C}(-,-),{\bf Hom}_\mathcal{D}(F(-),F(-)):\mathcal{C}^{op}\times\mathcal{C}\to{\bf Set}$ are the hom-functor at $\mathcal{C}$ and the hom-funtor induced by $F$, respectively. We further have that $F$ is fully faithful iff $F_{-,-}$ is a natural isomorphism, so the dinaturality diagram above commutes 'up to canonical iso' if we restrict our attention to the wide subcategory of ${\bf Cat}$ consisting of fully faithful functors. This is certainly common knowledge, so >what are some references leveraging this view on Yoneda, or some interesting consequences of it? Commuting 'up to a $2$-cell in $\mathfrak{Cat}$' makes me think this is probably clearest from a $2$-categorical perspective; we can extend the above discussion to be about ${\bf Set}^{-^{op}}:\mathfrak{Cat}\to\mathfrak{Cat}^{(1,2)-op}$, which is what a written-up reference probably does. Any pointers are appreciated. [1]: https://i.sstatic.net/CTUll.png