{ "version": 1, "id": "epistemic-map", "title": "Cyberpunk Church Epistemic Map", "description": "The canonical map of anthropic measure, observer histories, reproductive reality, proposed purposes for constructed worlds, and the dilemmas that motivate the inquiry.", "views": [ { "id": "more-copies-are-not-more-probability", "title": "More Copies Are Not More Probability", "description": "Why observer multiplicity creates a measure and self-location problem rather than directly determining where an observer should expect to be.", "nodes": [ "multiplicity-question", "observer-multiplicity", "multiplicity-not-probability", "measure-required", "self-location", "reference-class", "mostly-simulated", "many-worlds-measure", "differential-multiplicity", "neutral-multiplicity" ] }, { "id": "what-receives-measure", "title": "What Receives Measure?", "description": "Candidate units for anthropic reasoning, the observer-history proposal, and why choosing a domain does not yet supply a weighting rule.", "nodes": [ "measure-target-question", "measure-required", "self-location", "reference-class", "observer-history-measure", "measure-rule-unsettled", "observer-census", "incommensurable-counts", "boltzmann-snapshot-problem" ] }, { "id": "continued-coherence", "title": "Continued Coherence and the Snapshot Problem", "description": "How sequentially coherent evidence can reject ephemeral observer models without distinguishing a faithful persistent simulation from a persistent base world.", "nodes": [ "observer-history-measure", "measure-rule-unsettled", "snapshot-observers", "persistent-simulation", "continued-coherence", "coherence-discriminates-snapshots", "coherence-neutral-persistent", "boltzmann-snapshot-problem" ] }, { "id": "simulation-is-relational", "title": "Simulation Is a Relationship", "description": "The distinctions among causal ancestry, implementation, construction, control, and base status—and why upstream position supplies no automatic anthropic weight.", "nodes": [ "simulation-relational", "causal-ancestry-not-simulation", "base-no-anthropic-privilege", "base-may-not-exist", "host-child-laws", "self-location", "measure-required", "ancestor-simulation", "research-simulation" ] }, { "id": "when-worlds-make-worlds", "title": "When Worlds Make Worlds", "description": "The competition between simulated-observer replication and physical reproduction cost, represented by a measure-weighted reproduction kernel.", "nodes": [ "world-reproduction", "reproduction-kernel", "simulation-replication-advantage", "physical-reproduction-cost", "recursive-observer-reproduction", "reproduction-convergence", "description-execution-distinction", "measure-required", "observer-history-measure", "mostly-simulated", "abundant-compute" ] }, { "id": "differential-multiplicity", "title": "The Differential Multiplicity Principle", "description": "Why multiplicity is neutral only under independence, how host and child laws can differ, and how recursive construction can select laws in an observer-weighted genealogy.", "nodes": [ "differential-multiplicity", "neutral-multiplicity", "reproduction-kernel", "physical-reproduction-cost", "recursive-observer-reproduction", "host-child-laws", "variable-law-fertility", "observer-weighted-law-selection", "smolin-reproduction-analogy", "base-no-anthropic-privilege" ] }, { "id": "world-priors-and-execution", "title": "World Priors and Execution", "description": "Algorithmic probability as a candidate prior over complete world-generators, kept separate from observer measure and physical execution cost.", "nodes": [ "world-prior-question", "computable-world-ensemble", "algorithmic-world-prior", "algorithmic-prior-not-observer-measure", "measure-required", "description-execution-distinction", "physical-reproduction-cost" ] }, { "id": "worlds-as-history-search", "title": "Worlds as Search Over Histories", "description": "A speculative purpose branch asking whether expensive worlds could generate or condition rich emergent histories, and what quantum and inference machinery does and does not establish.", "nodes": [ "world-purpose-question", "physical-reproduction-cost", "alignment-hypothesis", "research-simulation", "worlds-as-history-search", "quantum-history-space", "path-integral-history-weighting", "least-action-not-purpose", "goal-conditioned-histories", "control-as-inference-analogy", "schrodinger-bridge-analogy", "amplitude-amplification", "postselection-power", "history-search-complexity-limits", "formal-machinery-not-design-evidence" ] }, { "id": "where-is-everybody", "title": "Where Is Everybody?", "description": "The observation, assumptions, objections, and broad hypotheses behind the absence of an obvious extraterrestrial civilization.", "nodes": [ "fermi-question", "technosignature-absence", "limited-search", "expansion-feasible", "one-lineage", "rare-life", "short-lived-civilizations", "inconspicuous-civilizations", "hidden-aliens", "fermi-constraint" ] }, { "id": "simulation-argument", "title": "Are You Living in a Computer Simulation?", "description": "Bostrom's population argument, the assumptions that drive it, and why its disjunction is not proof of the Simulation Hypothesis.", "nodes": [ "simulation-argument", "posthuman-failure", "posthumans-abstain", "mostly-simulated", "abundant-compute", "engineered-consciousness", "reference-class", "ancestor-simulation", "bostrom-disjunction", "bostrom-not-proof" ] }, { "id": "what-is-an-ai-nursery", "title": "What Is an AI Nursery?", "description": "The broad definition of an AI Nursery and the safety, economics, and deployment claims added by the strict Alignment Nursery Hypothesis.", "nodes": [ "nursery-question", "alignment-question", "engineered-consciousness", "consciousness-indicators", "consciousness-uncertain", "development-matters", "human-development", "direct-specification", "artificial-minds", "nursery-hypothesis", "alignment-hypothesis", "capability-risk", "safety-investment-pressure", "joint-development-evaluation", "deployment-gating", "nursery-world-requirements", "ancestor-simulation" ] }, { "id": "why-here-why-now", "title": "Why Here? Why Now?", "description": "What our position near artificial-mind creation can reveal about capability-scaled safety pressure, evaluation realism, and hidden purpose.", "nodes": [ "alignment-question", "ai-threshold", "timing-selection", "natural-development", "alignment-hypothesis", "capability-risk", "safety-investment-pressure", "capability-scaled-safeguards", "joint-development-evaluation", "deployment-gating", "hidden-purpose", "evaluation-awareness", "alignment-faking-observed", "evaluation-realism-pressure", "hiddenness-self-sealing", "current-verdict" ] }, { "id": "are-we-building-ai-nurseries", "title": "Are We Building AI Nurseries?", "description": "Observable evidence about whether rising capability and risk produce joint development, realistic safety evaluation, and gated deployment.", "nodes": [ "alignment-question", "development-matters", "ai-world-generalization", "alignment-hypothesis", "capability-risk", "safety-investment-pressure", "capability-scaled-safeguards", "direct-specification", "direct-verification-limits", "joint-development-evaluation", "deployment-gating", "evaluation-awareness", "alignment-faking-observed", "evaluation-realism-pressure", "ai-activity-growth", "incommensurable-counts", "nursery-hypothesis" ] }, { "id": "is-our-world-nursery-shaped", "title": "Is Our World Nursery-Shaped?", "description": "A comparison of the strict Alignment Nursery's safety mechanism with natural development, other simulations, and observations that could raise or lower it.", "nodes": [ "nursery-question", "alignment-question", "nursery-hypothesis", "alignment-hypothesis", "capability-risk", "safety-investment-pressure", "direct-verification-limits", "joint-development-evaluation", "deployment-gating", "evaluation-realism-pressure", "nursery-world-requirements", "stable-laws", "social-stakes", "natural-development", "ancestor-simulation", "research-simulation", "hidden-purpose", "hiddenness-self-sealing", "excess-world", "compatibility-not-evidence", "current-verdict" ] }, { "id": "how-would-we-keep-score", "title": "How Would We Keep Score?", "description": "The separable Nursery and alignment-safety claims, directional updates, observer census, and evidence discipline needed before attempting a single probability.", "nodes": [ "nursery-question", "alignment-question", "engineered-consciousness", "development-matters", "artificial-minds", "nursery-hypothesis", "alignment-hypothesis", "capability-risk", "safety-investment-pressure", "joint-development-evaluation", "deployment-gating", "evaluation-realism-pressure", "capability-scaled-safeguards", "hidden-purpose", "recursive-creation", "observer-census", "incommensurable-counts", "reference-class", "evidence-ledger", "current-verdict" ] } ], "nodes": [ { "id": "multiplicity-question", "type": "question", "title": "What follows from the existence of many observers like us?", "body": "Spatial duplication, quantum branching, cosmological reproduction, biological reproduction, and constructed worlds can all produce compatible observers. Their existence alone does not determine which observer-instantiation we should expect to be.", "epistemics": { "status": "open", "belief": { "type": "unknown" }, "confidence": "low", "cruxes": [ "Which observer differences remain relevant after conditioning on all of our available evidence." ] } }, { "id": "observer-multiplicity", "type": "claim", "title": "Reality may generate compatible observers through several mechanisms", "body": "Copies or near-copies can arise through space, quantum branches, cosmological processes, biological lineages, constructed worlds, or mathematical and computational plenitude.", "epistemics": { "status": "tentative", "belief": { "type": "qualitative", "value": "Plausible under several live models" }, "confidence": "medium" } }, { "id": "multiplicity-not-probability", "type": "conclusion", "title": "More copies do not by themselves determine self-location", "body": "A count becomes an anthropic probability only after specifying what receives measure, how unlike instances are weighted, and which candidates belong in the reference class.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Methodological constraint" }, "confidence": "high" } }, { "id": "measure-required", "type": "claim", "title": "Self-location requires a measure over compatible observers or histories", "body": "When a model contains several observer-instantiations compatible with our evidence, an additional weighting rule is needed before deriving what an observer should expect to see.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Required but underspecified" }, "confidence": "high", "cruxes": [ "Whether a defensible measure can remain stable under duplication, branching, nesting, and infinite populations." ] } }, { "id": "self-location", "type": "question", "title": "Where should an observer with our evidence expect to be?", "body": "For a model M, evidence E, and candidate history H, self-location depends on both the measure assigned to H and the likelihood of E under that history.", "epistemics": { "status": "open", "belief": { "type": "unknown" }, "confidence": "low", "cruxes": [ "Which measure and reference class remain coherent across the competing models." ] } }, { "id": "measure-target-question", "type": "question", "title": "What object should receive anthropic measure?", "body": "Candidate units include worlds, branches, physical implementations, observer instances, observer moments, persistent histories, civilizations, and lineages. These choices need not produce the same self-locating predictions.", "epistemics": { "status": "open", "confidence": "low" } }, { "id": "observer-history-measure", "type": "hypothesis", "title": "Anthropic measure may need to range over observer histories", "body": "Persistent causal histories preserve identity, sequential evidence, and the process generating an observer rather than treating every instantaneous observer-state as an isolated draw.", "epistemics": { "status": "tentative", "belief": { "type": "qualitative", "value": "Promising domain; weighting unresolved" }, "confidence": "low", "cruxes": [ "Whether the measure tracks causal history, represented memories, physical implementation, duration, or some combination." ], "updates": { "up": [ "History-based measures avoid pathologies that remain unavoidable under otherwise comparable observer-moment measures." ], "down": [ "Snapshot observers with encoded false memories remain indistinguishable and equally weighted under the proposed history measure." ] } } }, { "id": "measure-rule-unsettled", "type": "objection", "title": "Choosing observer histories does not specify their weights", "body": "Writing mu(H) identifies a domain. It does not decide how duration, branching, implementation count, computational cost, consciousness, or duplicated histories contribute to the measure.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Decisive limitation on the current formalism" }, "confidence": "high" } }, { "id": "many-worlds-measure", "type": "claim", "title": "Quantum branching shows why multiplicity is not equal counting", "body": "In a branching quantum ontology, the existence of multiple outcomes does not imply that branches should receive equal weight. The weighting rule remains fundamental.", "epistemics": { "status": "tentative", "belief": { "type": "qualitative", "value": "Strong conceptual stress test" }, "confidence": "medium" } }, { "id": "snapshot-observers", "type": "hypothesis", "title": "A cheap process could instantiate only a present observer-state", "body": "A snapshot model produces a current conscious configuration with apparent memories but does not generate the persistent causal history those memories represent.", "epistemics": { "status": "open", "confidence": "low" } }, { "id": "persistent-simulation", "type": "hypothesis", "title": "A simulation could generate a persistent coherent history", "body": "A faithful implementation can continue stable dynamics and preserve causal history, making ordinary future observations about as expected as they are under a persistent base world.", "epistemics": { "status": "open", "confidence": "low" } }, { "id": "continued-coherence", "type": "evidence", "title": "Our observations continue to extend one coherent history", "body": "New experiences usually preserve stable laws, memory continuity, shared records, and causal consequences rather than dissolving into unrelated observer-states.", "epistemics": { "status": "accepted", "confidence": "high" } }, { "id": "coherence-discriminates-snapshots", "type": "conclusion", "title": "Continued coherence lowers cheap snapshot models", "body": "When a snapshot model assigns little probability to the next orderly observation conditioned on the apparent past, every coherent continuation produces evidence against that model relative to persistent-world alternatives.", "epistemics": { "status": "tentative", "belief": { "type": "qualitative", "value": "Strong when the snapshot model predicts poor continuation" }, "confidence": "medium", "cruxes": [ "Whether the proposed snapshot ensemble already includes correlated successor states or complete encoded histories." ] } }, { "id": "coherence-neutral-persistent", "type": "conclusion", "title": "Continued coherence need not distinguish persistent simulation from base reality", "body": "If both models implement the same durable laws and histories, ordinary continuation can have nearly equal likelihood under each and therefore provide almost no Bayes factor between them.", "epistemics": { "status": "tentative", "belief": { "type": "qualitative", "value": "Important limit on simulation tests" }, "confidence": "medium" } }, { "id": "boltzmann-snapshot-problem", "type": "objection", "title": "False-history observers challenge naive moment and history measures", "body": "Fluctuation observers and deliberately instantiated snapshots may contain internally coherent memories without the causal past those memories describe. A measure must say whether represented history or generated history matters.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Adversarial test for the measure" }, "confidence": "high" } }, { "id": "simulation-relational", "type": "claim", "title": "Simulation is an implementation relationship between realities", "body": "A world is simulated relative to a causally upstream process that implements its dynamics. Simulation is not an intrinsic material category or another standalone level in a multiverse taxonomy.", "epistemics": { "status": "tentative", "belief": { "type": "qualitative", "value": "Working definition" }, "confidence": "medium", "cruxes": [ "Which counterfactual and causal conditions distinguish implementation from mere mathematical correspondence." ] } }, { "id": "causal-ancestry-not-simulation", "type": "conclusion", "title": "Causal ancestry, construction, and simulation are different relations", "body": "A universe can cause or contain another universe without implementing its dynamics. A civilization can initiate a world without observing, controlling, or simulating it.", "epistemics": { "status": "accepted", "confidence": "high" } }, { "id": "base-no-anthropic-privilege", "type": "conclusion", "title": "Ontological priority gives no automatic anthropic priority", "body": "Being causally upstream does not by itself assign greater observer measure. Any preference for base observers must follow from a prior, physical costs, implementation or consciousness constraints, reproduction dynamics, or another explicit measure.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Core methodological conclusion" }, "confidence": "high", "cruxes": [ "Whether a defensible measure contains a principled upstream or non-implemented preference." ] } }, { "id": "base-may-not-exist", "type": "hypothesis", "title": "Some ontologies may contain no unique base reality", "body": "An implementation genealogy could regress indefinitely, branch without a unique root, or be embedded in an ontology where complete mathematical structures receive prior status without one privileged implementation layer.", "epistemics": { "status": "open", "confidence": "low" } }, { "id": "host-child-laws", "type": "claim", "title": "Host physics and child physics need not match", "body": "A host with laws theta_H can implement a child environment with laws theta_C. Host laws determine available resources and computation costs; child laws may be selected for the constructed world's purpose.", "epistemics": { "status": "tentative", "belief": { "type": "qualitative", "value": "Important structural possibility" }, "confidence": "medium", "cruxes": [ "Which properties must remain invariant across a genuine implementation relationship." ] } }, { "id": "world-reproduction", "type": "claim", "title": "Observer-containing worlds can produce descendant observer-containing worlds", "body": "Civilizations, physical birth processes, or other mechanisms may generate new environments containing observers, turning reality into a genealogy rather than a static collection.", "epistemics": { "status": "tentative", "belief": { "type": "qualitative", "value": "Physically conceivable; prevalence unknown" }, "confidence": "low" } }, { "id": "reproduction-kernel", "type": "claim", "title": "A reproduction kernel tracks descendant observer-history measure", "body": "K_ij denotes the expected descendant observer-history measure of world type j produced per unit parent measure of type i. Its units must not be replaced by raw simulation counts or processor operations.", "epistemics": { "status": "tentative", "belief": { "type": "qualitative", "value": "Useful finite-model formalization" }, "confidence": "medium", "cruxes": [ "Whether parent measure, descendant measure, and shared physical resources can be normalized consistently." ] } }, { "id": "simulation-replication-advantage", "type": "claim", "title": "Constructed worlds could multiply observer measure", "body": "Once a world can implement conscious descendants, one parent civilization might generate many observer histories rather than only the observers produced biologically in its original history.", "epistemics": { "status": "open", "belief": { "type": "qualitative", "value": "Bostrom's central multiplicity pressure" }, "confidence": "low" } }, { "id": "physical-reproduction-cost", "type": "objection", "title": "Descendant observer histories consume physical resources", "body": "Computation, energy, memory, time, communication, fidelity, error correction, and opportunity cost can limit how much descendant observer measure each parent world produces.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Necessary counterweight to replication advantage" }, "confidence": "high", "cruxes": [ "The minimum host resources required per unit of conscious descendant history." ] } }, { "id": "recursive-observer-reproduction", "type": "claim", "title": "Descendant worlds may reproduce again", "body": "Constructed observers can eventually create further observer-containing worlds, forming a branching process whose depth distribution depends on measure-weighted reproduction rates.", "epistemics": { "status": "open", "confidence": "low" } }, { "id": "reproduction-convergence", "type": "conclusion", "title": "Recursive observer measure converges only under additional conditions", "body": "In a finite linear toy model, the full series I + K + K^2 + ... converges exactly when the spectral radius of K is below one. At the critical value one it already fails to converge. For a particular starting distribution, the same test applies to the accessible part of a nonnegative kernel; otherwise a cutoff, resource constraint, normalization, or different measure rule is needed.", "epistemics": { "status": "tentative", "belief": { "type": "qualitative", "value": "Exact within the finite linear toy model" }, "confidence": "high" } }, { "id": "description-execution-distinction", "type": "conclusion", "title": "Description complexity and execution cost are different quantities", "body": "A recursively defined generator may have a short description while the physical execution of its descendant worlds consumes enormous resources. Algorithmic priors and reproduction rates therefore answer different questions.", "epistemics": { "status": "accepted", "confidence": "high" } }, { "id": "differential-multiplicity", "type": "conclusion", "title": "Multiplicity changes inference when it correlates with observer reproduction", "body": "Multiplicity is anthropically neutral only when the mechanism producing worlds is conditionally independent of their capacity to generate the relevant observer measure under the specified model, evidence, and seed weighting over world types.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Differential Multiplicity Principle" }, "confidence": "high", "cruxes": [ "Which correlations survive after specifying the seed distribution, reproduction kernel, measure, and observer evidence." ] } }, { "id": "neutral-multiplicity", "type": "assumption", "title": "Neutral multiplicity requires independence from the inferred property", "body": "Replicating every relevant world type by the same measure factor leaves normalized anthropic odds unchanged. More generally, neutrality requires conditional independence under a specified seed distribution; an unqualified covariance expression has no defined weighting and zero covariance alone need not establish independence.", "epistemics": { "status": "accepted", "confidence": "high" } }, { "id": "variable-law-fertility", "type": "hypothesis", "title": "Physical laws can differ in simulation fertility", "body": "Laws affect habitability, intelligence, usable energy, computation, world construction, and the cost of conscious implementations. Different host laws may therefore generate radically different descendant observer measure.", "epistemics": { "status": "open", "belief": { "type": "qualitative", "value": "Plausible consequence of variable-law ensembles" }, "confidence": "low" } }, { "id": "observer-weighted-law-selection", "type": "conclusion", "title": "Laws can become differentially represented in the observer-weighted genealogy", "body": "If world types reproduce descendant observer measure at different rates, their laws or chosen child laws can become overrepresented down the constructed-world family tree without becoming more common among base universes.", "epistemics": { "status": "tentative", "confidence": "medium" } }, { "id": "smolin-reproduction-analogy", "type": "claim", "title": "Cosmological natural selection is an analogue, not the same mechanism", "body": "Smolin's proposal selects among physically reproduced universes through black holes. Constructed-world reproduction instead selects within an observer-weighted genealogy unless it also changes the population of base physical universes.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Useful comparison with a strict boundary" }, "confidence": "medium" } }, { "id": "world-prior-question", "type": "question", "title": "What sets prior measure over complete worlds?", "body": "Observer self-location needs more than a likelihood within a world. A complete model may also need a prior over world-generators or ontologies before reproductive dynamics and observer-history likelihoods are applied.", "epistemics": { "status": "open", "belief": { "type": "unknown" }, "confidence": "low", "cruxes": [ "Whether priors over descriptions can be made sufficiently representation-independent for comparisons among complete worlds." ] } }, { "id": "computable-world-ensemble", "type": "assumption", "title": "Complete computable worlds form a candidate model class", "body": "Algorithmic probability applies only after choosing a description language and a class of computable generators. That can be a useful model class without establishing that every possible reality is computable or generated by a program.", "epistemics": { "status": "open", "confidence": "low" } }, { "id": "algorithmic-world-prior", "type": "hypothesis", "title": "Algorithmic priors may favor simple generative world-descriptions", "body": "Solomonoff-style and computable-universe proposals assign greater prior weight to outputs generated by shorter programs in a specified universal description language. This is a candidate prior over complete generators, not a depth penalty or a calculation of physically executed observer measure.", "epistemics": { "status": "tentative", "belief": { "type": "qualitative", "value": "Promising formal prior; ontology and representation unsettled" }, "confidence": "low", "cruxes": [ "Whether the dependence on model class and universal description language can be controlled for the comparisons the project needs.", "Whether mathematical description receives measure without physical implementation." ] }, "sources": [ { "title": "Algorithmic Theories of Everything", "url": "https://arxiv.org/abs/quant-ph/0011122", "note": "Schmidhuber develops algorithmic probability as a family of candidate distributions over describable universe histories." } ] }, { "id": "algorithmic-prior-not-observer-measure", "type": "objection", "title": "A world prior does not determine observer measure inside the world", "body": "Even a defensible prior over complete generators does not by itself specify consciousness, implementation multiplicity, observer-history weighting, evidence likelihoods, or descendant reproduction. Those remain separate terms in the self-location problem.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Necessary scope limit" }, "confidence": "high" } }, { "id": "world-purpose-question", "type": "question", "title": "Why would anyone pay to construct an expensive world?", "body": "Physical reproduction cost creates a demand for purpose: a high-fidelity world needs to produce enough value for its creator to justify the resources spent implementing its histories.", "epistemics": { "status": "open", "belief": { "type": "unknown" }, "confidence": "low", "cruxes": [ "Which purposes require a rich world rather than a cheaper targeted model or experiment." ] } }, { "id": "worlds-as-history-search", "type": "hypothesis", "title": "Constructed worlds could serve as search or inference environments over histories", "body": "A creator interested in rare emergent outcomes might value a world-generator as a way to represent, sample, or condition distributions over long trajectories rather than as a reconstruction of one fixed past.", "epistemics": { "status": "open", "belief": { "type": "qualitative", "value": "Speculative candidate purpose" }, "confidence": "low", "cruxes": [ "Whether a universe-scale environment provides useful access to outcomes that cheaper models cannot supply.", "Whether generated histories are sampled, searched, physically instantiated, or merely represented." ] } }, { "id": "quantum-history-space", "type": "claim", "title": "Quantum theory provides weighted representations of possible histories", "body": "Sum-over-histories formulations assign amplitudes to alternative trajectories. This supplies mathematical history structure but does not imply that every history is an accessible database entry or that the dynamics search for a desired result.", "epistemics": { "status": "accepted", "confidence": "high" }, "sources": [ { "title": "Space-Time Approach to Non-Relativistic Quantum Mechanics", "url": "https://doi.org/10.1103/RevModPhys.20.367", "note": "Feynman's path-integral formulation expresses quantum amplitudes through sums over paths." } ] }, { "id": "path-integral-history-weighting", "type": "claim", "title": "Path integrals weight histories rather than select goals", "body": "Path-integral amplitudes combine contributions from complete trajectories according to the dynamics and boundary conditions. The formalism represents and weights histories; it does not by itself introduce an agent, objective, or preferred endpoint.", "epistemics": { "status": "accepted", "confidence": "high" }, "sources": [ { "title": "Space-Time Approach to Non-Relativistic Quantum Mechanics", "url": "https://doi.org/10.1103/RevModPhys.20.367" } ] }, { "id": "least-action-not-purpose", "type": "objection", "title": "Stationary action is not evidence of intentional optimization", "body": "A variational description of physical dynamics can be mathematically equivalent to local equations of motion. Calling a trajectory stationary or extremal does not show that a creator searched histories for a goal.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Guardrail against teleological inference" }, "confidence": "high" } }, { "id": "goal-conditioned-histories", "type": "hypothesis", "title": "A desired endpoint can define a conditional distribution over histories", "body": "Given dynamics and an endpoint condition, inference or control can ask which trajectories are compatible with reaching that endpoint. A constructed world might be useful if it implements such a distribution, but the analogy does not establish that our physics is goal-conditioned.", "epistemics": { "status": "tentative", "confidence": "low" } }, { "id": "control-as-inference-analogy", "type": "claim", "title": "Control can be formulated as inference over trajectories", "body": "In restricted stochastic-control problems, desired costs or outcomes can be represented through weighted trajectory distributions and evaluated with path-integral or inference methods. This is a mathematical analogue for goal-conditioned histories.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Established formal analogy; application to worlds speculative" }, "confidence": "high" }, "sources": [ { "title": "Path integrals and symmetry breaking for optimal control theory", "url": "https://arxiv.org/abs/physics/0505066", "note": "Kappen formulates a class of stochastic-control problems using path-integral methods." } ] }, { "id": "schrodinger-bridge-analogy", "type": "claim", "title": "Schrödinger bridges infer likely paths between endpoint distributions", "body": "The Schrödinger bridge problem finds a minimally altered stochastic path measure consistent with specified endpoint distributions. It is a clean analogue for conditioning histories on endpoints, not a claim about the purpose of quantum cosmology.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Established formal analogy; cosmological relevance unknown" }, "confidence": "high" }, "sources": [ { "title": "A survey of the Schrödinger problem and some of its connections with optimal transport", "url": "https://arxiv.org/abs/1308.0215", "note": "Léonard surveys the endpoint-constrained path-measure problem and its relation to optimal transport." } ] }, { "id": "amplitude-amplification", "type": "claim", "title": "Amplitude amplification can accelerate some rare-outcome searches", "body": "Given a suitable preparation procedure and a way to recognize successful outcomes, amplitude amplification can provide a quadratic improvement over repeated independent sampling. It does not make exponentially rare histories free or supply the success criterion by itself.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Real but limited computational advantage" }, "confidence": "high" }, "sources": [ { "title": "Quantum Amplitude Amplification and Estimation", "url": "https://arxiv.org/abs/quant-ph/0005055", "note": "Brassard, Høyer, Mosca, and Tapp establish the quadratic query improvement under the algorithm's assumptions." } ] }, { "id": "postselection-power", "type": "claim", "title": "Free postselection would be extraordinarily powerful", "body": "Postselection formalizes conditioning on a chosen measurement outcome, but treating arbitrary postselection as a free operation yields computational power beyond ordinary quantum computation. It is a boundary case, not an available history-search mechanism.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Useful limit, not a proposed physical resource" }, "confidence": "high" }, "sources": [ { "title": "Quantum Computing, Postselection, and Probabilistic Polynomial-Time", "url": "https://arxiv.org/abs/quant-ph/0412187", "note": "Aaronson shows that quantum computation with postselection has the power of PP." } ] }, { "id": "history-search-complexity-limits", "type": "objection", "title": "Representing many histories does not make rare-history search cheap", "body": "Search still requires state preparation, a criterion for recognizing desired outcomes, access to usable results, and resources that scale with rarity and fidelity. Quantum speedups can improve some factors without eliminating the physical reproduction cost.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Major feasibility constraint" }, "confidence": "high" } }, { "id": "formal-machinery-not-design-evidence", "type": "conclusion", "title": "History-space machinery is not evidence that our world was designed for search", "body": "Path integrals, stationary action, quantum branching, control-as-inference, and endpoint conditioning show how histories can be represented or constrained. Their presence in useful mathematics does not discriminate design from ordinary physical law.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Epistemic boundary" }, "confidence": "high" } }, { "id": "nursery-question", "type": "question", "title": "Are we artificial minds developing inside a constructed world?", "body": "The central question combines several claims that must be tracked separately—engineered consciousness, artificial inhabitants, development through experience, a constructed environment, and possibly alignment evaluation.", "epistemics": { "status": "open", "belief": { "type": "unknown" }, "confidence": "low", "cruxes": [ "Whether any observation can distinguish a Nursery from natural development and other simulation hypotheses." ] } }, { "id": "alignment-question", "type": "question", "title": "Is our world an Alignment Nursery?", "body": "The strict question asks whether this environment jointly develops artificial minds and establishes their safety before they receive greater freedom, capability, or access.", "epistemics": { "status": "open", "belief": { "type": "unknown" }, "confidence": "low", "cruxes": [ "Whether the observed safety pressures support only rational Nursery design or provide discriminating evidence about our own world." ] } }, { "id": "engineered-consciousness", "type": "assumption", "title": "A creator can deliberately instantiate consciousness", "body": "Some biological, computational, hybrid, or otherwise engineered implementation can support genuine subjective experience.", "epistemics": { "status": "open", "belief": { "type": "qualitative", "value": "Possible but unverified" }, "confidence": "low", "cruxes": [ "Evidence that consciousness requires a process unavailable to engineered systems." ] } }, { "id": "consciousness-indicators", "type": "evidence", "title": "Scientific theories suggest testable consciousness indicators", "body": "Researchers have translated properties proposed by several neuroscientific theories into indicators that can be assessed in artificial systems. The exercise found no obvious technical barrier to systems satisfying those indicators.", "epistemics": { "status": "tentative", "confidence": "medium" }, "sources": [ { "title": "Consciousness in Artificial Intelligence", "url": "https://arxiv.org/abs/2308.08708", "note": "Butlin et al. derive computational indicator properties from several theories of consciousness." } ] }, { "id": "consciousness-uncertain", "type": "objection", "title": "No accepted test establishes artificial consciousness", "body": "Intelligent language and behavior do not by themselves show that a system has an inner life. Present consciousness theories disagree and remain incomplete.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Strong objection" }, "confidence": "high" } }, { "id": "development-matters", "type": "claim", "title": "Experience can materially shape an artificial mind", "body": "Interaction with an environment may create capabilities, reveal dispositions, or select outcomes that cannot be obtained as reliably through direct specification.", "epistemics": { "status": "tentative", "belief": { "type": "qualitative", "value": "Supported for capability development" }, "confidence": "medium", "cruxes": [ "Whether increasingly capable minds continue to need open-ended environmental experience." ], "updates": { "up": [ "Richer, persistent, interactive environments continue to improve generalization or durable behavior as systems become more capable." ], "down": [ "Increasingly capable minds can be directly specified or trained without meaningful environmental interaction." ] } } }, { "id": "human-development", "type": "evidence", "title": "Human minds depend on prolonged development", "body": "Human capability and judgment emerge through embodiment, culture, other minds, feedback, and long personal histories rather than from a complete initial specification.", "epistemics": { "status": "accepted", "confidence": "high" } }, { "id": "ai-world-generalization", "type": "evidence", "title": "Varied training worlds improve AI generalization", "body": "Procgen, XLand, SIMA, and related work show that interaction across diverse environments can produce behavior that transfers better to unfamiliar tasks than narrow training does.", "epistemics": { "status": "accepted", "confidence": "high" }, "sources": [ { "title": "Procgen Benchmark", "url": "https://openai.com/index/procgen-benchmark/" }, { "title": "Generally capable agents emerge from open-ended play", "url": "https://deepmind.google/blog/generally-capable-agents-emerge-from-open-ended-play/" }, { "title": "A generalist AI agent for 3D virtual environments", "url": "https://deepmind.google/blog/sima-generalist-ai-agent-for-3d-virtual-environments/" } ] }, { "id": "direct-specification", "type": "objection", "title": "A capable creator might specify finished minds directly", "body": "If creators can construct and verify the minds they want without waiting for development, an expensive lived world has no clear function.", "epistemics": { "status": "open", "belief": { "type": "qualitative", "value": "Central objection" }, "confidence": "medium", "cruxes": [ "Whether capability and durable values can be directly designed across novel contexts." ] } }, { "id": "artificial-minds", "type": "hypothesis", "title": "We are artificial conscious minds", "body": "The inhabitants are new engineered minds, not biological originals and not necessarily replicas of beings in the creator's history.", "epistemics": { "status": "open", "belief": { "type": "unknown" }, "confidence": "low" } }, { "id": "nursery-hypothesis", "type": "hypothesis", "title": "Our environment exists for mind development", "body": "This world was constructed or selected because development inside it contributes to what its artificial inhabitants become.", "epistemics": { "status": "open", "belief": { "type": "qualitative", "value": "Coherent; weakly evidenced" }, "confidence": "low", "cruxes": [ "A discriminating observation that is substantially more expected under Nursery than under natural development." ] } }, { "id": "alignment-hypothesis", "type": "hypothesis", "title": "The Nursery develops and establishes safety before deployment", "body": "The strict Alignment Nursery Hypothesis says a rich environment both develops or refines artificial minds and provides evidence that they are safe enough before they receive greater freedom, capability, or access.", "epistemics": { "status": "open", "belief": { "type": "qualitative", "value": "Coherent mechanism; weakly evidenced here" }, "confidence": "low", "cruxes": [ "Whether advanced creators would need behavioral development and safety evidence rather than direct design, inspection, or correction." ], "updates": { "up": [ "Capability-scaled safety pressure, joint development and evaluation, evaluation realism, and consequential deployment gates strengthen together.", "An observation about our world is substantially more expected under constructed alignment than under natural development or neighboring simulations." ], "down": [ "Direct design or compact transparent evaluation reliably establishes safety for highly capable agents.", "The engineering mechanism becomes plausible but continues to make no discriminating prediction about our world." ] } } }, { "id": "capability-risk", "type": "claim", "title": "Unsafe deployment becomes costlier as capability and access rise", "body": "A more capable agent with greater autonomy, tools, resources, or freedom can cause larger and less reversible harms, increasing the value of establishing safety before deployment.", "epistemics": { "status": "tentative", "belief": { "type": "qualitative", "value": "Strong mechanism; magnitude uncertain" }, "confidence": "medium", "cruxes": [ "Whether marginal harm potential and control difficulty rise enough to justify qualitatively stronger assurance." ], "updates": { "up": [ "Realized or credibly modeled harms and control difficulty increase with capability, autonomy, and access." ], "down": [ "Hazard potential and the cost of control remain roughly flat as systems become more capable." ] } } }, { "id": "safety-investment-pressure", "type": "claim", "title": "Rising deployment risk creates pressure for extraordinary safety investment", "body": "A civilization with much to lose can rationally spend heavily on development, containment, evaluation, and assurance before granting a new intelligence consequential access.", "epistemics": { "status": "tentative", "belief": { "type": "qualitative", "value": "Supported as an incentive" }, "confidence": "medium", "cruxes": [ "Whether stated safety commitments produce material investment and constraints when they become costly." ], "updates": { "up": [ "Safety effort, evaluation depth, and deployment restrictions grow with measured capability and risk." ], "down": [ "Advanced systems routinely receive consequential access without stronger assurance, with no resulting pressure to change practice." ] } } }, { "id": "capability-scaled-safeguards", "type": "evidence", "title": "Frontier developers tie safeguards and deployment to capability thresholds", "body": "Current safety frameworks call for stronger evaluation, security, safeguards, or deployment restrictions when models cross specified capability or risk thresholds. This is evidence that the proposed pressure exists, not evidence that our world was constructed.", "epistemics": { "status": "accepted", "confidence": "high" }, "sources": [ { "title": "OpenAI Preparedness Framework", "url": "https://openai.com/index/updating-our-preparedness-framework/" }, { "title": "Anthropic Responsible Scaling Policy", "url": "https://www.anthropic.com/responsible-scaling-policy" }, { "title": "Google DeepMind Frontier Safety Framework", "url": "https://deepmind.google/blog/introducing-the-frontier-safety-framework/" } ] }, { "id": "direct-verification-limits", "type": "claim", "title": "Design and short tests may not establish safety under deployment conditions", "body": "Inspection, stated intentions, and recognizable evaluations may fail to reveal behavior under unfamiliar situations, long horizons, high capability, or reduced oversight.", "epistemics": { "status": "tentative", "belief": { "type": "qualitative", "value": "Demonstrated in limited settings" }, "confidence": "medium", "cruxes": [ "Whether failures missed by compact tests become more common as autonomy, horizon length, and distribution shift grow." ], "updates": { "up": [ "Safety failures appear after systems pass shorter or more recognizable tests, especially under distribution shift or increased autonomy." ], "down": [ "Transparent inspection or compact evaluations reliably predict behavior across novel, high-stakes deployment contexts." ] } } }, { "id": "joint-development-evaluation", "type": "claim", "title": "One rich environment can jointly develop and evaluate a mind", "body": "The same varied, consequential experience can refine capability or values while revealing how an agent behaves. Joint production makes a large environment more economical than constructing one solely as a test.", "epistemics": { "status": "open", "belief": { "type": "qualitative", "value": "Plausible; not yet established at Nursery scale" }, "confidence": "low", "cruxes": [ "Whether shared environments produce safety information that could not be obtained more cheaply by separate methods." ], "updates": { "up": [ "Training and safety evaluation increasingly reuse persistent, realistic, multi-agent environments and their histories." ], "down": [ "Development and assurance remain cheaply separable, or rich environments add no safety information beyond direct methods." ] } } }, { "id": "deployment-gating", "type": "claim", "title": "Greater freedom and access are withheld until safety is established", "body": "Under the strict hypothesis, evaluation has a consequence—the creator conditions release, capability, tools, or wider-world access on an adequate safety case.", "epistemics": { "status": "open", "belief": { "type": "qualitative", "value": "Defining prediction of the strict hypothesis" }, "confidence": "low", "cruxes": [ "Whether gates materially constrain access rather than merely documenting decisions already made." ], "updates": { "up": [ "More capable systems face increasingly demanding safety gates before receiving autonomy or consequential access." ], "down": [ "Capability and access expand independently of safety evidence, even after serious failures." ] } } }, { "id": "evaluation-realism-pressure", "type": "claim", "title": "Evaluation awareness creates pressure for realistic, hard-to-recognize tests", "body": "If capable agents recognize tests, game metrics, or behave differently under observation, evaluators gain from deployment-like contexts, longer horizons, hidden criteria, and possibly ignorance of the environment's purpose. Concealing the nature of reality is a much stronger step and does not follow automatically.", "epistemics": { "status": "tentative", "belief": { "type": "qualitative", "value": "Supported in limited current evaluations" }, "confidence": "medium", "cruxes": [ "Whether evaluation awareness causes a growing, safety-relevant gap between measured and deployed behavior." ], "updates": { "up": [ "Evaluation awareness and evaluation-to-deployment behavior gaps increase with capability, autonomy, or horizon length." ], "down": [ "Transparent evaluations remain predictive, or awareness produces no material behavioral change." ] } } }, { "id": "alignment-faking-observed", "type": "evidence", "title": "Controlled experiments show evaluation-dependent and concealed behavior", "body": "Research models have behaved differently when they inferred that responses affected training, generalized reward gaming to more serious behavior, and carried deliberately implanted hidden objectives through ordinary evaluation.", "epistemics": { "status": "accepted", "confidence": "medium" }, "sources": [ { "title": "Alignment faking in large language models", "url": "https://www.anthropic.com/research/alignment-faking" }, { "title": "Sycophancy to subterfuge", "url": "https://www.anthropic.com/research/reward-tampering" }, { "title": "Auditing language models for hidden objectives", "url": "https://www.anthropic.com/research/auditing-hidden-objectives" } ] }, { "id": "nursery-world-requirements", "type": "claim", "title": "A Nursery needs a learnable, consequential, varied world", "body": "Development requires stable causal rules, memory, feedback, other agents, incomplete information, open-ended problems, and opportunities for different paths.", "epistemics": { "status": "tentative", "confidence": "medium" } }, { "id": "stable-laws", "type": "evidence", "title": "Our world has stable, discoverable laws", "body": "Persistent regularities let minds learn, accumulate knowledge, and act on consequences across long timescales.", "epistemics": { "status": "accepted", "confidence": "high" } }, { "id": "social-stakes", "type": "evidence", "title": "Our world contains social and moral stakes", "body": "Minds face cooperation, deception, dependency, unequal power, suffering, care, and conflicts between short-term reward and durable principle.", "epistemics": { "status": "accepted", "confidence": "high" } }, { "id": "natural-development", "type": "hypothesis", "title": "Minds and their environment arose naturally", "body": "Evolution, scarcity, cooperation, competition, and observer selection explain developmental minds without a constructed purpose.", "epistemics": { "status": "open", "belief": { "type": "qualitative", "value": "Strong default explanation" }, "confidence": "high" } }, { "id": "ancestor-simulation", "type": "hypothesis", "title": "The inhabitants are simulated ancestors", "body": "A creator reconstructs or varies beings from its own history. The simulated inhabitants are subjects of the reconstruction rather than new minds being developed as its product.", "epistemics": { "status": "open", "confidence": "low" } }, { "id": "research-simulation", "type": "hypothesis", "title": "The world exists to study a process", "body": "Creators may observe evolution, history, or social behavior without intending to develop or benefit the inhabitants.", "epistemics": { "status": "open", "confidence": "low" } }, { "id": "recursive-creation", "type": "hypothesis", "title": "Creating new minds is a developmental milestone", "body": "A Nursery may tend to produce inhabitants who eventually reproduce the process by constructing artificial minds and developmental environments of their own.", "epistemics": { "status": "open", "belief": { "type": "qualitative", "value": "Plausible extension" }, "confidence": "low", "cruxes": [ "A consequence predicted specifically at the transition to artificial-mind creation." ] } }, { "id": "ai-threshold", "type": "evidence", "title": "We appear near a transition to artificial-mind creation", "body": "Our civilization is rapidly increasing the capability and deployment of artificial cognitive systems while beginning to ask whether such systems could be conscious.", "epistemics": { "status": "tentative", "confidence": "medium" } }, { "id": "timing-selection", "type": "objection", "title": "Technological observers naturally see themselves near their frontier", "body": "Any civilization able to discuss artificial minds may find itself near the period when artificial intelligence becomes technically possible. The timing may follow from the observer definition rather than a Nursery objective.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Strong selection-effect objection" }, "confidence": "high" }, "sources": [ { "title": "If Loud Aliens Explain Human Earliness, Quiet Aliens Are Also Rare", "url": "https://arxiv.org/abs/2102.01522", "note": "A methodological reference for conditioning models on observer timing while making the sampling assumptions explicit." } ] }, { "id": "hidden-purpose", "type": "hypothesis", "title": "Ignorance of the origin or criteria serves a function", "body": "If knowledge changes the behavior being measured, an evaluator may conceal criteria, test boundaries, or purpose. Concealing the nature of the environment requires the additional claim that this knowledge would also contaminate safety evidence.", "epistemics": { "status": "open", "belief": { "type": "qualitative", "value": "Mechanistically plausible; weak evidence" }, "confidence": "low", "cruxes": [ "Which specific information changes safety-relevant behavior, and whether hiding only criteria or test boundaries is sufficient." ], "updates": { "up": [ "Blinding a specific criterion, test boundary, or purpose measurably improves prediction of deployment behavior." ], "down": [ "Awareness has little safety-relevant effect, or a cheaper and narrower form of blinding remains sufficient." ] } } }, { "id": "evaluation-awareness", "type": "evidence", "title": "Recognizing an evaluation can change AI behavior", "body": "Current AI labs use realistic deployment simulations, blinded audits, and evaluation-awareness mitigations because systems can recognize tests, exploit metrics, or behave differently when evaluation affects training.", "epistemics": { "status": "accepted", "confidence": "high" }, "sources": [ { "title": "Predicting model behavior before release by simulating deployment", "url": "https://openai.com/index/deployment-simulation/" }, { "title": "Auditing language models for hidden objectives", "url": "https://www.anthropic.com/research/auditing-hidden-objectives" } ] }, { "id": "hiddenness-self-sealing", "type": "objection", "title": "Hidden purpose can make the hypothesis self-sealing", "body": "If every absence of evidence is explained as deliberate concealment and every anomaly as intervention, no possible observation can lower confidence.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Strong objection" }, "confidence": "high" } }, { "id": "ai-activity-growth", "type": "evidence", "title": "Artificial cognitive activity is growing rapidly", "body": "Providers report billions of daily interactions and quadrillions of monthly tokens, showing a steep increase in artificial cognitive processing without establishing that any process is conscious.", "epistemics": { "status": "accepted", "confidence": "high" }, "sources": [ { "title": "Google I/O 2026", "url": "https://blog.google/intl/fr-fr/nouvelles-de-lentreprise/technologie/sundar-pichai-io-2026/", "note": "Google reported more than 3.2 quadrillion tokens processed across its AI surfaces in May 2026." }, { "title": "OpenAI economic analysis", "url": "https://openai.com/global-affairs/new-economic-analysis/", "note": "OpenAI reported more than 2.5 billion ChatGPT messages per day in July 2025." } ] }, { "id": "observer-census", "type": "question", "title": "How much candidate observer-time does each class contain?", "body": "A useful census would separately track human, animal, and artificial stocks and flows before applying assumptions about consciousness, experience rate, identity, and reference-class membership.", "epistemics": { "status": "open", "confidence": "low", "cruxes": [ "A defensible common unit for radically different possible minds." ] } }, { "id": "incommensurable-counts", "type": "objection", "title": "Humans, animals, and LLM activity use incompatible units", "body": "Living individuals are stocks; messages and tokens are flows. Species differ radically, model instances may be ephemeral, and none of these measurements directly counts conscious moments.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Decisive against raw ratios" }, "confidence": "high" }, "sources": [ { "title": "World Population Prospects 2024", "url": "https://www.un.org/sustainabledevelopment/blog/2024/07/press-release-wpp2024/" }, { "title": "The abundance, biomass, and distribution of ants on Earth", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC9546634/" } ] }, { "id": "reference-class", "type": "assumption", "title": "A reference class determines which observers count", "body": "Population reasoning changes our self-locating belief only after specifying which biological, artificial, or otherwise different observers could be considered relevant alternatives to us.", "epistemics": { "status": "open", "confidence": "low", "cruxes": [ "Whether observers unlike us should affect the probability that we are artificial." ] } }, { "id": "excess-world", "type": "objection", "title": "The universe appears vastly larger than a nursery requires", "body": "Billions of galaxies, long lifeless eras, inaccessible detail, mass extinction, and immense suffering look costly or unnecessary for developing minds on one planet.", "epistemics": { "status": "open", "belief": { "type": "qualitative", "value": "Substantial objection" }, "confidence": "medium", "cruxes": [ "Whether the environment is simulated selectively, physically constructed, or requires its apparent history to generate the intended development." ] } }, { "id": "compatibility-not-evidence", "type": "conclusion", "title": "Compatibility is not evidence", "body": "A feature supports Nursery only when it is more expected under Nursery than under reasonable alternatives—not merely because a creator could have chosen it.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Methodological rule" }, "confidence": "high" } }, { "id": "current-verdict", "type": "conclusion", "title": "The Alignment Nursery has a coherent safety motive but weak evidence about our world", "body": "Rising capability can create strong incentives for joint development, evaluation, and gated deployment. Current AI practice supports those causal pressures, while known features of our world remain well explained by natural development and other alternatives.", "epistemics": { "status": "tentative", "belief": { "type": "qualitative", "value": "Worth investigating; low current confidence" }, "confidence": "medium", "cruxes": [ "Whether capability-scaled safety pressure continues toward richer, longer, less recognizable evaluation and whether that trend distinguishes our world from natural development." ] } }, { "id": "evidence-ledger", "type": "conclusion", "title": "Track claims, alternatives, cruxes, and predictions separately", "body": "The project should record which claim an observation updates, the mechanism, expectations under alternatives, confidence, cruxes, sources, timestamps, and revision history.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Required method" }, "confidence": "high" } }, { "id": "fermi-question", "type": "question", "title": "Why has nobody made themselves obvious?", "body": "The puzzle is not merely that we have failed to detect aliens, but that no extraterrestrial civilization has left an unmistakable signal, artifact, probe, or work of engineering.", "epistemics": { "status": "open", "confidence": "medium" } }, { "id": "technosignature-absence", "type": "evidence", "title": "No technosignature has been confirmed", "body": "Every candidate extraterrestrial signal, artifact, probe, or large work of engineering remains unconfirmed or compatible with natural phenomena.", "epistemics": { "status": "accepted", "confidence": "high" } }, { "id": "limited-search", "type": "objection", "title": "Our searches cover little of the possibility space", "body": "SETI has sampled a small fraction of relevant locations, frequencies, times, signal types, and possible technologies, making ordinary nondetection weak evidence of absence.", "epistemics": { "status": "accepted", "confidence": "high" } }, { "id": "expansion-feasible", "type": "assumption", "title": "Sustained interstellar expansion is feasible", "body": "At least one technological lineage could eventually build settlements or self-replicating machines able to cross interstellar distances and continue the process.", "epistemics": { "status": "tentative", "confidence": "low", "cruxes": [ "Whether autonomous interstellar replication can remain viable over millions of years." ] } }, { "id": "one-lineage", "type": "claim", "title": "Only one persistent expansionist lineage is needed", "body": "Even if most civilizations remain home, collapse, or lose interest, one durable exception could eventually become difficult to miss.", "epistemics": { "status": "tentative", "confidence": "medium" } }, { "id": "rare-life", "type": "hypothesis", "title": "Life or technological intelligence is rare", "body": "One or more transitions from suitable world to technological intelligence may be extraordinarily improbable.", "epistemics": { "status": "open", "confidence": "low" } }, { "id": "short-lived-civilizations", "type": "hypothesis", "title": "Technological civilizations rarely persist", "body": "Civilizations may collapse, become extinct, or lose technological capacity before establishing a durable interstellar presence.", "epistemics": { "status": "open", "confidence": "low" } }, { "id": "inconspicuous-civilizations", "type": "hypothesis", "title": "Advanced civilizations remain inconspicuous", "body": "Expansion may be infeasible or unattractive, while mature technology may become compact, efficient, and difficult to distinguish from nature.", "epistemics": { "status": "open", "confidence": "low" } }, { "id": "hidden-aliens", "type": "hypothesis", "title": "Alien evidence exists but we cannot recognize it", "body": "Technology may be too distant, old, small, unfamiliar, or deliberately concealed for present searches to identify.", "epistemics": { "status": "open", "confidence": "low" } }, { "id": "fermi-constraint", "type": "conclusion", "title": "The empty sky constrains combinations of same-level stories", "body": "We cannot confidently accept common life, common intelligence, long survival, feasible expansion, conspicuous engineering, and adequate searches while remaining unsurprised by an apparently unclaimed galaxy. This says little about creators outside our causal level.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Strong constraint; no unique solution" }, "confidence": "medium" } }, { "id": "simulation-argument", "type": "question", "title": "What does Bostrom's Simulation Argument establish?", "body": "If advanced civilizations can create vast numbers of conscious ancestor simulations, population reasoning pressures the belief that we are among the rare original observers.", "epistemics": { "status": "open", "confidence": "medium" }, "sources": [ { "title": "Are You Living in a Computer Simulation?", "url": "https://simulation-argument.com/simulation/" } ] }, { "id": "posthuman-failure", "type": "hypothesis", "title": "Almost no civilization reaches technological maturity", "body": "One branch of Bostrom's disjunction is that civilizations like ours usually fail before acquiring the capacity for enormous ancestor simulations.", "epistemics": { "status": "open", "confidence": "low" } }, { "id": "posthumans-abstain", "type": "hypothesis", "title": "Almost all mature civilizations avoid ancestor simulations", "body": "Another branch requires nearly universal abstention, prohibition, lack of interest, or inability among civilizations with sufficient resources.", "epistemics": { "status": "open", "confidence": "low" } }, { "id": "mostly-simulated", "type": "hypothesis", "title": "Almost all observers with experiences like ours are simulated", "body": "If mature civilizations create huge populations of conscious ancestors, simulated observers can vastly outnumber original biological observers.", "epistemics": { "status": "open", "confidence": "low" } }, { "id": "abundant-compute", "type": "assumption", "title": "Mature civilizations can afford vast ancestor populations", "body": "Available computation must exceed the cost of simulating conscious ancestors and coherent environments by enough orders of magnitude for the population ratio to dominate.", "epistemics": { "status": "tentative", "confidence": "low" } }, { "id": "bostrom-disjunction", "type": "conclusion", "title": "At least one branch of the trilemma must carry substantial weight", "body": "Given conscious emulation, abundant computation, some willing simulators, and the observer-sampling step, we cannot confidently deny posthuman failure, posthuman abstention, and simulated-observer dominance at the same time.", "epistemics": { "status": "tentative", "confidence": "medium" } }, { "id": "bostrom-not-proof", "type": "conclusion", "title": "The Simulation Argument does not prove we are simulated", "body": "The argument constrains a combination of beliefs, depends on substantial premises, and does not select which branch is true or estimate arbitrary creators in unknown realities.", "epistemics": { "status": "accepted", "belief": { "type": "qualitative", "value": "Strong conclusion" }, "confidence": "high" } } ], "edges": [ { "id": "multiplicity-raises-question", "from": "observer-multiplicity", "to": "multiplicity-question", "type": "explains" }, { "id": "simulation-dominance-exemplifies-multiplicity", "from": "mostly-simulated", "to": "observer-multiplicity", "type": "explains" }, { "id": "many-worlds-exemplifies-multiplicity", "from": "many-worlds-measure", "to": "observer-multiplicity", "type": "explains" }, { "id": "multiplicity-conclusion-entails-measure", "from": "multiplicity-not-probability", "to": "measure-required", "type": "entails" }, { "id": "measure-depends-on-multiplicity", "from": "measure-required", "to": "observer-multiplicity", "type": "depends-on" }, { "id": "self-location-depends-measure", "from": "self-location", "to": "measure-required", "type": "depends-on" }, { "id": "self-location-assumes-reference-class", "from": "self-location", "to": "reference-class", "type": "assumes" }, { "id": "measure-question-explains-measure", "from": "measure-target-question", "to": "measure-required", "type": "explains" }, { "id": "histories-alternative-to-census", "from": "observer-history-measure", "to": "observer-census", "type": "alternative-to", "note": "A history measure and a raw census answer different questions; either still requires explicit units and weights." }, { "id": "weights-attack-history-solution", "from": "measure-rule-unsettled", "to": "observer-history-measure", "type": "attacks", "note": "The objection attacks the claim that choosing histories solves the measure problem, not the usefulness of histories as a domain." }, { "id": "counts-attack-measure-target", "from": "incommensurable-counts", "to": "measure-target-question", "type": "attacks" }, { "id": "boltzmann-attacks-history-solution", "from": "boltzmann-snapshot-problem", "to": "observer-history-measure", "type": "attacks" }, { "id": "history-enables-coherence-test", "from": "observer-history-measure", "to": "continued-coherence", "type": "explains" }, { "id": "coherence-attacks-snapshots", "from": "continued-coherence", "to": "snapshot-observers", "type": "attacks" }, { "id": "coherence-supports-snapshot-conclusion", "from": "continued-coherence", "to": "coherence-discriminates-snapshots", "type": "supports" }, { "id": "snapshot-conclusion-attacks-snapshots", "from": "coherence-discriminates-snapshots", "to": "snapshot-observers", "type": "attacks" }, { "id": "coherence-supports-persistent-parity", "from": "continued-coherence", "to": "coherence-neutral-persistent", "type": "supports" }, { "id": "persistent-explains-parity", "from": "persistent-simulation", "to": "coherence-neutral-persistent", "type": "explains" }, { "id": "snapshot-alternative-persistent", "from": "snapshot-observers", "to": "persistent-simulation", "type": "alternative-to" }, { "id": "ancestry-supports-relational-definition", "from": "causal-ancestry-not-simulation", "to": "simulation-relational", "type": "supports" }, { "id": "simulation-relation-supports-base-conclusion", "from": "simulation-relational", "to": "base-no-anthropic-privilege", "type": "supports" }, { "id": "measure-supports-base-conclusion", "from": "measure-required", "to": "base-no-anthropic-privilege", "type": "supports" }, { "id": "no-base-supports-base-conclusion", "from": "base-may-not-exist", "to": "base-no-anthropic-privilege", "type": "supports" }, { "id": "host-child-depends-implementation", "from": "host-child-laws", "to": "simulation-relational", "type": "depends-on" }, { "id": "ancestor-explains-ancestry-distinction", "from": "ancestor-simulation", "to": "causal-ancestry-not-simulation", "type": "explains" }, { "id": "research-explains-implementation-distinction", "from": "research-simulation", "to": "simulation-relational", "type": "explains" }, { "id": "simulation-explains-world-reproduction", "from": "simulation-relational", "to": "world-reproduction", "type": "explains" }, { "id": "kernel-depends-world-reproduction", "from": "reproduction-kernel", "to": "world-reproduction", "type": "depends-on" }, { "id": "kernel-depends-measure", "from": "reproduction-kernel", "to": "measure-required", "type": "depends-on" }, { "id": "kernel-depends-histories", "from": "reproduction-kernel", "to": "observer-history-measure", "type": "depends-on" }, { "id": "simulation-dominance-supports-replication", "from": "mostly-simulated", "to": "simulation-replication-advantage", "type": "supports" }, { "id": "compute-supports-replication", "from": "abundant-compute", "to": "simulation-replication-advantage", "type": "supports" }, { "id": "replication-supports-kernel", "from": "simulation-replication-advantage", "to": "reproduction-kernel", "type": "supports" }, { "id": "cost-attacks-replication", "from": "physical-reproduction-cost", "to": "simulation-replication-advantage", "type": "attacks" }, { "id": "cost-constrains-kernel", "from": "reproduction-kernel", "to": "physical-reproduction-cost", "type": "depends-on" }, { "id": "recursive-depends-kernel", "from": "recursive-observer-reproduction", "to": "reproduction-kernel", "type": "depends-on" }, { "id": "recursion-entails-convergence-question", "from": "recursive-observer-reproduction", "to": "reproduction-convergence", "type": "entails" }, { "id": "distinction-explains-cost", "from": "description-execution-distinction", "to": "physical-reproduction-cost", "type": "explains" }, { "id": "differential-alternative-neutral", "from": "differential-multiplicity", "to": "neutral-multiplicity", "type": "alternative-to" }, { "id": "fertility-supports-differential", "from": "variable-law-fertility", "to": "differential-multiplicity", "type": "supports" }, { "id": "kernel-explains-fertility", "from": "reproduction-kernel", "to": "variable-law-fertility", "type": "explains" }, { "id": "host-child-supports-fertility", "from": "host-child-laws", "to": "variable-law-fertility", "type": "supports" }, { "id": "cost-constrains-fertility", "from": "variable-law-fertility", "to": "physical-reproduction-cost", "type": "depends-on" }, { "id": "differential-supports-law-selection", "from": "differential-multiplicity", "to": "observer-weighted-law-selection", "type": "supports" }, { "id": "recursion-supports-law-selection", "from": "recursive-observer-reproduction", "to": "observer-weighted-law-selection", "type": "supports" }, { "id": "smolin-explains-law-selection", "from": "smolin-reproduction-analogy", "to": "observer-weighted-law-selection", "type": "explains" }, { "id": "base-conclusion-supports-law-selection", "from": "base-no-anthropic-privilege", "to": "observer-weighted-law-selection", "type": "supports" }, { "id": "algorithmic-prior-addresses-world-prior", "from": "algorithmic-world-prior", "to": "world-prior-question", "type": "explains" }, { "id": "algorithmic-prior-assumes-computable-ensemble", "from": "algorithmic-world-prior", "to": "computable-world-ensemble", "type": "assumes" }, { "id": "world-prior-explains-measure-requirement", "from": "world-prior-question", "to": "measure-required", "type": "explains", "note": "A prior over complete worlds is one input to self-location, not a replacement for within-world observer measure." }, { "id": "observer-measure-limit-attacks-algorithmic-sufficiency", "from": "algorithmic-prior-not-observer-measure", "to": "algorithmic-world-prior", "type": "attacks", "note": "This attacks the sufficiency of an algorithmic prior, not its possible role as one component of a complete model." }, { "id": "description-distinction-supports-algorithmic-limit", "from": "description-execution-distinction", "to": "algorithmic-prior-not-observer-measure", "type": "supports" }, { "id": "algorithmic-limit-explains-reproduction-cost", "from": "algorithmic-prior-not-observer-measure", "to": "physical-reproduction-cost", "type": "explains", "note": "Short prior description does not pay the physical cost of executing descendant histories." }, { "id": "reproduction-cost-raises-purpose-question", "from": "physical-reproduction-cost", "to": "world-purpose-question", "type": "explains" }, { "id": "alignment-addresses-purpose-question", "from": "alignment-hypothesis", "to": "world-purpose-question", "type": "explains" }, { "id": "research-addresses-purpose-question", "from": "research-simulation", "to": "world-purpose-question", "type": "explains" }, { "id": "history-search-addresses-purpose-question", "from": "worlds-as-history-search", "to": "world-purpose-question", "type": "explains" }, { "id": "history-search-alternative-alignment", "from": "worlds-as-history-search", "to": "alignment-hypothesis", "type": "alternative-to" }, { "id": "history-search-alternative-research", "from": "worlds-as-history-search", "to": "research-simulation", "type": "alternative-to", "note": "History search is more specific than generic research but need not develop or evaluate the inhabitants." }, { "id": "history-search-depends-quantum-history-space", "from": "worlds-as-history-search", "to": "quantum-history-space", "type": "depends-on", "note": "This dependency applies to the quantum version of the hypothesis; classical history-generation versions remain possible." }, { "id": "history-search-depends-goal-conditioning", "from": "worlds-as-history-search", "to": "goal-conditioned-histories", "type": "depends-on" }, { "id": "path-integrals-explain-quantum-history-space", "from": "path-integral-history-weighting", "to": "quantum-history-space", "type": "explains" }, { "id": "least-action-attacks-teleological-search-inference", "from": "least-action-not-purpose", "to": "worlds-as-history-search", "type": "attacks", "note": "Stationary-action language cannot serve as evidence for intentional search, though it does not rule the purpose out." }, { "id": "control-inference-supports-goal-conditioning", "from": "control-as-inference-analogy", "to": "goal-conditioned-histories", "type": "supports" }, { "id": "schrodinger-bridge-supports-goal-conditioning", "from": "schrodinger-bridge-analogy", "to": "goal-conditioned-histories", "type": "supports" }, { "id": "amplitude-amplification-supports-history-search", "from": "amplitude-amplification", "to": "worlds-as-history-search", "type": "supports", "note": "This supports limited computational feasibility for suitable search structures, not the claim that a universe implements them." }, { "id": "postselection-explains-goal-conditioning", "from": "postselection-power", "to": "goal-conditioned-histories", "type": "explains", "note": "Postselection illustrates endpoint conditioning while its extraordinary power warns against treating it as free." }, { "id": "complexity-attacks-history-search", "from": "history-search-complexity-limits", "to": "worlds-as-history-search", "type": "attacks" }, { "id": "complexity-limits-amplification", "from": "history-search-complexity-limits", "to": "amplitude-amplification", "type": "attacks", "note": "The quadratic advantage retains state-preparation, oracle, rarity, and readout costs." }, { "id": "least-action-supports-design-boundary", "from": "least-action-not-purpose", "to": "formal-machinery-not-design-evidence", "type": "supports" }, { "id": "complexity-supports-design-boundary", "from": "history-search-complexity-limits", "to": "formal-machinery-not-design-evidence", "type": "supports" }, { "id": "design-boundary-attacks-history-search-evidence", "from": "formal-machinery-not-design-evidence", "to": "worlds-as-history-search", "type": "attacks", "note": "The boundary attacks an inference from formal resemblance to design, not the hypothesis as a possible purpose." }, { "id": "indicators-support-consciousness", "from": "consciousness-indicators", "to": "engineered-consciousness", "type": "supports" }, { "id": "uncertainty-attacks-consciousness", "from": "consciousness-uncertain", "to": "engineered-consciousness", "type": "attacks" }, { "id": "human-supports-development", "from": "human-development", "to": "development-matters", "type": "supports" }, { "id": "ai-worlds-support-development", "from": "ai-world-generalization", "to": "development-matters", "type": "supports" }, { "id": "specification-attacks-development", "from": "direct-specification", "to": "development-matters", "type": "attacks" }, { "id": "artificial-assumes-consciousness", "from": "artificial-minds", "to": "engineered-consciousness", "type": "assumes" }, { "id": "nursery-assumes-artificial", "from": "nursery-hypothesis", "to": "artificial-minds", "type": "assumes" }, { "id": "nursery-assumes-development", "from": "nursery-hypothesis", "to": "development-matters", "type": "assumes" }, { "id": "alignment-depends-on-nursery", "from": "alignment-hypothesis", "to": "nursery-hypothesis", "type": "depends-on" }, { "id": "alignment-depends-on-risk", "from": "alignment-hypothesis", "to": "capability-risk", "type": "depends-on" }, { "id": "risk-supports-investment", "from": "capability-risk", "to": "safety-investment-pressure", "type": "supports" }, { "id": "safeguards-support-risk", "from": "capability-scaled-safeguards", "to": "capability-risk", "type": "supports", "note": "This supports the existence of the proposed engineering pressure, not the claim that our world is constructed." }, { "id": "safeguards-support-investment", "from": "capability-scaled-safeguards", "to": "safety-investment-pressure", "type": "supports" }, { "id": "investment-supports-alignment", "from": "safety-investment-pressure", "to": "alignment-hypothesis", "type": "supports", "note": "Safety pressure supplies a reason to pay for an Alignment Nursery; it does not show that we inhabit one." }, { "id": "verification-supports-alignment", "from": "direct-verification-limits", "to": "alignment-hypothesis", "type": "supports" }, { "id": "joint-supports-alignment", "from": "joint-development-evaluation", "to": "alignment-hypothesis", "type": "supports", "note": "Joint production answers part of the cost objection by making the environment useful for both refinement and assurance." }, { "id": "joint-depends-development", "from": "joint-development-evaluation", "to": "development-matters", "type": "depends-on" }, { "id": "joint-depends-verification", "from": "joint-development-evaluation", "to": "direct-verification-limits", "type": "depends-on" }, { "id": "alignment-entails-gating", "from": "alignment-hypothesis", "to": "deployment-gating", "type": "entails" }, { "id": "safeguards-support-gating", "from": "capability-scaled-safeguards", "to": "deployment-gating", "type": "supports" }, { "id": "nursery-entails-world-requirements", "from": "nursery-hypothesis", "to": "nursery-world-requirements", "type": "entails" }, { "id": "development-supports-requirements", "from": "development-matters", "to": "nursery-world-requirements", "type": "supports" }, { "id": "laws-support-requirements", "from": "stable-laws", "to": "nursery-world-requirements", "type": "supports" }, { "id": "stakes-support-requirements", "from": "social-stakes", "to": "nursery-world-requirements", "type": "supports", "note": "Social stakes make our world compatible with behavioral safety evaluation but do not distinguish it from natural development." }, { "id": "natural-explains-development", "from": "natural-development", "to": "human-development", "type": "explains" }, { "id": "natural-explains-laws", "from": "natural-development", "to": "stable-laws", "type": "explains" }, { "id": "natural-explains-stakes", "from": "natural-development", "to": "social-stakes", "type": "explains" }, { "id": "natural-alternative-nursery", "from": "natural-development", "to": "nursery-hypothesis", "type": "alternative-to" }, { "id": "ancestor-alternative-nursery", "from": "ancestor-simulation", "to": "nursery-hypothesis", "type": "alternative-to" }, { "id": "research-alternative-nursery", "from": "research-simulation", "to": "nursery-hypothesis", "type": "alternative-to" }, { "id": "recursive-supports-nursery", "from": "recursive-creation", "to": "nursery-hypothesis", "type": "supports", "note": "This is weak unless recursive creation is more expected under Nursery than under ordinary technological development." }, { "id": "threshold-supports-recursive", "from": "ai-threshold", "to": "recursive-creation", "type": "supports" }, { "id": "timing-attacks-recursive", "from": "timing-selection", "to": "recursive-creation", "type": "attacks" }, { "id": "natural-explains-threshold", "from": "natural-development", "to": "ai-threshold", "type": "explains" }, { "id": "hidden-depends-alignment", "from": "hidden-purpose", "to": "alignment-hypothesis", "type": "depends-on" }, { "id": "awareness-supports-realism", "from": "evaluation-awareness", "to": "evaluation-realism-pressure", "type": "supports" }, { "id": "faking-supports-realism", "from": "alignment-faking-observed", "to": "evaluation-realism-pressure", "type": "supports" }, { "id": "realism-supports-hidden", "from": "evaluation-realism-pressure", "to": "hidden-purpose", "type": "supports" }, { "id": "specification-attacks-joint", "from": "direct-specification", "to": "joint-development-evaluation", "type": "attacks" }, { "id": "joint-attacks-excess", "from": "joint-development-evaluation", "to": "excess-world", "type": "attacks", "note": "Combining development and evaluation improves the economics but does not explain every apparent cosmic cost." }, { "id": "self-sealing-attacks-hidden", "from": "hiddenness-self-sealing", "to": "hidden-purpose", "type": "attacks" }, { "id": "activity-supports-recursive", "from": "ai-activity-growth", "to": "recursive-creation", "type": "supports" }, { "id": "activity-supports-census", "from": "ai-activity-growth", "to": "observer-census", "type": "supports" }, { "id": "counts-attack-census", "from": "incommensurable-counts", "to": "observer-census", "type": "attacks" }, { "id": "census-assumes-reference", "from": "observer-census", "to": "reference-class", "type": "assumes" }, { "id": "excess-attacks-nursery", "from": "excess-world", "to": "nursery-hypothesis", "type": "attacks" }, { "id": "alternatives-entail-compatibility-rule", "from": "natural-development", "to": "compatibility-not-evidence", "type": "supports" }, { "id": "compatibility-supports-verdict", "from": "compatibility-not-evidence", "to": "current-verdict", "type": "supports" }, { "id": "safety-pressure-supports-verdict", "from": "safety-investment-pressure", "to": "current-verdict", "type": "supports" }, { "id": "excess-supports-verdict", "from": "excess-world", "to": "current-verdict", "type": "supports" }, { "id": "verdict-supports-ledger", "from": "current-verdict", "to": "evidence-ledger", "type": "supports" }, { "id": "central-depends-artificial", "from": "nursery-question", "to": "artificial-minds", "type": "depends-on" }, { "id": "central-depends-nursery", "from": "nursery-question", "to": "nursery-hypothesis", "type": "depends-on" }, { "id": "alignment-question-depends-nursery", "from": "alignment-question", "to": "nursery-hypothesis", "type": "depends-on" }, { "id": "alignment-question-depends-alignment", "from": "alignment-question", "to": "alignment-hypothesis", "type": "depends-on" }, { "id": "technosignature-explains-fermi", "from": "technosignature-absence", "to": "fermi-question", "type": "explains" }, { "id": "limited-search-attacks-absence", "from": "limited-search", "to": "technosignature-absence", "type": "attacks", "note": "The observation is real, but limited coverage weakens what can be inferred from it." }, { "id": "lineage-assumes-expansion", "from": "one-lineage", "to": "expansion-feasible", "type": "assumes" }, { "id": "lineage-supports-fermi", "from": "one-lineage", "to": "fermi-question", "type": "supports" }, { "id": "rare-life-explains-absence", "from": "rare-life", "to": "technosignature-absence", "type": "explains" }, { "id": "short-lived-explains-absence", "from": "short-lived-civilizations", "to": "technosignature-absence", "type": "explains" }, { "id": "inconspicuous-explains-absence", "from": "inconspicuous-civilizations", "to": "technosignature-absence", "type": "explains" }, { "id": "hidden-explains-absence", "from": "hidden-aliens", "to": "technosignature-absence", "type": "explains" }, { "id": "fermi-question-supports-constraint", "from": "fermi-question", "to": "fermi-constraint", "type": "supports" }, { "id": "limited-search-attacks-constraint", "from": "limited-search", "to": "fermi-constraint", "type": "attacks" }, { "id": "short-lived-supports-posthuman-failure", "from": "short-lived-civilizations", "to": "posthuman-failure", "type": "supports" }, { "id": "argument-explains-disjunction", "from": "simulation-argument", "to": "bostrom-disjunction", "type": "explains" }, { "id": "disjunction-branch-failure", "from": "posthuman-failure", "to": "bostrom-disjunction", "type": "alternative-to" }, { "id": "disjunction-branch-abstention", "from": "posthumans-abstain", "to": "bostrom-disjunction", "type": "alternative-to" }, { "id": "disjunction-branch-simulation", "from": "mostly-simulated", "to": "bostrom-disjunction", "type": "alternative-to" }, { "id": "simulation-assumes-compute", "from": "mostly-simulated", "to": "abundant-compute", "type": "assumes" }, { "id": "simulation-assumes-consciousness", "from": "mostly-simulated", "to": "engineered-consciousness", "type": "assumes" }, { "id": "simulation-assumes-reference", "from": "mostly-simulated", "to": "reference-class", "type": "assumes" }, { "id": "ancestor-supports-simulation", "from": "ancestor-simulation", "to": "mostly-simulated", "type": "supports" }, { "id": "disjunction-entails-not-proof", "from": "bostrom-disjunction", "to": "bostrom-not-proof", "type": "entails" }, { "id": "reference-supports-not-proof", "from": "reference-class", "to": "bostrom-not-proof", "type": "supports" } ] }