Phenomenal transparency, cognitive extension, and predictive processing

I discuss Clark’s predictive processing/extended mind hybrid, diagnosing a problem: hybrid suggests that, when we use them, pay attention to mind-extending external resources. This clashes with commonly accepted necessary condition of cognitive extension; namely, that resources must be phenomenally transparent used. then propose solution this problem claiming the phenomenal transparency should rejected. To do so, put forth parity argument effect cannot on extension: roughly, since internal can fail used, by parity, too. Further, argue is not even reliable indicator hence its absence considered for extended mind-predictive processing hybrid. Lastly, consider and allay number worries my proposal might raise, conclude paper.