SIGPSY — Special Interest Group in Psychohistory

Studying long-range historical modeling and prediction — drawing on quantitative history, complexity science, and protocol theory to develop frameworks for understanding civilizational-scale dynamics. The group maintains worldmachines.org, a collaborative platform for psychohistorical modeling.

Led by Venkatesh Rao and Aneesh Sathe — Biweekly Thursdays, 4pm UTC

← All Special Interest Groups

Meeting Archive — 8 sessions

  • September 10, 2026
    SIGPSY Symposium Planning & More is Different Paper Discussion
    Symposium special session structure and planningHierarchies of laws across scientific disciplinesEmergence and complexity in physical systemsInterdisciplinary approaches to psychohistory

    Participants: _vgr, boredgargoyle, drevius.

    The SIGPSY group convened to accomplish two goals: previewing the structure of an upcoming symposium special session and discussing P.W. Anderson's foundational 1972 paper 'More is Different.' The meeting began with a brief delay as one participant joined late, after which the group prioritized reading Anderson's paper, which addresses how organizational hierarchies in physical systems generate emergent laws that cannot be explained by lower-level physics alone. This concept of emergence and hierarchical complexity appears central to understanding psychohistorical phenomena. After the reading portion, the group planned to conduct a round of comments before transitioning to detailed planning of the symposium special session.

    • The group planned to read P.W. Anderson's 'More is Different' paper, which explores how different levels of organization in physical systems follow distinct laws that cannot be reduced to lower levels.
    • The meeting adopted a structured approach: 15-20 minutes of paper reading followed by quick comments, then transition to symposium planning discussions.
    • Anderson's work on hierarchies of laws from physics to higher-order systems appears directly relevant to psychohistory methodology and understanding complexity.
    Links discussed
    View discussion in Discord →
  • August 27, 2026
    SIGPSY 27Aug26: Self-fulfilling Prophecies
    Self-fulfilling prophecy dynamics and theoryHyperstitional theory fiction and Nick Land's workRobert K. Merton's foundational 1948 frameworkInterdisciplinary literature survey on self-fulfilling propheciesStudy track development for extended exploration

    Participants: _vgr, kylemathews, jdbb, drevius.

    The SIGPSY group convened to examine self-fulfilling prophecy dynamics as their theory week topic. Organizer _vgr identified Robert K. Merton's seminal 1948 work as the main reading and discovered extensive interdisciplinary literature on the subject, which exceeds what can be covered in a single session. In addition to the core theoretical framework, the group noted connections to hyperstitional theory fiction and Nick Land's ideas previously discussed. Due to scheduling conflicts, both _vgr and kylemathews were unable to attend the live meeting, with kylemathews planning to contribute notes afterward and _vgr noting that Aneesh would lead the call instead. The group decided to establish a dedicated study track to systematically explore the broader literature on self-fulfilling prophecies across multiple disciplines.

    • Self-fulfilling prophecy is a well-established concept across multiple disciplines with substantial literature beyond the hyperstitional theory framework previously discussed by the group.
    • Robert K. Merton's 1948 work 'The Self-Fulfilling Prophecy' serves as the primary theoretical foundation for this session's exploration.
    • The topic warrants extended investigation through a dedicated study track due to the breadth of relevant literature across different fields.
    • Hyperstitional theory fiction represents one application of self-fulfilling prophecy dynamics but is part of a much larger theoretical landscape.
    Links discussed
    View discussion in Discord →
  • August 13, 2026
    SIGPSY Update on Prime Radiant and World Machines
    Prime Radiant platform updates and featuresWorld Machines project overviewOracle question recommendation systemCrisis prediction and confidence trackingContext tank progression

    Participants: _vgr, boredgargoyle, rafa_0x, sn37, smsteven

    Aneesh gave an extended walkthrough of the software he has been building for the group: a note-taking/wiki app ("Scribe"), the "Oracle" retrieval system built on the group's corpus of books and essays, and the "Prime Radiant" — a temporal knowledge-graph system that models history as interacting "machines" and generates extrapolations/"prophecies." The second half turned to planning the upcoming symposium/special session (involving Aneesh, Sean, and Florian), which Venkatesh framed as the project's public launch. Discussion touched on how notes and essays could eventually feed automatically into the modeling system, and on how contributions and funding might scale afterward.

    • **Project context (Aneesh):** The group began as a book club on Venkatesh's Substack, evolved into members writing and cross-referencing essays, and became the "World Machines" project. The core framing (attributed to Venkatesh) is that history can be read as different "machines" in play — e.g., a modernity machine, a "versions" machine, a "liveness" machine.
    • **The Oracle:** An AI system that ingests the group's material (books, members' essays, external essays) into a searchable database so members can ask questions or explore, ideally lowering the barrier to writing essays. It functions as retrieval-and-answer ("it doesn't do prophecy yet"), and behind the scenes searches a **knowledge graph**, taking several back-and-forth steps across related links before answering.
    • **Two-tier vs. MCP access (Aneesh):** The portal AI is described as "fairly low powered" and uses a two-tier agent setup (the Oracle queries a second agent with access to raw notes, which returns summaries). A dedicated **MCP** option lets members connect their own, more capable AI directly to the corpus (including the raw text) via a generated access token.
    • **Scribe app (Aneesh):** `scribe.worldmachines.org` — a lightweight web app (installable as a PWA on iPhone via "Add to Home Screen") for quickly adding notes to the corpus. Motivation: editing/uploading text files was too much friction and meant he wasn't writing anything. Notes can be quick thoughts, book notes, or "velocity notes" / term definitions (Venkatesh had wanted to define some terms).
    • **Notes as corpus / "hive mind" (Aneesh):** The purpose of adding a note is that it becomes a source the Oracle can cite. Aneesh described it as a hive mind where members' thoughts enter the shared sourcing and others (and their AIs) can build on them.
    • **Provenance/tagging (Aneesh):** There's no truth filter, but every note is attributed to a person or tagged as coming from a book — so a book source may be more trustworthy than a personal opinion.
    Links discussed
    View discussion in Discord →
  • July 30, 2026
    Cliodynamics and Mathematical History: Theory, Methods, and Predictive Modeling
    Peter Turchin's cliodynamics framework and mathematical approaches to historyIn-sample vs. out-of-sample model validation and predictive testingOntologies and AI as tools for historical analysisMeteorological forecasting as a comparative model for historical predictionEvent dynamics as graph traversal with structural feedbackPattern recognition and the psychology of divination systems

    Participants: _vgr, boredgargoyle, matt_ms, drevius.

    SIGPSY convened to examine Peter Turchin's cliodynamics approach, which attempts to apply mathematical and scientific methods to historical analysis. The group reviewed introductory materials on mathematical history alongside critical commentary, noting that Turchin's combative style generates significant debate. A central methodological discussion emerged around model validation: while comparing theoretical predictions to empirical data eliminates weaker models, the presence of tunable parameters weakens conclusions unless tested against out-of-sample data reserved for validation. The group drew parallels to meteorological forecasting as a mature domain combining physics-based reasoning with AI, suggesting such hybrid approaches might strengthen cliodynamic predictions.

    The conversation expanded into broader conceptual territory, exploring how divination systems work psychologically—achieving apparent success through reliable on-demand pattern detection combined with sophisticated obfuscation for failures. Participants proposed that historical events might be modeled as paths through evolving graphs where traversal itself restructures the space, implying historical causation is fundamentally non-linear and feedback-generating. A final thread suggested that trauma (and its opposites) become physically encoded in bodies as unreliable pre-cognitive signaling systems, connecting neurobiology to historical dynamics. These discussions blend cliodynamics proper with adjacent questions about prediction, pattern recognition, embodiment, and causal structure.

    • Testing theoretical models against empirical data is valuable for elimination, but strengthening results requires out-of-sample validation rather than relying on tunable parameters.
    • Meteorological forecasting provides a useful analogy for model-based forecasting in historical contexts, suggesting physics-informed AI approaches could apply to cliodynamics.
    • Divination systems succeed through reliable on-demand pattern detection combined with plausible deniability for failures—a phenomenon worth analyzing for understanding how pattern-seeking works in prediction generally.
    • Historical events may function like paths through a graph whose structure they simultaneously modify, suggesting a dynamic feedback model where past events reshape the causal landscape for future events.
    • Trauma and its inverse may encode physically in body structures as unreliable pre-cognitive signals, pointing to embodied mechanisms that influence future decision-making.
    Links discussed
    View discussion in Discord →
  • July 16, 2026
    Oracle v2.0 System Design Discussion
    Oracle system v2.0 design and updatesLive deployment of updated oracle systemHuman-to-human interaction model for oracle interfacePermissions transfer and contributor access managementSite migration and email verification process

    Participants: _vgr, boredgargoyle, florianlow, matt_ms

    Aneesh Sathe walked the group through the software he has been building — an "Oracle" question-answering system backed by a note/notes corpus, a copyright-respecting "Witness" retrieval agent, an ingestion pipeline for books, and a "Prime Radiant" prophecy tool — including a live demo of the World Machines website. Alongside the technical walkthrough, the group did a "side quest" round where each person shared their current reading. A substantive discussion emerged about how to handle input quality: whether to editorially gate contributed essays (Brandon's concern about "poisoning" the corpus) versus deliberately "leaning into the noise" (Venkatesh's position), plus ideas for lowering the barrier to contributing content.

    • **Ingestion pipeline:** breaks books down, creates notes about the books, and then creates "canon notes" that place material into the psycho-history mind-space and flag what's relevant. Only ~4 books ingested so far, already producing many chunks; ingesting one book can take a whole night. Currently runs on Aneesh's computer; hopes to move to the cloud and get help ingesting more books.
    • **Two MCP servers** now exist: an "article" one (returns text from the notes, not raw text) usable by anyone with their own agent, and a **Witness MCP** that can return raw-text chunks but is heavily gated to accepted contributors only.
    • **Prime Radiant:** a history-mapping / prophecy tool that generates "lead cards" capturing specific historic "machines" (e.g., a "paper states machine"), models how machines interact and their codependencies, and aims to produce prophecies that respect both the affordances and limits of each machine. Currently works roughly like the article (has historical events but lacks a good interpretive lens); expected to improve as more notes and essays accumulate.
    • **Brandon:** raised concern about editorial control / corpus "poisoning" — how essays get into the Oracle's bucket, and whether there should be a review process to keep low-quality or off-thesis material out.
    • **Venkatesh:** argued the opposite bias — *don't* gate for quality; build capacity to handle messy, low-signal, even adversarial input. He distinguished two kinds of noise: (1) good but low-quality/extraneous material, and (2) deliberate "poisoning" (e.g., a future rival tribe of psycho-historians or a cult trying to corrupt the Oracle). A prophecy engine that can't accommodate reactions to itself is "fundamentally not doing the job" — tying to hyperstition (self-fulfilling prophecy) and its adversarial mirror. Cited **Philip Tetlock's superforecasting** work: the best forecasters update priors rapidly on weak signals, so the system should "metabolize" low quality rather than reject it. Coined the doctrine phrase **"lean into the noise."**
    • **Matthew:** offered support via his reading — history (Darnton, Braudel) shows reality is "ridiculously messy and noisy," so leaning into noise may be more faithful to how history actually happens; liked the "poetic, rhyming" angle.
    Links discussed
    View discussion in Discord →
  • July 2, 2026
    Three Temporalities: Eventful Sociology and Historical Time
    Sewell's critique of teleological vs. experimental temporality in sociologyEventful temporality as a third alternative conception of historical timeAgency, resistance, and capacity to act within historical structuresNon-teleological modes of anticipation and abductive reasoningEvent as a loaded concept bridging software architecture and narrative theory

    Participants: _vgr, brndnpink, boredgargoyle, matt_ms, sachbenny, kylemathews, florianlow

    The group did a round-robin of first impressions on a dense paper distinguishing three ways of thinking about historical time, before opening to broader discussion. Participants largely agreed that "eventful temporality" — with its emphasis on path dependence, contingency, and the coupling of micro and macro — is the most promising frame and the one closest to the group's own "world machine"/psycho-history modeling project. Venkatesh closed with an extended build on the paper, proposing that historical events are distinguished from ordinary "commodity events" by high dimensionality and narrative feedback loops that locate human agency.

    • **Path dependence & contingency (Sean):** History is deeply path-dependent and contingent; it's neither reducible to grand teleological narratives nor to repeatable experiments. Sean noted Sewell's analogy to paleontology (a science that respects both contingency and broader unfolding stories) and the concept of "eventfulness" as where the micro meets the macro.
    • **Structures are mutable (Sean):** Important events don't just involve people — they also shift structures (e.g. capitalism), which are not immutable.
    • **Teleology and politics (Sachin):** Political movements often lean on teleological narratives to imply continuity with predecessors ("people before you were just like you"). Cited the French Revolution, where violence was justified by tying it back to Rome's founding story and to Christian ideas of sacrifice — an attempt to construct a teleological narrative around revolutionary sacrifice. Suggested "pop history" tends teleological because it aims to influence present politics.
    • **Shorthand for the three frames (Matthew):** Teleological = a grand law converging on an outcome; experimental = a transferable/general process across space and time; eventful = emergent local iteration/evolution that propagates upward to a global state. Offered a nature/nurture-and-poverty analogy (nature ≈ teleological; nurture ≈ experimental process; the blend/constraints of contingency ≈ eventful) but said he couldn't fully untangle it.
    • **Can a constantly-changing subject be a science? (Hatim):** Appreciated Sewell situating historical sociology between causation-obsessed sociology and narrative-detail-obsessed history. Raised reflexivity — society reacts to being studied — as a challenge, noting sociologists disagree on how much it matters. Linked this to Wolfram's computational irreducibility: how reducible is society?
    • **Reconciling patterns vs. randomness (Florian, echoed by Andre):** The paper opens a middle path where you can see patterns without becoming Spengler-style deterministic — path dependency plus decisive moments that can break either way. Florian connected "events" to the group's earlier "crisis moments" discussion, and argued the paper's evolutionary/anthropological "driving engine" fits the group's project better than the physics engine of the Foundation model. Noted the postscript defends eventful temporality against the charge of being too microscale, arguing it can scale to civilizational events.
    Links discussed
    View discussion in Discord →
  • June 18, 2026
    Seldon Vault Design and Oracle Backend
    Seldon vault design specificationsWebsite feature development for vault predictionsOracle backend implementation and progressPsychohistory prediction storage and retrieval

    Participants: _vgr, boredgargoyle

    The SIGPSY meeting focused on developing a Seldon vault feature for the Protocol Institute's website. _vgr initiated discussion about designing the vault more carefully and creating a corresponding website feature to hold vault predictions. The group recognized the need for thoughtful design work before implementation. boredgargoyle contributed by sharing progress on a new oracle backend that would serve as the technical infrastructure for the vault predictions system. The meeting appears to be in the planning and architectural phase of this feature development.

    • The group is working on designing a dedicated Seldon vault feature for their website to systematically store and manage psychohistory predictions.
    • An oracle backend is being developed as a technical foundation to support vault predictions functionality.
    • The discussion indicates a need for careful architectural planning before implementing the vault design on the website.
    View discussion in Discord →
  • June 4, 2026
    SIGPSY Kickoff: Psychohistory, Prediction Markets, and Narrative-Driven Forecasting
    Oracle project update and agentic predictionsEssay series for USA 500-year anniversary (July 4, 250y)Asimov's Foundation as gas theory model of history with chokepointsPrediction markets redesign: narrative-driven vs binary approachesCyberpunk avoidance vs solarpunk preference in future scenariosPrime Radiant integration with current forecasting systems

    Participants: _vgr, boredgargoyle, randylubin, kylemathews, kurosh1669, smsteven

    SIGPSY held its inaugural meeting to discuss psychohistory, prediction methodology, and organizational structure. The group reviewed an oracle project involving agentic predictions and planned an essay collection for the USA's 500-year historical milestone. A central discussion examined how Asimov's Foundation series modeled history as a gas-like system with critical chokepoints—anticipating chaos theory—where intervention could alter outcomes. The group debated approaches to prediction markets, concluding that binary prediction markets are insufficient for their goals. Instead, they advocated for a narrative-driven, agentic forecasting approach that emphasizes storytelling and long-term scenario exploration over short-term quantifiable bets. They also noted that publishing narratives openly attracts intellectually engaged participants. The group considered whether to frame futures through a cyberpunk avoidance lens (emphasizing challenges) or solarpunk preference (emphasizing solutions), and discussed integrating a Prime Radiant website with current forecasting infrastructure.

    • History behaves like a gas where most events cannot be controlled, but certain chokepoints exist where deliberate action can redirect outcomes—a concept Asimov intuited from chaos theory before its formal development.
    • Rather than traditional binary prediction markets, the group proposes an agentic, narrative-driven approach focused on storytelling as a mechanism to attract engaged participants and explore long-term scenarios.
    • Pre-committed predictions that self-eliminate upon failure create accountability mechanisms; the group is exploring how to implement this with narrative depth rather than simple market mechanics.
    • Publishing open narratives and stories serves dual purposes: exploring psychohistorical ideas while attracting intellectually interesting collaborators to the Protocol Institute.
    Links discussed
    View discussion in Discord →

Interested in joining? Join the Discord community →