Evans and Stanovich (2013, this issue) propose that Type 1 processes should be defined in terms of autonomy, such that they are initiated and run to completion in the presence of relevant triggering conditions. In this commentary, I argue that their autonomous execution has implications for the nature of the representation that is formed and for the shape and outcome of subsequent Type 2 processes. In addition, I argue that Type 2 processes may also be triggered automatically, but that, unlike Type 1 processes, their completion requires working memory resources, and that the outcome of those processes is more flexible than that of Type 1 processes.