Remark 2.24. Definition 2.21 makes sense for any stable category \(E\) and full stable subcategories \(C_1,\dots ,C_n \subseteq E\), omitting condition (1). No presentability or size assumption is needed. Parts (1) and (2) of Proposition 2.22 hold in this setting with the same proof, and the sequence \(D \xrightarrow {i_D} E \xrightarrow {i_C^L} C\) is a Verdier sequence: \(i_C^L\) is a localization with kernel \(D\), and it inverts precisely the maps whose cofiber lies in \(D\). If \(E\) is idempotent complete, then so are \(C\) and \(D\), being thick subcategories of \(E\) (as kernels of the exact functors \(i_D^R\) and \(i_C^L\)), and the sequence is short exact. (In the talk, the definition was given simultaneously for presentable stable categories, for dualizable categories and for small stable categories.)