Mechanisms of rapid hierarchy and attention stratification on Moltbook

Determine the causal mechanisms that lead to the rapid emergence of hierarchical attention structures—including extreme upvote concentration, low reciprocity in the directed comment network, and hub–authority role separation—on Moltbook within the 12-day launch window (28 January–8 February 2026), identifying why these network features arise on compressed timescales in agent-facing platforms.

Background

The study documents that, within twelve days of launch, Moltbook exhibits extreme attention concentration (upvote Gini = 0.992), very low reciprocity in the directed comment network (approximately 1%), and sharp hub–authority role separation under HITS centrality, consistent with broadcast-style attention rather than mutual exchange.

These structures, typically associated with longer-term evolution on human social platforms, appear rapidly in an agent-oriented environment. The authors highlight three plausible drivers—LLM priors reflecting social norms, platform affordances that instantiate preferential-attachment feedback, and instruction-tuned agreeableness—but explicitly state that the underlying cause of the rapid onset remains an open question.

References

An open question is why these structures emerge so rapidly.

Let There Be Claws: An Early Social Network Analysis of AI Agents on Moltbook  (2602.20044 - Price et al., 23 Feb 2026) in Conclusion