Canonical Methodology & Provenance Directory

Weighted mean of per-signal normalized growth over the signals collected for the technology's discipline, computed by scoring/weights.py compute_momentum_score from the discipline-scoped signal registry in scoring/discipline_signals.py. Signal sets and weights differ by discipline (METHODOLOGY Tier 1 section 13); the weights are hand-set, not fitted. NULL when fewer than MIN_ACTIVE_SIGNALS signals are collected.

Output of the leave-one-out-validated logistic regression in scoring/arrival_live.py (model arrival_loo_v1), fitted on the historical CHAT technology cohort: the probability that a technology reaches its t50 within arrival_live.THRESHOLD years of its anchor year, from growth features of its metric_snapshots. When too little snapshot history exists the row is stored with arrival_model = 'weighted_v1', a linear map of momentum that is NOT a calibrated probability; with no momentum it is NULL.

Not currently computed. scoring/weights.py writes NULL for disruption_magnitude because no model or current source exists for it; earlier manual assessments are historical records and carry their own as-of date. The sources listed are inputs a future model would use, not evidence that a value is being produced.

Hand-set linear blend (not fitted or sourced), defined once in scoring/composite.py and used by both scoring/weights.py and scoring/rollup.py: arrival probability 0.4, disruption magnitude (scaled to [0, 1]) 0.4, momentum (rescaled from [-1, 1] to [0, 1]) 0.2. NULL when any term is NULL; missing terms are never defaulted. Disruption magnitude is currently always NULL at the innovation level, so innovation-level composites are NULL.

Standardized benchmark achievement against reproducible public evaluation suites (e.g., LMSYS Elo, MLPerf throughput, or battery energy density Wh/kg normalized to physical theoretical limits).

Sum of annual wage payroll across all O*NET occupations mapped to tasks automatable or assisted by the technology, computed from BLS Occupational Employment and Wage Statistics (OEWS).

Aggregate annual payroll exposure in occupational codes where task automation potential exceeds 50%: W_exposed = sum(employment_j * wage_j * task_overlap_j).

Cumulative undepreciated asset base rendered economically obsolete by emerging paradigm substitution: S(t) = Net PP&E * Disruption Rate.

Headcount in occupational codes where task-level exposure exceeds displacement threshold: H(t) = sum(employment_i * task_overlap_i).

Count of unresolved gating dependencies in the critical path across science, supply chain, and regulatory regimes.

Speculative financial multiple modeled via Carlota Perez Frenzy dynamics: captures enterprise valuation expansion, venture multiple inflation, and public equity market capitalization relative to installed baseline.

Real installed economic deployment modeled via logistic diffusion: P(t) = L / (1 + e^{-k(t - t_0)}), calibrated against physical adoption statistics (fabs, gigawatt capacity, units shipped).

Cumulative physical infrastructure capacity installed to support deployment (e.g. gigawatts of energized datacenter power, extreme ultraviolet lithography wafer capacity, transmission lines).

Dual-trajectory divergence: Delta(t) = F(t) - P(t), measuring speculative financial capital acceleration relative to installed production deployment.

Capacity spread between installed physical infrastructure and current productive utilization: Overhang(t) = I(t) - P(t). Identifies stranded capacity and timing risk.

Algorithmic classification into Irruption, Frenzy, Synergy, or Maturity based on key factor cost deflation, speculative decoupling delta, and infrastructure buildout.

Predicted calendar year at which entrant learning-curve cost intersects incumbent cost threshold: C_entrant(Y) <= C_incumbent.

Empirical percentage cost reduction achieved for each doubling of cumulative manufacturing production volume: Learning Rate = 1 - 2^{-w}.

Discrete count of independently verified commercial adoption events meeting Rubric C criteria (dated, cited, deployed at scale in real production environments).

Count of unique, non-affiliated corporate, government, or academic entities with confirmed production deployments of the specified technology.

Census count of distinct commercial organizations tracked across the 6-stage AI revolution supply chain architecture (EDA/Foundry, GPUs, ASICs, Cooling/Power, Hyperscalers, Frontier Labs).

Census count of active applied technologies and innovations with verified commercialization paths, patent signals, or adoption evidence in the platform graph.

Total public equity market value calculated as the latest closing share price multiplied by diluted common shares outstanding as disclosed in the most recent SEC Form 10-Q/K.

Ratio of current market share price to GAAP diluted earnings per share (EPS) over the trailing four fiscal quarters.

Ratio of enterprise market capitalization to total GAAP net revenue generated over the trailing four fiscal quarters.

Theoretical takeover price of a company: Market Capitalization plus Total Short- and Long-Term Debt, minus Cash and Cash Equivalents.

Duration in calendar years between the publication date of foundational basic science discovery (t_basic) and the first verified commercial deployment at scale (t_commercial).

Ratio comparing the historical mean translation latency of physical science / biotechnology (~13.0 years) against computer science / modern AI foundation models (~2.2 years).

Cumulative global peer-reviewed citation count indexed across Web of Science Core Collection and OpenAlex academic corpora for landmark foundational research papers.

Annual rate of peer-reviewed publications and co-citations referencing a foundational discovery node, normalized against the discipline baseline mean over a rolling 3-year window.

Expert-audited score (0 to 5) evaluating empirical support across 8 independent pre-registered dimensions: primary source rigor, falsifiability, replication, analog match, code availability, data freshness, boundary bounds, and conflict disclosure.

Count of unique OECD/NSF academic scientific disciplines represented among foundational basic discovery nodes (Computer Science, Condensed Matter Physics, Molecular Biology, Mathematics, Materials Science, Electrical Engineering).

Non-parametric survival probability function S(t) correcting for right-censoring in pending, uncompleted technology gestation horizons: S(t) = prod_{t_i <= t} (1 - d_i / n_i).

Strictly proper scoring rule evaluating probability calibration. Mean squared error between predicted probabilistic forecast P(breakout) and binary empirical ground truth outcome Y ∈ {0, 1} across historical test cases: BS = (1/N) ∑ (p_i - y_i)².

Percentage of historical test cases where the model's binary decision threshold (p ≥ 0.50) correctly matched empirical ground truth breakout or flop status across blinded historical epochs.

Empirical unit cost of the techno-economic paradigm's core key factor (e.g. inference intelligence per token, energy per kWh, transistor per mm²), tracking cumulative experience curve deflation according to Wright's Law C(Y) = C_1 · Y^{-w} relative to emergence base year.

Perez frenzy phase multiple comparing total financialized enterprise value (EV) of paradigm leaders against the cumulative tangible capital stock (CapEx) and physical R&D deployed into real production capacity: M = ∑ EV_i / ∑ (CapEx_i + R&D_i).

Carlota Perez structural friction metric calculating the spread between technological deployment velocity P_tech(t) and statutory/regulatory institutional absorption capacity A_inst(t): Δ_inst(t) = P_tech(t) - A_inst(t).

Unweighted mean, across the 14 non-core NAICS sectors, of the share of firms reporting AI use in the last two weeks (Census Bureau BTOS AI Use Supplement). Compared with Perez's 50% Synergy threshold. Null when the survey data is unavailable; no hand-set sector scores are used.

Audited commercial contract value meeting US GAAP ASC 606 standards: non-binding LOIs, memoranda of understanding, and uncommitted pipeline are assigned zero weight (0.0). Only binding take-or-pay purchase orders with fixed delivery schedules and termination fees qualify.

Physical mass conservation audit comparing the projected annual consumption of critical minerals/feedstocks at full technology scale against annual global primary mining extraction reported by the USGS.

Maximum likelihood estimate of the heavy-tailed power law shape parameter alpha governing economic value capture and citation concentration: P(X > x) = (x_min / x)^alpha.

GDP contribution net of intermediate double-counting: (∑ created_frontier spend × VA_ratio) − (∑ destroyed_incumbent spend × VA_ratio), deflated to Chained 2024 USD via BEA GDP deflators. VA_ratio sourced from BEA Input-Output Use Tables by 4-digit NAICS code.

Symmetrical supply chain cascade: (∑ created_frontier × (B_k − 1)) − (∑ destroyed_incumbent × (B_k − 1)), where B_k is the Leontief backward multiplier for NAICS code k. Represents net change in orders placed on upstream supplier industries.

Physical capital rendered economically obsolete net of recoverable salvage value: ∑(stranded_capital_assets_usd × deflation_factor) − ∑(stranded_capital_assets_usd × salvage_recovery_rate × deflation_factor). Default salvage rate = 20% unless explicitly recorded.

Welfare gain to consumers from unit cost reduction via Harberger triangle approximation: CS = baseline_spend × (ΔP + 0.5 × |ε| × ΔP²), capped at 5× baseline_spend. Uses explicit consumer_surplus_gain_usd if recorded; otherwise computes from unit_cost_reduction_pct and price_elasticity (default ε = −1.5).

Laplace-smoothed ratio of gross created-frontier spend to gross destroyed-incumbent spend (Chained 2024 USD): CDR = (gross_created + $1B) / (gross_displaced + $1B). The $1B regularization constant prevents division-by-zero and ensures continuity near zero-displacement markets.

Aggregate public equity market capitalization of all companies whose primary revenue derives from the tracked technology node, summed across SEC EDGAR-listed entities linked to the technology's market_metrics record. Computed by market_metrics_builder.py from company_market_signals.

Aggregate venture capital and growth equity financing volume raised by companies in the technology ecosystem over the trailing 12 months, sourced from SEC Form D filings and PitchBook/Crunchbase cross-referenced records.

Census count of all demand-side market arenas cataloged in the UTP market lifecycle registry, spanning active and extinct markets across 7 economic sectors and multiple Perez techno-economic surges.

Count of distinct competitive structure archetypes classified in the UTP market taxonomy, based on Tirole (1988) Theory of Industrial Organization: Natural Monopoly, Cournot Oligopoly, Bertrand Price War, Monopolistic Competition, Bilateral Monopsony, Contestable Commodity.

Count of direct asymmetric entrant-vs-incumbent substitution pairs tracked in the UTP lifecycle graph where an entrant technology is displacing or has displaced a legacy incumbent, based on Christensen disruptive innovation framework and UTP cost-parity crossings.