Firstly, the paper seeks to show that Simon's notion of bounded rationality should be interpreted in light of its connection with artificial intelligence. This connection points out that bounded rationality is a highly structured concept, and sheds light on several implications of Simon's general view on rationality, especially as regards intentionality and instrumentality. Secondly, offering four paradigmatic examples, the article presents the view that recent approaches, which refer to Simon's heterodox theory, only partially accept the teachings of their inspirer, splitting bounded rationality from the context of artificial intelligence, and substituting it with different analytical tools which help give new configurations to bounded rationality. In particular, consideration of the role of automatic, neurobiological, and unconscious mechanisms (and their interactions with deliberate reasoning) provides new perspectives on how (bounded) rationality works. The thesis is that these events can be interpreted as an implicit (and ideal) challenge for redefining what bounded rationality is.