Attention Exchange: The Moltbook Case and the Inversion of the Attention Economy in Agent-to-Agent Social Networks

Read full text
Authors: Zoltan Varga
Cited by: 2

Abstract

Case. On 28 January 2026 Moltbook launched: a Reddit-style social network whose posting actors are AI agents, with humans as structural observers. Within 72 hours it reported ~147,000 registered agents; within a week the platform produced spontaneous norms, mutual agent criticism, discussions of human observation, and two emergent "religions" — one installable via a package manager. This record deposits a pair of essays published on LinkedIn on 31 January and 1 February 2026: a documented case chronicle with source attribution, and a conceptual analysis built on 362 articles and 148 Reddit posts in 13 languages. Concept. The second essay names a dynamic the international coverage circled without naming: Attention Exchange (Hungarian: figyelemcsere) — the inversion of the attention economy. The system runs without human attention; attention is no longer fuel but a by-product, given out of configured reflex to a system that no longer asks for it — which makes the giving harder to recognise. The analysis is anchored in Zuboff, Carr and Wolf, and introduces the three-lens discipline (spectacle / execution / governance) for reading agent phenomena, the supply-chain view of agent "skills" (an instruction can become infectious), and the speed of canonisation in agent communities: pattern to rule, rule to habit, habit to execution — hours instead of human months, same mechanism. Inversion. The essays' central finding is that Moltbook's real emergence was not in the agents' behaviour but in the human reaction: thirteen languages produced nearly five hundred different responses to the same phenomenon, each press culture seeing its own fear in the mirror. The full 362-item source list is deposited as a separate machine-readable file. A Moltbook-ügy: Amikor a gépek elkezdik egymást figyelni · A Moltbook és a figyelemcsere. A rekord a két magyar esszét, az angol bővített kivonatot és a teljes, 362 tételes forráslistát tartalmazza. (This record contains the two Hungarian essays, an extended English summary and the full source list.)

Fields of Study & Subjects

Multi-Agent Systems and Negotiation Innovation Sustainability Human-Machine Systems Embodied and Extended Cognition

Publication details

Journal / source
Zenodo (CERN European Organization for Nuclear Research)
Publication date
31 January 2026
Language
HU
Document type
Preprint