The paper explores the duality of user engagement on prominent social media platforms, analysing interactions among 834 users across audio-based and text-based modalities. By employing social network analysis, it investigates how user interests align differently in real-time, audio interactions versus traditional, text-based discussions. Our findings reveal that social networks built from X Spaces activity data exhibit characteristics of a broad-scale network, as commonly seen in other types of online social networks. The results also reveal that audio-based environments foster focused conversations among users with shared interests, exhibiting high topic similarity. In contrast, text-based interactions encompass a broader diversity of topics, with a weak correlation between the discussions in both modes, reflecting the varied interests of users. This research underscores the complex dynamics of digital communication and enhances our understanding of how distinct communication forms impact user behaviour and social interactions.