Internet Studies and Social Theory
The Unraveling of “Social Media” as a Category
The dominant arc across this cluster is one of paradigmatic exhaustion. A remarkable convergence emerges in the 2026 anniversary essays of Social Media + Society: the very object that founded the field is dissolving. Tornberg2026-lc and its companion piece Tornberg2025-ir make the most programmatic version of this claim, declaring the need for a “post-social media studies” organized around three emergent formations — algorithmic broadcasting platforms, semi-private micro-communities, and AI-mediated communication. Boyd2026-op reaches a kindred conclusion through a different route, proposing that we redescribe these platforms as “parasocial media” to register the collapse of reciprocal sociality into one-sided attention to influencers. Gerbaudo2026-fo supplies the morphological diagnosis: a shift from “networked publics” structured by explicit interpersonal ties to “clustered publics” assembled by algorithmic inference from implicit behavioral signals. Read together, these papers form a coordinated obituary for the social-graph paradigm.
Periodization and the Politics of Inevitability
A second thread treats this transformation as historically contingent rather than technologically inevitable. Marwick2026-ss anchors this argument autoethnographically, contrasting LiveJournal’s reciprocal intimacy with today’s extractive logics and insisting that different design and governance choices could have produced different outcomes. Baym2026-tr revisits her 2015 manifesto to document the empirical worsening of the four threats she identified — wealth concentration, algorithmic opacity, precarious labor, data extraction — translating diagnosis into a call for engaged, cross-sector intervention rather than critique-for-critique’s-sake. Bruns2026-yv performs a comparable post-mortem for Twitter specifically, narrating its mismanagement, weaponization by reactionary actors, and Muskian “enshittification,” while warning that simply cloning its form (Threads, Bluesky, Mastodon) without rethinking moderation reproduces the same vulnerabilities. The shared framing is political-economic: a contingent trajectory, not an unfolding logic.
Infrastructures, Platformization, and the AI Turn
Where the anniversary essays mostly diagnose decline, Helmond2026-ll re-theorizes the underlying sociotechnical process. Revisiting “platformization” a decade after coining it, Helmond argues that the concept must be sharpened — redefined as a platform-specific form of Manovich’s “transcoding” — and that it is now entering an AI-centered phase symbolized by the retirement of Facebook’s Like button and Google’s space-based compute infrastructure. This piece dovetails directly with the Törnberg–Rogers thesis: the Like button’s death is precisely the infrastructural correlate of the shift from social-graph to interest-based recommendation, and from user-generated to AI-generated content. Helmond’s contribution is to insist that this is not a clean rupture but an uneven, collaborative, sectorally specific reorganization involving new actors (“Big AI,” defense startups, energy and space firms).
Theoretical Resources: From Networked Publics to Systems Theory
The cluster also reveals the theoretical vocabularies being mobilized to make sense of the transition. Gerbaudo2026-fo reaches back to Simmel’s “social forms” and Weberian ideal-types to give morphological precision to the networked/clustered distinction. Tornberg2026-lc explicitly extends and seeks to supersede the lineage of boyd, Papacharissi, Jenkins, van Dijck, and Zuboff. An older but theoretically generative counterpoint is Giglietto2019-e9be81c1, which deploys second-order cybernetics (von Foerster, Bateson) and Luhmann’s systems theory to reconceptualize false-information cascades as observer-dependent propagation chains rather than properties of originating intent. Its move — from creator-centric to propagation-centric analysis, organized across micro/meso/macro levels — anticipates the broader analytical reorientation the 2026 essays now demand: when the social graph dissolves and algorithmic curation, AI synthesis, and clustered publics take over, the unit of analysis can no longer be the posting user but must become the circulatory dynamics of an opaque, multi-level system.
Open Questions
Several tensions remain productive across the set. First, the relationship between the “parasocial” diagnosis (Boyd2026-op) and the “clustered publics” diagnosis (Gerbaudo2026-fo) is underspecified — are these the same phenomenon described at different levels, or distinct logics? Second, the post-social media studies program (Tornberg2026-lc, Tornberg2025-ir) and the nostalgic-reformist register (Marwick2026-ss, Baym2026-tr) imply different research agendas: paradigm replacement versus political-economic intervention. Third, Helmond2026-ll’s warning about conceptual erosion through overuse applies reflexively to the new vocabulary now proliferating — “clustered publics,” “user-spectators,” “parasocial media,” “Big AI” — and invites a disciplined effort to specify what each term does and does not capture. The structure note this cluster supports is one of a field self-consciously rebuilding its conceptual foundations in real time.