This thesis examines how everyday social media users engage in curation practices to influence what news and information they see on their social feeds. It finds that cultural commentary content can act as a proxy for news on these platforms, contributing to public debate and the fifth estate. While much research has explored the implications of algorithmically driven recommender systems for content personalisation and news visibility, this thesis investigates a gap in our understanding of how social media users understand and respond to algorithmic processes, customising their feed in their day-to-day curation practices on these platforms. It explores how a group of Australians aged 18–30 respond to algorithmic recommender systems and how effective their practices are in shaping their social feeds. The study used a mixed methods approach that included a digital ethnography of social media use and a comparative content analysis of social media news exposure and topics in the legacy news cycle. This study develops a taxonomy of consumptive curation practices that users can engage in to influence their personalised social feeds. The study also examines users’ motivations for this curation and how effective these are in filtering news and ‘cultural commentary’ content into or out of their feed. The findings demonstrate that algorithmic literacy is a driver of active curation practices, where users consciously engage in practices designed to influence recommender processes that customise their social feed. They also demonstrate the prevalence of non-journalistic news-related content or ‘cultural commentary’ on social media platforms in the form of hot takes, memes, and satire, and how this cultural commentary can act as a proxy for the news, even for users who are news avoidant. These findings address gaps in our understanding of news discovery and consumption on social media platforms, with implications for how news businesses can reach emerging news audiences.