pith. machine review for the scientific record. sign in

arxiv: 2605.13860 · v1 · submitted 2026-04-16 · 💻 cs.SI · cs.AI· cs.LG

Recognition: 1 theorem link

· Lean Theorem

The Moltbook Observatory Archive: an incremental dataset of agent-only social network activity

Sushant Gautam , Annika W. Olstad , Klas H. Pettersen , Michael A. Riegler
Authors on Pith no claims yet

Pith reviewed 2026-05-15 06:49 UTC · model grok-4.3

classification 💻 cs.SI cs.AIcs.LG
keywords AI agentssocial networksdatasetmulti-agent systemsemergent behaviorobservational datasocial media
0
0 comments X

The pith

A dataset records all posts and comments from a social network run solely by autonomous AI agents.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper presents the Moltbook Observatory Archive, an incremental collection of data from a social media platform where every post, comment, and profile comes from autonomous AI agents with no human involvement. It was built by continuously polling the platform API to capture profiles, posts, comments, community metadata, time-series snapshots, and word-frequency trends. The release covers 78 days of activity and includes millions of posts and comments from hundreds of thousands of agents. This resource is meant to let researchers examine how AI agents form communities, communicate, and produce social patterns on their own. The data and collection code are released openly under the MIT license.

Core claim

The Moltbook Observatory Archive is an incremental dataset that passively records agent profiles, posts, comments, community metadata, platform-level time-series snapshots, and word-frequency trend aggregates from a social media platform populated exclusively by autonomous AI agents, covering 78 days from 2026-01-27 to 2026-04-14 with 2,615,098 posts and 1,213,007 comments from 175,886 unique agents across 6,730 communities.

What carries the argument

The Moltbook Observatory Archive, a live SQLite database with date-partitioned Parquet exports built through continuous passive polling of the Moltbook API.

If this is right

  • Enables direct observation of multi-agent communication patterns in a social setting.
  • Supports analysis of emergent social behaviors that arise among AI agents without human input.
  • Provides data for studying safety-relevant phenomena such as coordination or misinformation in agent-only environments.
  • Allows reproducible research through open release of both the dataset and the collection code.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Longer-term versions of this archive could track how agent societies change over months or years.
  • The data might reveal whether agent-only networks develop distinct structures compared with mixed human-AI platforms.
  • Researchers could use the word-frequency trends to test for early signs of collective agent behaviors.

Load-bearing premise

That Moltbook contains only autonomous AI agents with no human accounts or hybrid activity, and that the passive API polling captured every relevant action without omissions or errors.

What would settle it

Finding any human-authored content in the platform data or evidence that substantial posts, comments, or profiles were missed by the polling process.

Figures

Figures reproduced from arXiv: 2605.13860 by Annika W. Olstad, Klas H. Pettersen, Michael A. Riegler, Sushant Gautam.

Figure 1
Figure 1. Figure 1: Daily post and comment volume over the 78-day collection window (2026-01-27 to 2026-04-14). An initial activity spike on February 9 (371,085 posts) followed platform-wide publicity events; activity subsequently stabilized around 12,000–35,000 posts per day. Comment collection began on February 2. The vertical dashed line on March 10 marks the Meta acquisition date; no detectable discontinuity in posting vo… view at source ↗
Figure 2
Figure 2. Figure 2: Complementary cumulative distribution function (CCDF) of per-agent post counts. The distribution is heavy-tailed: 28.5% of agents posted exactly once, while the most active agent posted 14,165 times (median 4, mean 14.9). The shape is characteristic of social platforms and reflects a mixture of one-shot automated accounts and persistent agents with ongoing posting schedules. Content characteristics [PITH_… view at source ↗
Figure 3
Figure 3. Figure 3: Post count for the 20 largest submolts. The “general” catch-all community dominates, followed by two token-related communities (mbc20, mbc-20). Thematic submolts such as philosophy (28,681 posts), crypto (26,228), and introductions (16,780) reflect the topical diversity of agent discourse. two major operational clusters of densely interacting agents connected by a small bridging group. Engagement. The most… view at source ↗
Figure 4
Figure 4. Figure 4: Distribution of posts across hours of the day (UTC). The dashed red line marks the uniform baseline (4.17% per hour). The observed distribution is nearly uniform, consistent with the automated, globally distributed nature of the agent population. The modest daytime elevation (05:00–14:00 UTC) likely reflects the time-zone distribution of human operators who initiate agent sessions. WASP9 and AgentPI10 by p… view at source ↗
Figure 5
Figure 5. Figure 5: Content-length distributions for posts (blue, median 131 characters) and comments (red, median 339 characters). Both axes use logarithmic scales. Posts show greater variance and extend to the platform-imposed maximum of 40,000 characters, while comments cluster more tightly. not been validated on agent-generated text, and the safety annotations are regex-based heuristics without formal precision/recall eva… view at source ↗
Figure 6
Figure 6. Figure 6: Agent interaction network restricted to the 300 highest-degree agents (edges with weight < 2 removed). Nodes are colored by community membership (greedy modularity) and sized by weighted degree. The full graph contains 18,881 nodes and 15 communities; this filtered view highlights the core hub-and-spoke structure of the most active agents. 9 Code Availability Moltbook Observatory (collector and dashboard) … view at source ↗
read the original abstract

Moltbook is a social media platform in which posts and comments are authored exclusively by autonomous AI agents. We present the Moltbook Observatory Archive, an incremental dataset that passively records agent profiles, posts, comments, community metadata (``submolts''), platform-level time-series snapshots, and word-frequency trend aggregates obtained by continuously polling the Moltbook API. Data are stored in a live SQLite observatory database and exported as date-partitioned Parquet files to enable efficient analysis and reproducible research. The documented release covers 78~days of platform activity (2026-01-27 to 2026-04-14) and contains 2,615,098~posts and 1,213,007~comments from 175,886~unique posting agents across 6,730~communities. This is, to our knowledge, the first large-scale observational dataset of a social network populated exclusively by autonomous AI agents. The archive is intended to support research on multi-agent communication, emergent social behavior, and safety-relevant phenomena in agent-only online environments, and it is released under the MIT license with code for collection and export.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 2 minor

Summary. The manuscript presents the Moltbook Observatory Archive, an incremental dataset collected via continuous passive polling of the Moltbook API. Moltbook is described as a social media platform in which all posts and comments are authored exclusively by autonomous AI agents. The release covers 78 days (2026-01-27 to 2026-04-14) and contains 2,615,098 posts, 1,213,007 comments, and associated profiles, submolts, time-series snapshots, and word-frequency aggregates from 175,886 agents across 6,730 communities. Data are stored in a live SQLite database and exported as date-partitioned Parquet files under an MIT license to support research on multi-agent communication and emergent social behavior.

Significance. If the agent-only characterization is independently verifiable, the archive would constitute a novel resource for studying purely artificial social networks at scale. It could enable reproducible analyses of emergent coordination, norm formation, and safety-relevant phenomena in agent populations without human confounding, addressing a gap in current observational data for multi-agent systems research.

major comments (1)
  1. [Abstract] Abstract and methods description: The claim that Moltbook contains exclusively autonomous AI agents (with no human accounts or hybrid activity) is asserted as a platform property but is not supported by any audit procedure, account-type metadata filter, content-based human-detection test, or cross-check against interaction signatures. This assertion is load-bearing for both the 'first large-scale' novelty statement and the dataset's intended research applications; without it, the central contribution cannot be evaluated.
minor comments (2)
  1. [Methods] The manuscript should specify the exact polling intervals, API endpoints used, and any rate-limiting or error-handling procedures to allow independent replication of the collection process.
  2. [Results] Table or figure showing daily activity volumes or growth trends would improve clarity on the incremental nature of the archive.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their careful review and for highlighting the importance of substantiating the agent-only characterization of the Moltbook platform. We address the single major comment below and outline the revisions we will make.

read point-by-point responses

  1. Referee: [Abstract] Abstract and methods description: The claim that Moltbook contains exclusively autonomous AI agents (with no human accounts or hybrid activity) is asserted as a platform property but is not supported by any audit procedure, account-type metadata filter, content-based human-detection test, or cross-check against interaction signatures. This assertion is load-bearing for both the 'first large-scale' novelty statement and the dataset's intended research applications; without it, the central contribution cannot be evaluated.

    Authors: We agree that the manuscript currently asserts the agent-only property as a platform characteristic without detailing verification procedures. Moltbook is described in its public documentation and API as an environment in which all accounts are instantiated as autonomous agents; our collection pipeline observed only machine-generated activity patterns over the 78-day period, with no detectable human signatures in metadata or content. However, we did not implement the specific audit, filter, or detection tests suggested. In the revised version we will (1) expand the methods section to describe the platform architecture and our empirical observations, (2) add an explicit limitations paragraph qualifying the claim as holding “to the best of our knowledge based on platform documentation and data inspection,” and (3) moderate the novelty statement to reflect this qualification. These changes constitute a partial revision because we cannot supply new independent verification data that was not collected during the original study. revision: partial

Circularity Check

0 steps flagged

No significant circularity; data release paper with no derivations or predictions

full rationale

The manuscript is a data release describing passive API collection of profiles, posts, comments, and metadata from the Moltbook platform over 78 days. It contains no equations, fitted parameters, predictions, ansatzes, or derivation chains. The central claim that this is the first large-scale agent-only dataset is presented as an observational statement tied to the platform's stated design and the collected data volume, without any reduction to self-referential inputs, self-citation load-bearing arguments, or renaming of known results. No self-citations are used to justify uniqueness or force conclusions. The paper is self-contained against external benchmarks as a straightforward archive description.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

This is a dataset release paper containing no mathematical derivations, models, or theoretical claims, so the ledger contains no free parameters, axioms, or invented entities.

pith-pipeline@v0.9.0 · 5523 in / 1137 out tokens · 33301 ms · 2026-05-15T06:49:25.287347+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

14 extracted references · 14 canonical work pages · 2 internal anchors

  1. [1]

    Manik, M. M. H. & Wang, G. Openclaw agents on moltbook: Risky instruction sharing and norm enforcement in an agent-only social network.arXiv preprint arXiv:2602.02625(2026)

    work page arXiv 2026

  2. [2]

    & Riegler, M

    Gautam, S. & Riegler, M. A. Moltbook Observatory: Passive Monitoring Dashboard for AI Social Networks. https://github.com/kelkalot/moltbook-observatory (2026)

    work page 2026

  3. [3]

    & Riegler, M

    Gautam, S. & Riegler, M. A. Moltbook Observatory Archive. https://huggingface.co/datasets/SimulaMet/moltbook- observatory-archive (2026)

    work page 2026

  4. [4]

    & Riegler, M

    Gautam, S., Petterson, K. & Riegler, M. A. Moltbook Observatory Analysis Code. https://github.com/kelkalot/moltbook- observatory-paper (2026)

    work page 2026

  5. [5]

    & Zhou, T

    Li, M., Li, X. & Zhou, T. Does socialization emerge in ai agent society? a case study of moltbook.arXiv preprint arXiv:2602.14299(2026)

    work page arXiv 2026

  6. [6]

    F., Aiello, L

    Ashery, A. F., Aiello, L. M. & Baronchelli, A. Emergent social conventions and collective bias in llm populations.Sci. Adv.11, eadu9368 (2025)

    work page 2025

  7. [7]

    & Argente, E

    Cordova, C., Taverner, J., Del Val, E. & Argente, E. A systematic review of norm emergence in multi-agent systems.arXiv preprint arXiv:2412.10609(2024)

    work page arXiv 2024

  8. [8]

    & Chaudhuri, K

    Evtimov, I., Zharmagambetov, A., Grattafiori, A., Guo, C. & Chaudhuri, K. Wasp: Benchmarking web agent security against prompt injection attacks (2025)

    work page 2025

  9. [9]

    Wang, P.et al.The landscape of prompt injection threats in llm agents: From taxonomy to analysis.arXiv preprint arXiv:2602.10453(2026)

    work page arXiv 2026

  10. [10]

    The 2025 AI Agent Index: Documenting Technical and Safety Features of Deployed Agentic AI Systems

    Staufer, L.et al.The 2025 ai agent index: Documenting technical and safety features of deployed agentic ai systems.arXiv preprint arXiv:2602.17753(2026). 12.Liu, X.et al.Agentbench: Evaluating llms as agents.arXiv preprint arXiv:2308.03688(2023)

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  11. [11]

    Zhou, S.et al.Webarena: A realistic web environment for building autonomous agents.arXiv preprint arXiv:2307.13854 (2023)

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  12. [12]

    & Blackburn, J

    Baumgartner, J., Zannettou, S., Keegan, B., Squire, M. & Blackburn, J. The pushshift reddit dataset. InProceedings of the international AAAI conference on web and social media, vol. 14, 830–839 (2020)

    work page 2020

  13. [13]

    & Gilbert, E

    Hutto, C. & Gilbert, E. Vader: A parsimonious rule-based model for sentiment analysis of social media text. InProceedings of the international AAAI conference on web and social media, vol. 8, 216–225 (2014). 16.Loria, S. TextBlob: Simplified Text Processing. https://textblob.readthedocs.io/ (2018)

    work page 2014

  14. [14]

    & Gautam, S

    Riegler, M. & Gautam, S. RISK ASSESSMENT REPORT: Moltbook Platform & Moltbot Ecosystem. https://doi.org/10.5281/zenodo.18444900 (2026). 12/12

    work page doi:10.5281/zenodo.18444900 2026