papers

the infrastructure.

research in progress on disclosure, judgment, and coordination.

this is my workbench. i bootstrap it outside any university, and right now it is moving fast, each paper an idea pushed until it either held or broke. when i find something i give it a working name, often my own, so that if it breaks, the failure has a name on it, and if it holds, the record shows when it was struck. when a paper goes to a journal, it takes standard terms. where an idea has broken, the entry says so, and the broken version stays on zenodo under its doi.

three places to start

recent papers

2026bg. the judgment record conjecture: coordination-dependence and the value of attributed price records for machine learning. https://doi.org/10.5281/zenodo.23325026 read online

current AI models are trained on raw text and public market data, like stock prices. a price record, though, compresses human behavior into a single scalar. it throws away who made the judgment, when they made it, how they reasoned, and their historical accuracy track record. in this paper i let b in [0, 1] be the coordination-dependence fraction of an outcome, the share of its variance that coordination among the parties around it produces. i conjecture, in what i call the judgment record conjecture, that the gain in model decision quality from training on attributed, timestamped, and later-scored judgment records rather than on the blind public record vanishes as b approaches zero in deep markets, and rises both with b and with the depth of the scoring record. i set out a companion conjecture, price-AI coupling, for the loop that opens once models set prices as well as learn from them: under blind prices, price informativeness and model quality thin together as the machine share of price formation rises, while attributed records let the human-originated signal be isolated.

2026bf. singular models and moving varieties: the algebraic geometry of coordination-dependent data. https://doi.org/10.5281/zenodo.21832685 read online

current machine learning models rely on statistical assumptions that treat data structures as fixed geometric shapes. in human-dominated domains like markets and governance, the act of observing and predicting data causes the underlying data-generating structure to shift, a property i quantify with the coordination-dependence fraction b. in this paper i connect singular learning theory, algebraic statistics, and performative prediction through b. first, i argue that conservation laws in training data can induce fisher rank deficiency along probed directions (a note: this inherits the reversal in 2026as, where the probed direction is stiff rather than flat; AG-1 is affected), and that training on attributed, timestamped, six-facet disclosures shapes the singularity structure and local learning coefficient of a model measurably differently from unstructured text about the same domain. second, i replace the fixed-variety assumption of algebraic statistics with a family of moving varieties fibered over a coordination-state manifold whose primary structure is the closed elliptope of similarity matrices in its bures-wasserstein geometry, which gives the fraction b a candidate geometric referent.

2026be. the machine-origination boundary and the differential filtration audit: a criterion and protocol for certifying origination. https://doi.org/10.5281/zenodo.21460914 read online

as frontier AI models become more opaque, it is nearly impossible to tell from the outside whether a model has produced an original insight or is recombining its training corpus. in this paper i set out an operational mathematical criterion and a synthesis protocol for certifying machine origination without access to model weights, training data, or internal states. i let Wt be the ambient public filtration, Y a claim struck at time t, and Z its resolution at t + delta. an opaque system with Wt-measurable inputs achieves origination relative to Wt if and only if the conditional mutual information is strictly positive, I(Y; Z | Wt) > 0. the estimator i propose is a proper-scoring differential against a synthesized Wt-measurable baseline with a pre-registered degeneracy tripwire, so that every verdict is a computation on public and declared objects.

2026bd. formal extensions to the ocot (origin chain of thought) corpus: language, growth, and the limits of physics contact. https://doi.org/10.5281/zenodo.21198571 read online

in this consolidated volume i extend the formal foundations of origin chain of thought (ocot) across six papers:

  • the language mediation channel: i characterize how physical constraint geometry and conservation laws attenuate when converted into generic prose, and i argue that six-facet structured disclosure preserves fisher rank-deficiency signatures with strictly higher geometric fidelity than unstructured text, amplified by the scaling facet (a note: LM-1 to LM-3 and the first hypothesis of the test protocol inherit the reversal in 2026as and change sign).

  • the dynamic calibration improvement model: i supply the formal learning dynamics equation and argue that six-facet disclosure under proper scoring acts as a deliberate-practice engine, driving individual calibration monotonically toward capacity and shifting the population quality distribution rightward at a rate that compounds with the age of the commons.

  • the scaling facet as a distinct information type: i argue that because aggregate constraint states combine nonlinearly, the scaling facet carries system-scale interaction information that cannot be derived from individual-transaction outputs, which completes the argument that the six facets are each necessary and jointly sufficient.

  • joint dynamics of coupled constraint states: i model intersecting domains (such as aquifer depletion and grid stability) through block-structured covariances, argue that treating adjacent domains as fixed boundaries loses cross-domain mutual information exactly when threshold sensitivity diverges, and derive the optimal sequence for connecting domains to the disclosure network.

  • information architecture and total factor productivity: i endogenize coordination quality inside the solow productivity residual and argue that because tracked constraint interactions grow quadratically in network size, combinatorial disclosure reduces capital misallocation with increasing returns to aggregate output.

  • non-adversarial failure modes: i analyze four endogenous failure modes that occur without an attacker, ontology drift, scoring lag, platform fragmentation, and quality collapse, and i state for each the mechanism, dominance conditions, early-warning signature, and structural defense.

2026bc. the six-facet tensor (the current specification of the beier tensor): consolidated foundations of the beier record. https://doi.org/10.5281/zenodo.21198473 read online

in this seven-paper consolidated volume i bring the formal specification of the beier tensor to its canonical six-facet form, defining the extended price E = (p, T, w), which keeps the transaction price p as its nucleus and adds structure around it rather than replacing it, and building the metadata layer around it:

  • the six-facet beier tensor: i upgrade the five-field record of 2026z to a nucleus wrapped in six linearly independent facets (identity, input, method, output, scaling, and transmit), recover the earlier five-field treatment as the metadata-suppressed, transmit-constant limit, and generalize blackwell dominance to the six-facet extended price.

  • strike-time and transmit-time: i argue that no single-timestamp record can preserve strict propriety of scoring while permitting an author-chosen publication delay. separating tamper-evident strike-time (when the cognitive act is committed) from transmit-time (when it becomes readable) resolves the trade-off between competitive secrecy and honest scoring, and it makes the disclosure lag an informative conditioning variable.

  • the transmit facet: i argue that the scope test for choosing who may read a record is equivalent to a coordination-completeness condition, and that a mandatory public option enables permissionless redundancy, without which private coalitions can enclose the information commons.

  • typed nulls: i replace binary non-disclosure with a four-symbol alphabet of absence (inapplicable, unknown to author, withheld under disposition, or carried by standing default positions), and i argue that the kind of blank is a strictly richer channel policed by proper scoring.

  • the substrate marker: i record whether reasoning originated from a person, a model, or a hybrid across facets, so that uncalibrated, zero-cost synthetic outputs cannot borrow human track-record presumptions and drive an adverse-selection lemons collapse.

  • the engagement level: i record the depth of cognitive work from a single glance to a multi-month study, and i argue that engagement is orthogonal to confidence and historical calibration, most informative in novel long-horizon domains, and must remain a reader-side conditioning signal outside the automated scoring rule.

  • the composition algebra: i formalize labeled citations (builds on, endorses, refutes, references) as a directed acyclic graph on which convergence, divergence, and constraint-space gaps become decidable predicates. i argue that confirmation value between agreeing records requires disjoint ancestor sets, and i set out inheritance rules that close privacy-laundering routes.

2026bb. extensions of the extropy measure: technology, restoration, institutions, distribution, history, and perception. https://doi.org/10.5281/zenodo.21198383 read online

in this seven-paper volume i extend beier extropy across six socio-technical dimensions: technology scaling, environmental and structural restoration, institutional longevity, wealth distribution, historical path-dependency, and human perception. through the extropy-distortion bridge i argue that the displacement of multi-dimensional constraint variance by scalar price signals operates as an entropy-generating channel (on the thermodynamic framing, see the note at 2026i), and that institutional decay, resource depletion, and unpriced externalities can follow from running complex systems on low-bandwidth scalar indicators.

2026ba. the hosting conditions: political topology, sustainability, and coalition incentives. https://doi.org/10.5281/zenodo.21198335 read online

in this paper i argue that an origin chain of thought (ocot) judgment layer is not politically neutral infrastructure. it imposes five mandatory conditions on its host environment: permissionless publication, a public-scope option, persistent identity that cannot be faked, tamper-evident records, and tolerance for dissenting structural assessments. i define a regime's participation capacity across democratic, authoritarian, and totalitarian topologies, and i predict that partial authoritarian hosting is unstable because the economic value of an open coordination channel makes suppressing adjacent dissent increasingly costly.

2026az. the ordering channel: strike-time, permutation space, and what the beier factorial names. https://doi.org/10.5281/zenodo.21198137 read online

traditional price series treat market observations as interchangeable, unordered sets. in this paper i define the ordering channel and distinguish three growth rates in a disclosure network: a quadratic floor on pairwise interaction information among n disclosers, an exponential space of subsets, and a factorial space of n! distinguishable chronological histories, unlocked when tamper-evident strike-time converts sets of disclosures into permutations. scalar prices carry this ordering channel at zero. the beier factorial, as i set it out here, formalizes the permutation geometry through which models can learn the chronological sequence of multi-party coordination. a correction note is in preparation: the quadratic floor on interaction information that this paper cites as proven does not hold as stated (see 2026s); the paper's own three results do not depend on it.

2026ay. a missing coordinate in the classification of quantitative data: the coordination-dependence fraction. https://doi.org/10.5281/zenodo.21154565 read online

standard statistical practice assumes that the data-generating process of a measured quantity remains invariant under observation. in this paper i argue that this assumption fails across microeconomic school capitalization, equity indices, and large language model benchmarks, and i introduce the coordination-dependence fraction b to separate observation-independent physical variance from variance constituted by coordinating observers. i supply an empirical estimator b-hat built on meta-regression residuals, link the secular rise in equity intangibles to rising aggregate b, and argue that producers of quantitative data owe downstream users a report of b alongside the series.

2026ax. productive failure and goodhart dislocation: how systematic error maps the topology of ignorance and predicts where the next crisis originates. https://doi.org/10.5281/zenodo.20012184 read online

standard forecasting treats good proper scores as success and poor scores as noise. in this paper i argue that systematically poor scores on specific constraint dimensions are more informative than good ones, because they map the orthogonal complement of a forecaster's observation subspace, which i call productive failure. aggregating these systematic errors across a population maps the observation architecture's topology of ignorance. i then derive the goodhart dislocation, the quantifiable, discontinuous repricing crash that occurs when a neglected constraint dimension becomes visible after prolonged scalar metric targeting, and i argue that early disclosure bias signatures predict both the location and magnitude of the crash before it occurs.

2026aw. basin transition economics: how constraint dimensions cross thresholds, why the price cannot see them coming, and what the disclosure mechanism changes. https://doi.org/10.5281/zenodo.20012093 read online

in this paper i distinguish the observation-architecture phase transition (where stability S crosses 1) from a basin transition, which i formalize as a saddle-node bifurcation in which a hidden constraint parameter theta governs whether an asset's value basin persists. i argue three results. the first is that basin transitions occur with zero warning time under scalar prices because prices track smooth movement inside the basin rather than the boundary parameter theta. the second is that basin transitions exhibit hysteresis, where the recovery threshold exceeds the collapse threshold in proportion to irreversibility. the third is that the two discontinuities are bidirectionally coupled. structured disclosure breaks the coupling by making theta observable, converting zero-warning collapses into forecastable trajectories. a correction note is in preparation: in the paper's own normal form the price does carry a faint, late and unreliable warning, so the zero-warning claim will be restated.

2026av. reverse percolation on observation lattices: how disclosure fragments the blind zone and why the cascade threshold is a phase transition. https://doi.org/10.5281/zenodo.20011994 read online

in this paper i map the illumination of unpriced constraint interactions to reverse site percolation on the line graph of the complete constraint graph. initially the blind zone forms a giant connected component spanning the interaction lattice. as multi-dimensional disclosures arrive and remove vertices from the blind zone, i argue that the lattice undergoes a sharp fragmentation phase transition at a computable critical disclosure density (corresponding to the $5.8 trillion cascade threshold of 2026q), above which the giant blind component vanishes into isolated pockets and each disclosure yields strictly increasing marginal reductions in remaining blind connectivity. a correction note is in preparation: the blind zone shrinks smoothly rather than through a phase transition; the increasing marginal returns result stands.

2026au. topological protection of blind spots under scalar observation: a grassmannian framework for information architecture fragility. https://doi.org/10.5281/zenodo.20011788 read online

in this paper i explain why excess distortion persists regardless of how many scalar analysts or esg ratings a market adds: blind spots under scalar observation are topologically protected. scalar channels operate on the grassmannian Gr(1, K), leaving the constraint interactions in Gr(2, K) that represent nontrivial cohomology and schubert classes completely invisible for any economy with K >= 3 constraint dimensions. i argue that the multi-dimensional scaling facet lifts the observation architecture to Gr(k, K) with k >= 2, discontinuously collapsing those schubert classes and changing what the market can see. a correction note is in preparation: the containment theorem stands, but the conclusion that scalar observers cannot see interactions does not follow from it; what protects the blind spot is correlated methods, not one-number outputs.

core papers

2026at. symbolic time crystals: temporal invariants in human-generated data streams and their detection through weight-space geometry. https://doi.org/10.5281/zenodo.19440255 read online

in this paper i identify recurring temporal invariants, meaning periodic and quasi-periodic coordination structures in human-generated judgment streams that persist across regime shifts, and i argue that these temporal symmetries appear as detectable geometric invariants in the weight space and fisher geometry of models trained on timestamped disclosure corpora. a note: the weight-space candidates rest on CG-1 to CG-3 of 2026as, which are corrected, and the windowing protocol is a band-stop test requiring a sweep over window lengths (see 2026ac); the detection claim is a conjecture.

2026as. the geometry of constraint knowledge: how physical laws impose structure on learned representations, and what this means for coordination. https://doi.org/10.5281/zenodo.19439253 read online

in this paper i connect physical constraint geometry (2026ai) to neural network weight manifolds through the fisher information metric. i argue three propositions: that conservation laws in the data-generating process induce targeted rank deficiencies along probed conservation-violating directions in the fisher matrix (CG-1); that power-law scaling exponents beta are encoded in the sectional curvature of the weight manifold along the system-size direction, with sgn(K_N) = sgn(beta - 1) (CG-2); and that transformer attention weights decay spatially with the same characteristic length as the physical propagation kernel (CG-3). i also specify a five-tool software suite and a four-phase empirical research program. a correction note is in preparation: CG-1 is reversed; in the corrected form (CG-1') the probed direction is stiff, scaling as 1/s^2, not flat. CG-2 is withdrawn and CG-3 is a conjecture.

2026ar. the economics of judgment: why human assessment is an unlimited resource the price system throws away. https://doi.org/10.5281/zenodo.19439637 read online

in this paper i argue that human judgment is the most abundant and most wasted resource in the global economy. it is present in every transaction, and scalar market clearing discards it through cross-agent aggregation, temporal compression, and format incompatibility. i develop a structural parallel to kary mullis's polymerase chain reaction (pcr). pcr amplified pre-existing dna to detectable concentrations without creating new genes, and in the same way the disclosure incentive preserves, structures, scores, and amplifies the judgment that agents already produce, while engineering high-validity intermediate milestones through the scaling field so that capturing the resource actively improves its calibration quality.

2026aq. the institutional substrate for knowledge coordination: what must exist before disclosure can work. https://doi.org/10.5281/zenodo.19439299 read online

in this paper i identify and analyze the four public-good components of epistemic infrastructure that a judgment market needs in order to operate: outcome observation systems, scoring computation platforms, persistent tamper-resistant track record registries, and shared constraint domain ontologies. each component is non-rival, driven by network effects, and valuable only when the others exist, so private markets undersupply them. drawing on frischmann and ostrom, i propose a three-stage sequencing strategy and a hybrid public-consortium-open-source funding model to break the chicken-and-egg interdependence.

2026ap. design principles for disclosure systems: what the mechanism design literature tells us about building judgment markets. https://doi.org/10.5281/zenodo.19439488 read online

in this paper i map the six-dimensional mechanism design space for judgment market architectures: scoring rule choice (bounded brier vs. unbounded logarithmic vs. peer prediction), commitment structure, track record aggregation, disclosure format, verification architecture, and market structure. i analyze the cross-cutting tradeoffs between truth-telling strength, participation cost, information richness, and adversarial robustness, and i derive four invariant design principles: outcome-based scoring, public portable track records, a structured format with a scaling component, and an open-source specification.

2026ao. the performative price reset: how disclosure requirements create reasoning that was not previously occurring. https://doi.org/10.5281/zenodo.19439805 read online

in this paper i extend austin's and callon's performativity frameworks to information architecture. when the disclosure format requires an analyst to complete the scaling field, assessing how a transaction interacts with shared constraints at aggregate scale, it does not merely uncover a pre-existing hidden assessment. it brings into existence analytical reasoning that was not previously occurring at all. because the format generates new constraint cognition rather than merely capturing an existing stock, i argue that its economic yield exceeds the bounds of static disclosure models.

2026an. adversarial disclosure economics: attacks on the judgment data stream and the cross-disciplinary defense that unification provides. https://doi.org/10.5281/zenodo.19438664 read online

in this paper i classify four attack vectors against a valuable judgment data stream: sybil identity fabrication, coordinated scoring collusion, AI-generated hollow disclosures (including variance-based persona farming), and adversarial domain selection. i argue that no single discipline can defend against all four. economics detects market and domain anomalies, computer science detects statistical and kolmogorov-complexity fingerprints, mathematics detects violations of honest calibration bounds, and information theory detects compression-ratio mismatches on the scaling field. together they form an overlapping defense with no shared blind spot.

2026am. when constraint domains interact: the economics of cross-domain information failure. https://doi.org/10.5281/zenodo.19439157 read online

in this paper i extend within-domain scaling analysis across domain boundaries, with examples such as the ogallala aquifer-atmosphere-grid nexus, climate-infrastructure-finance cascades, and land-biodiversity-agriculture feedback. i argue that single-domain assessments compress coupled domains into static boundary conditions, losing cross-domain mutual information I(theta_A; theta_B) > 0 precisely when systems approach critical thresholds, which i call threshold coupling amplification. i then specify how cross-domain reading, boundary-sensitivity fields, and coupling maps recover the missing interaction structure.

2026al. how the improvement mechanism scales across populations: the economics of collective calibration improvement. https://doi.org/10.5281/zenodo.19439747 read online

in this paper i extend arrow's learning-by-doing to public learning-by-disclosing. i argue that when individual calibration improvements enter the public commons as scored disclosures, collective calibration improves faster than the average individual rate, through three network channels: higher-quality analytical inputs for downstream readers, empirical diffusion of high-scoring methodologies, and rising calibration benchmarks that compound across generations of analysts.

2026ak. how disclosure propagates through networks: the economics of judgment transmission across institutional boundaries. https://doi.org/10.5281/zenodo.19439003 read online

in this paper i contrast information-destroying herding cascades, in the sense of bikhchandani, hirshleifer, and welch, where agents discard private signals to imitate actions, with information-preserving disclosure cascades, where agents evaluate their predecessors' five-field reasoning against track records and add their own analytical layer. i analyze reach, absorption, and transformation, along with the three sources of transmission loss across scientific, financial, regulatory, and operational boundaries: attention scarcity, trust friction, and boundary compression.

2026aj. the information content of non-disclosure: silence as signal in structured judgment markets. https://doi.org/10.5281/zenodo.19440045 read online

in this paper i analyze how unraveling logic (grossman-milgrom) and discretionary disclosure thresholds (verrecchia-dye) operate once agents commit to public disclosure domains and cadences. within a committed domain, i argue, silence or a missed cadence is no longer an invisible absence but an attributable, informative signal that principals and counterparties price against the agent's track record.

2026ai. the mathematical structure of interaction effects across scale: why constraint environments have geometric properties that discipline the scaling field. https://doi.org/10.5281/zenodo.19439918 read online

in this paper i formalize four physical constraint regularities that govern valid scaling assessments across resource, grid, and ecological domains: conservation bounds, power-law scaling exponents, exponential spatial propagation kernels with characteristic length l_network, and critical threshold bifurcations. i argue that physics imposes strict geometric consistency conditions which discipline the scaling field and expose fabricated assessments.

2026ah. scaling risk as the fifth consideration: why interaction effects at scale constitute a distinct information type the price system cannot carry. https://doi.org/10.5281/zenodo.19439990 read online

in this paper i set out the case that input, method, output, and identity capture the structure of an individual inference, while the interaction effects across agents and time that arise when an activity is replicated at scale, the fifth field, scaling, constitute a mathematically distinct information type that bilateral market clearing structurally compresses to zero.

2026ag. factorial growth through coordination: why the interaction of n disclosers produces combinatorial information, not linear. https://doi.org/10.5281/zenodo.19439420 read online

in this paper i extend weitzman's recombinant growth framework from combining ideas as objects to reading disclosures as interacting perspectives mediated by a shared constraint environment. when n analysts assess related facets of the same constraint, say a hydrologist, an irrigation engineer, and an agricultural economist on a shared watershed, the non-empty subsets of their disclosures generate conditional implications that no individual disclosure contains. i argue that this causes network information value to grow combinatorially rather than linearly and that it generates S-curve tipping dynamics in domain adoption. a note: the combinatorial growth here is a count of pairwise comparisons, N(N-1)/2, not a floor on information (see 2026s, 2026az).

2026af. reality debt: the accumulated cost of decisions made on compressed information. https://doi.org/10.5281/zenodo.19439896 read online

in this paper i define reality debt as the cumulative stock of unpriced physical, biological, and institutional constraint degradation that builds up when capital is allocated on scalar prices that compress away constraint dimensions. it stays invisible in conventional accounting until constraint thresholds bind and force discontinuous repricing.

2026ae. the solow residual as coordination quality: how information architecture determines total factor productivity. https://doi.org/10.5281/zenodo.19440121 read online

in this paper i decompose total factor productivity (the solow residual) and argue that aggregate output growth depends not only on technological innovation but on coordination quality, which i define as the precision with which an economy's information architecture routes capital toward activities aligned with real constraint states rather than scalar compression artifacts.

2026ad. the judgment kernel: three propositions on the cognitive architecture for wise machines. https://doi.org/10.5281/zenodo.19439686 read online

in this paper i define the judgment kernel as the minimal information architecture at the intersection of market incentive compatibility (JK-1), AI weight-space geometric fidelity (JK-2), and single-stream cross-surface coupling (JK-3). i argue that the five-field scored disclosure format (the record now has six facets, see the six-facet tensor above) is the canonical minimal architecture satisfying all three (proposition 106), that it is self-improving across all three surfaces above critical mass N* (proposition 107), and that it forms a bidirectional, fidelity-preserving, incentive-compatible bridge between computationally irreducible human judgment and scalable machine processing (proposition 108). a note: proposition 107 is not established, because it rests on TB-2 of 2026v (withdrawn) and NR-1 of 2026x (reversed through 2026as); JK-2 is reversed for the same reason.

2026ac. formal foundations for symbolic time crystals: three propositions on temporal invariance in learned representations. https://doi.org/10.5281/zenodo.19440156 read online

in this paper i supply the formal mathematical propositions that characterize how periodic and time-translation-symmetry-breaking coordination structures in human disclosure streams are preserved as stable geometric invariants inside neural network representations. a correction note is in preparation: STC-1 holds as a gaussian bound with a stronger constant; STC-2 is withdrawn and replaced by a band-stop test (STC-2') that in general requires a sweep over window lengths; STC-3(ii) is withdrawn.

2026ab. the performative proposition: how structured disclosure formats generate information that unstructured formats cannot elicit. https://doi.org/10.5281/zenodo.19439858 read online

in this paper i argue formally (propositions PF-1 and PF-2) that structured multi-field disclosure formats elicit strictly more constraint-state and aggregate-interaction information than unstructured narrative or scalar prices do, because the explicit field decomposition prompts analytical evaluation of scaling interactions that stay uncomputed under unstructured reporting.

2026aa. the economics of cognitive trade: three propositions on the human-AI exchange rate. https://doi.org/10.5281/zenodo.19439330 read online

in this paper i frame the relationship between human cognition and AI processing as ricardian trade governed by a task's computational irreducibility index CI(T) in [0, 1]. i argue three results: that a unique threshold CI* partitions comparative advantage between AI pattern-matching on reducible tasks (CI < CI*) and human ab initio judgment on irreducible tasks (CI > CI*) (proposition EX-1); that the human-AI exchange rate E = (Q_H / C_H) / (Q_AI / C_AI) is directly observable in proper-scoring-rule units per dollar (proposition EX-2); and that as AI capability grows, CI* shifts upward monotonically at a decelerating rate but never reaches 1, because irreducible simulation costs impose a permanent floor (proposition EX-3).

2026z. the beier tensor: a formal geometric definition of the deep price (early five-field specification; the current specification is the six-facet tensor, above). https://doi.org/10.5281/zenodo.19438824

in this paper i set out the differential-geometric formalization of a deep price as the ordered triple T = (p, D, w) in R x R^K x [0, 1], where the blinding price is the rank-zero degenerate case (p, 0, 0). i argue that the total mutual information content of the disclosure vector D is invariant under a change of analytical framework, from fund manager to regulator to farmer (proposition BT-1); that the track record weights define a positive semi-definite, framework-invariant metric tensor g(D_i, D_j) measuring weighted consensus (proposition BT-2); and that blackwell welfare dominance (BE-3) is the decision-theoretic expression of tensor coordinate invariance (proposition BT-3).

2026y. adversarial detection superadditivity and geometric distinguishability: two results on the cross-disciplinary defense architecture. https://doi.org/10.5281/zenodo.19436310

in this paper i state two results that back the cross-disciplinary defense of the judgment stream. in proposition AD-1 (superadditivity) i argue that under OR-gate aggregation across K independent detection channels with convex evasion costs, an attacker's cost to reduce the overall detection probability to d_target grows superlinearly in K, because each channel tightens the per-channel evasion requirement to d_target / K. in proposition GD-1 (geometric distinguishability) i argue that the fisher information distance d_geom along conservation-violating directions distinguishes genuine physical constraint corpora from fabricated corpora with probability approaching 1 at rate O(K / sqrt(n)). a correction note is in preparation: the stated AD-1 formula contradicts its own proof; the corrected claim is that evasion cost grows at least linearly in the number of independent channels, and GD-1 is restated as a conjecture.

2026x. conservation geometry under measurement noise: a quantitative convergence rate for weight-space rank deficiency. https://doi.org/10.5281/zenodo.19439089

in this paper i extend proposition CG-1 (2026as) from exact conservation to realistic training data with measurement or simulation noise of magnitude delta. i argue that fisher information along the conservation-violating direction converges quadratically as I_{v_cons}(w*) = L^2 * delta^2 / 3 + O(delta^3) (proposition NR-1); i derive the fisher ratio FR = (sigma / delta)^2 between noisy and conservation-respecting models, which yields FR ~ 10^{28} in simulation and FR ~ 100 in empirical data (proposition NR-2); and i argue that K independent conservation laws induce K asymptotically orthogonal near-zero eigenvalues of order O(delta_k^2) in the fisher matrix (proposition NR-3). a note: NR-1 to NR-3 extend CG-1 of 2026as and inherit its reversal; along the probed direction the fisher information is stiff, not near zero.

2026w. the inexhaustibility of forward-looking constraint information: why the judgment market’s independence from price efficiency is permanent. https://doi.org/10.5281/zenodo.19439615

in this paper i argue that the forward-looking constraint component theta_C > 0, on which the judgment market's independence from price efficiency (condition J4) rests, never shrinks to zero even if disclosure succeeds completely. i set out three independent mechanisms. physical and institutional non-stationarity continuously shifts the governing parameters phi(t) (proposition IC-1). pairwise interaction configurations among N agents grow combinatorially as 2^{N(N-1)/2}, which outstrips the bounded bit rate R_price of scalar prices (proposition IC-2). and nonlinear constraint systems with positive maximal lyapunov exponents lambda_max > 0 amplify initial uncertainty exponentially as e^{lambda_max * h}, so there is a permanent prediction horizon beyond which future trajectories are computationally irreducible (proposition IC-3).

2026v. a tight bound on the calibration threshold for honest disclosure: order statistics refinement of the strategic selection analysis. https://doi.org/10.5281/zenodo.19440275

in this paper i refine the strategic withholding bound in the coverage-adjusted proper scoring mechanism using order statistics, and i set out tight parameter bounds on the coverage weight alpha under which the calibration threshold q* vanishes and honest comprehensive disclosure strictly dominates cherry-picking. a correction note is in preparation: the order-statistic lemma, TB-1 and TB-2 are withdrawn. the corrected TB-2' holds at alpha < 1/2, and the honesty condition gets harder, not easier, over a longer record.

2026u. compositional improvement beyond the gaussian case: what survives when the cara-normal assumption is relaxed. https://doi.org/10.5281/zenodo.19438910

in this paper i argue that the exact equality of total scalar precision for uninformed agents between the standard and augmented grossman-stiglitz equilibria (2026b, proposition 2) is a gaussian coincidence, and that the compositional welfare benefit of structured disclosure is strictly stronger once the cara-normal assumptions are relaxed. under fat-tailed distributions with excess kurtosis kappa > 0, i argue that disclosure carries tail structure that gaussian-optimal scalar prices miss, which yields strictly higher uninformed welfare (proposition BE-1). under ambiguity-averse preferences (gilboa-schmeidler, epstein-schneider), i argue that disclosure resolves model uncertainty about the data-generating process (proposition BE-2). and by blackwell's theorem, i argue that (P, D) strictly dominates P across all decision problems whenever I(V; D | P) > 0 (proposition BE-3).

2026t. the extropy-distortion bridge: how beier extropy maps to rate-distortion improvement in the price system. https://doi.org/10.5281/zenodo.19439374

in this paper i close the formal loop between beier extropy E_B = H(D | P) in bits (2026a) and excess distortion delta = D_price - D* in variance units (2026c). i derive the exact distortion reduction Delta_D(E_B) = sum_{k=2}^K lambda_k (1 - 2^{-2 r_k}) when E_B disclosure bits are allocated across constraint dimensions (proposition EB-1). i argue that the optimal allocation of disclosure bits follows reverse water-filling on the uncompressed constraint variances, so that active residuals are equalized at the water level theta_D (proposition EB-2). i also derive the excess-distortion closure ratio phi(E_B) and its marginal closure rate 2 ln(2) theta_D / delta (proposition EB-3).

2026s. supplementary formal foundations: additional results on learning dynamics, information growth, and constraint coupling. https://doi.org/10.5281/zenodo.19440198

in this paper i supply the formal arguments underlying the companion corpus. these include learning acceleration and bounded S-curve convergence (propositions LA-1, LA-2), collective cross-learning rates (CL-1), and superadditive mutual ignition between human disclosure and AI processing above a critical mass N* (MI-1, MI-2, MI-3). a correction note is in preparation: the quadratic information floor (FI-1) does not hold as stated; the pairwise counting identity survives, and that value tracks the count is a conjecture. the 67x reality-debt figure (RD-1) is withdrawn; at the paper's own rates the ratio is between 1 and 3.9. the variance corollary of CL-1 is false, the monotonicity in SI-1 is not established, CT-1 and IA-1(iii) are flawed, and CE-1 holds only qualitatively.

2026r. the incentive compatibility of the disclosure incentive: a mechanism design foundation for the judgment market. https://doi.org/10.5281/zenodo.19378673

in this paper i set out the hurwicz-myerson mechanism design foundation of the judgment market. i argue single-period and sequential truth-telling under strictly proper scoring rules. i solve the witkowski-parkes strategic selection (cherry-picking) problem via pre-committed domain coverage scoring C_i = alpha R_{D_i} + (1 - alpha)(|A_i| / |D_i|). i argue that the calibration threshold vanishes universally for bounded rules at alpha < 1/2 and for the brier score at alpha < 4/5, and i micro-found the principal's willingness to pay for forward-looking constraint calibration theta_C. a correction note is in preparation: sequential truth-telling holds only under affine or additively separable payoffs, and tournament payoffs break it; there is no calibration threshold q*, and under a bounded proper rule with alpha < 1/2, or the brier score with alpha < 4/5, every agent discloses its full belief.

2026q. the scalar lobby: political economy of the transition from scalar to multi-dimensional observation. https://doi.org/10.5281/zenodo.19378651

in this paper i model the transition from single-number metrics to multi-dimensional observation as a heterogeneous global game (sakovics-steiner) across asset managers, compliance officers, and scalar index providers. i argue that adoption is sequential, with asset managers moving first, that optimal policy subsidizes early adopters rather than holdouts, and that grossman-stiglitz information rents on unpriced constraint dimensions penalize legacy holdouts. to avoid a partial-adoption stall and trigger a full cascade, the fragility gap has to cross a computable threshold. a correction note is in preparation: the subsidy-targeting result is contradicted by the paper's own solved model; the ordered-thresholds result stands under its assumed orderings.

2026p. smith’s missing field: technology scaling, information architecture, and the origin of externalities. https://doi.org/10.5281/zenodo.19378638

in this paper i trace the origin of market externalities to an architectural omission in classical price theory. adam smith's invisible hand operated in low-scaling environments where bilateral prices approximated social cost. modern technology scaling generates nonlinear cross-agent constraint interactions that scalar prices compress away unless they are paired with a fifth disclosure field, scaling.

2026o. the constitution axiom: observation as the origin of coordination-dependent reality. https://doi.org/10.5281/zenodo.19337210

in this paper i state the foundational ontological axiom of the prolegomena. for coordination-dependent quantities (beierspace), sustained multi-agent observation does not merely measure a pre-existing structure, it actively constitutes and maintains the reality of the structure being measured.

2026n. five axioms for coordination under compression. https://doi.org/10.5281/zenodo.19259393

in this paper i identify five independent axioms, observation-maintenance identity, bounded capacity, compression inevitability, irreducible heterogeneity, and reflexivity, and i combine them with shannon's rate-distortion theory to set out five foundational theorems: universal excess distortion, the beier factorial (a quadratic O(N^2) lower bound and an O(2^N) combinatorial ceiling on discoverable constraint interactions), universal dissipation, universal goodhart, and the reflexive phase transition at S = 1. a correction note is in preparation: the bistability at S = 1 does not exist in the stated model, and the directional-diversity bound in theorem 10.1 needs a stronger assumption than the one given.

2026m. the observation floor: is planck’s constant the universe’s water level? https://doi.org/10.5281/zenodo.19205572

in this paper i explore the structural isomorphism between reverse water-filling in shannon rate-distortion theory, where dimensions below the water level theta* receive zero observation bandwidth, and quantum measurement limits. i ask whether physical action quanta represent a universal observation floor under bounded channel capacity.

2026l. shaky scalars rock the world: why single-number systems cannot maintain the structures they measure. https://doi.org/10.5281/zenodo.19184778

in this paper i synthesize the rate-distortion and coordination-thermodynamics results (on the thermodynamic framing, see the note at 2026i) for general and policy audiences. i explain why compressing multi-dimensional institutions, ecosystems, and balance sheets into single-number targets (prices, gdp, credit ratings, esg scores) starves unmeasured constraint dimensions of maintenance bandwidth until they fail discontinuously.

2026k. the beierspace fraction: measuring coordination fragility in asset prices. https://doi.org/10.5281/zenodo.19183038

in this paper i operationalize the mixing theorem by defining the beierspace fraction b(p, t), the share of an asset's valuation variance attributable to coordination-dependent beliefs rather than physical laws, and the fragility gap F(p, t) between actual and required maintenance bandwidth. i predict that assets with F > 0 accumulate hidden coordination decay until a discontinuous repricing crash, and i link rising constraint complexity to the secular rise in tobin's Q.

2026j. the mars test: how super-transparency solves the toughest life and death problems in real time. https://doi.org/10.5281/zenodo.19180830

in this paper i use high-stakes, closed-loop survival environments (such as deep-space habitats and acute crisis response) as a limiting case. i argue that real-time, multi-dimensional structured disclosure and intermediate milestone scoring resolve life-and-death constraint coordination problems where scalar price mechanisms are too slow or too blind to prevent system failure.

2026i. observation is maintenance: axiomatic foundations for coordination thermodynamics. https://doi.org/10.5281/zenodo.19176947

in this paper i unify lossy price compression (2026c) and covariance instability (2026e) under a single axiom: observing a coordination-dependent structure is inseparable from maintaining it. from this i derive a thermodynamic dissipation law, a reverse water-filling maintenance budget, the eigenspace-beierspace mixing spectrum, an institutional regress theorem, a formal argument for goodhart's law, and the creation-side beier factorial. a note: the thermodynamic framing is withdrawn in formal extensions to the ocot corpus, where i argue the resemblance is generic to convex optimization; the dissipation law stands under its stated conditions.

2026h. excess distortion in the price system: a rate-distortion analysis of scalar compression in multi-dimensional economies. https://doi.org/10.5281/zenodo.19162609

in this paper i analyze the economy-wide welfare and reconstruction costs of operating below the gaussian rate-distortion bound. i argue that because scalar market clearing allocates all of the rate R_price to the consensus dimension and zero bits to the K - 1 constraint dimensions, downstream participants spend hundreds of billions of dollars annually attempting to reconstruct at high cost what structured disclosure preserves at source. a correction note is in preparation: the partial-loading minimality claim (proposition 1') does not hold at the paper's own parameters; the corrected result is stronger and replaces it.

2026g. the ngakpa principle: why AI alignment requires a pre-kernel architecture of wisdom and compassion. https://doi.org/10.5281/zenodo.19155356

in this paper i argue that kernel-level AI alignment (RLHF, guardrails, constitutional overlays) can be fine-tuned away when it is separable from capability, and that durable alignment must descend to a functional rom (pre-kernel) layer where alignment is load-bearing and constitutive of competence. drawing on the vajrayana ngakpa tradition, i map prajna (wisdom) to reconstruction capacity, distortion awareness, and swarm position, and karuna (compassion) to constraint-directed signal enrichment and active flutter suppression under stress. i then argue the coupling advantage theorem: inseparable rom architectures (D = alpha L, yielding x = alpha L^2) achieve quadratic information capacity growth and a computable phase-separation interval where rom systems self-stabilize (S_ROM > 1) while separable kernel systems (x = D* L) collapse (S_ker < 1). a note: the three mechanisms the paper names for building a rom layer are proposals; none has been built.

2026f. every AI agent lives in beierspace. https://doi.org/10.5281/zenodo.19154577

in this paper i set out the ontological partition between eigenspace (quantities governed by physical law whose spectral decompositions are stable unless physical force is applied) and beierspace (quantities such as prices, spreads, and ratings that are constituted by human beliefs about beliefs). covariance matrices carry no label distinguishing the two. i argue that autonomous AI agents trained as if market data lives in static eigenspace co-constitute a reflexive loop, amplifying learning L while often reducing disclosure D, and that they trigger cliff-edge phase transitions when deployed at scale without disclosure requirements.

2026e. supple: how beierspace extends eigenspace. https://doi.org/10.5281/zenodo.19154460

in this paper i formalize the temporary nature of financial covariance matrices through the stability condition S = LD / (sigma_e + sigma_b(LD)), and i argue from a heterogeneous-agent network model that human-generated structural drift sigma_b is strictly decreasing in aggregate learning and disclosure LD. i argue a reflexive phase transition at S = 1, a saddle-node bifurcation at the critical re-entry intensity mu_crit that creates a bistable collapse trap, an informational impossibility theorem under which a system inside the collapse trap cannot identify from its own crisis data whether collapse is permanent or recoverable, and a cognitive suppleness multiplier g(s, sigma_k) that unconditionally amplifies stability and expands usable eigenspace dimensionality.

2026d. resolving the grossman-stiglitz paradox through track record capital. https://doi.org/10.5281/zenodo.19124054

in this paper i resolve, within a cara-normal model, the 1980 grossman-stiglitz paradox by relaxing the assumption that trading profit is the sole return to information production. when agents build portable track record capital by publishing structured judgments verified against outcomes, information production decouples from asset trading (proposition 1) and strictly dominates price-only signals for multi-dimensional decisions (propositions 3 and 4). i micro-found the judgment market by decomposing the principal's willingness to pay into a consensus premium that vanishes as prices become efficient and a constraint premium that remains permanently bounded above zero (proposition 7), with explicit near-diagonal perturbation bounds (proposition 8). within the hurwicz-myerson framework i argue that under the brier score with coverage weight alpha < 4/5, the calibration threshold vanishes (q* = 0) and honest comprehensive disclosure strictly dominates strategic selection for all agents (propositions 13 and 14).

2026c. the economics of lossy compression: how price formation destroys information and why it matters. https://doi.org/10.5281/zenodo.19123888

in this paper i model the price system as a lossy codec compressing a K-dimensional gaussian source V into a scalar price loading on the consensus dimension V_1. i argue via reverse water-filling that while the price is an optimal codec on its own dimension, its excess distortion delta = D_price - D*(R_price) is strictly positive above the threshold rate R_0 = (1/2) log_2(lambda_1 / lambda_2), strictly increasing in active constraint variance, and converges to the total constraint variance sum_{k=2}^K lambda_k even in the limit of perfect market efficiency (proposition C, extended to general non-diagonal covariance in corollary C'). i also argue that supplementary disclosure bits targeting uncompressed constraint dimensions produce strictly higher marginal distortion reduction than additional price-precision bits whenever lambda_2 > lambda_1 / (1 + SNR_1) (proposition D).

2026b. the disclosure incentive: how track record capital resolves the grossman-stiglitz paradox and reveals the competition-cooperation synthesis across four disciplines. https://doi.org/10.5281/zenodo.19078916

in this paper i introduce track record capital, a portable, compounding asset built by publishing five-field judgments (input, method, output, identity, scaling; the record now has six facets, see the six-facet tensor) scored against outcomes. in a cara-normal general equilibrium model with three agent types (uninformed, traders, disclosers) i argue that information production is bounded below by lambda_D* > 0 even when noise trading vanishes (proposition 1), that it preserves total uninformed scalar precision while shifting its composition toward structured public reasoning (proposition 2), that it pareto-dominates the standard equilibrium (proposition 3), and that it avoids any reflexive judgment-market paradox (proposition 4). i then set out a competition-cooperation synthesis, in which cooperative disclosure generates the signal and competitive trading transmits it into prices, and i argue that it appears simultaneously across economics, computer science (open-source peer production), mathematics (the nash-to-pareto equilibrium shift), and physics (source preservation versus landauer-style reconstruction cost). i apply the mechanism to real-time externality pricing via the cost of capital.

2026a. reconstruction from compressed signals: markets and neural networks as a single information-theoretic problem. https://doi.org/10.5281/zenodo.19055458

in this paper i identify a literal structural identity between financial markets and neural networks. both are reconstruction systems that receive lossy compressed signals (a scalar price or a stripped training example) and must reconstruct latent causal structure in order to act or generalize. i define beier extropy E_B = H(P, D) - H(P) = H(D | P) as the bits restored to a compressed signal by structured disclosure, and i argue the dual-surface proposition: attaching structured disclosure D to signal P reduces posterior variance on the market surface (Delta_market > 0) and reduces bayes-optimal prediction loss on the AI surface (Delta_AI > 0) simultaneously if and only if I(V; D | P) > 0. this couples market transparency and AI training quality into a single intervention on a shared nonrival data stream.

pre-core papers

superharddata: liability-grounded information as training substrate for aligned artificial intelligence. https://doi.org/10.5281/zenodo.18120608

in this paper i introduce liability-grounded, attributed, outcome-scored human disclosures as a high-fidelity training substrate for artificial intelligence. i set out how real-world accountability and reputational stakes filter out synthetic noise and ground symbolic representations in physical and institutional consequences.

the atlas of accountability: a general theory of incentive-compatible semantics (superseded by superharddata). https://doi.org/10.5281/zenodo.18112795

in this paper i develop the early theoretical framework that links semantic meaning in information systems to the incentive-compatible accountability of the human originators who stake their track records on declared claims.

the liability-grounded neuro-symbolic bridge: a socio-technical solution to the symbol grounding problem (superseded by superharddata). https://doi.org/10.5281/zenodo.18111762

in this paper i propose a socio-technical resolution to harnad's symbol grounding problem: symbols in machine learning models acquire grounded referents when they are anchored to timestamped, identity-verified human judgments that are exposed to real-world outcome scoring.

early articles

sustainability arbitrage is the winning strategy for esg (or how to make your hedge fund an impact fund). august 12, 2023. chartered alternative investment analyst association. https://caia.org/blog/2023/08/12/sustainability-arbitrage-winning-strategy-esg-or-how-make-your-hedge-fund-impact/

esg data is a public good. let’s open it up. january 12, 2022. chartered alternative investment analyst association. https://caia.org/blog/2022/01/12/esg-data-public-good-lets-open-it/

esg data is a public good. let’s open it up: climate risk analysis should not be an “investment edge,” but a “public good.” february 9, 2022 (published on greenbiz, now trellis). https://trellis.net/article/esg-data-public-good-lets-open-it/

esg data is a public good. let’s open it up. november 29, 2021. https://impactalpha.com/esg-data-is-a-public-good-lets-open-it-up/

u.s. s.e.c. letters

u.s. s.e.c. - letter 3 “re: file number s7-10-22: the enhancement and standardization of climate-related disclosures for investors”, june 17, 2022: https://www.sec.gov/comments/s7-10-22/s71022-20132229-302750.pdf

u.s. s.e.c. - letter 2 “linux esg: the logical next step for markets and regulators”, september 29, 2021: https://www.sec.gov/comments/climate-disclosure/cll12-9309465-259892.pdf

u.s. s.e.c. - letter 1 “full market disclosure of climate financial risks”, august 19, 2021: https://www.sec.gov/comments/climate-disclosure/cll12-9202257-249829.pdf

formative interviews on keeping wisdom alive

“a jewel of humanity that belongs to all of us.” the new york times, 2014, video.

tbrc - keep wisdom alive - english, 2013, six minutes. tibetan buddhist resource center.

archival record: harvard magazine, 2014 · new york times, 2014 · harvard gazette, 2013

formative chapter emphasizing judgment risk

beier, g.c. (1996). “understanding the difference between hedging and speculating.” in r. klein and j. lederman (eds.), derivatives risk and responsibility (pp. 25-41). chicago, il: irwin publishers. https://doi.org/10.5281/zenodo.23849929

alfred d. chandler, jr., harvard business school: “your piece should be required reading for all investors dealing with derivatives.”

a seed in the dark

holds what no one has been told

it grows into thought

https://orcid.org/0009-0008-3263-2447

https://scholar.google.com/citations?user=9JGrqB0AAAAJ&hl=en

about