PUBLICATIONS
Publications
TAI publishes research on AI economics, the compute transition, and institutional capacity across six series: annual scenario reports, governance reviews, lecture transcripts, equity indices, working papers, and policy briefs. All publications are open access. No paywalls. No embargoes.
FEATURED · COMPUTE 2030 · JUNE 2026
Compute 2030
TAI's inaugural annual scenario report. Four structural scenarios for the AI economy through 2030, each analyzed through the lens of Ashby's Law and the constraints that shape cost, market structure, and institutional capacity.
READ THE FULL REPORT →"The question is not whether compute will reshape the world. The question is whether our regulatory institutions will have sufficient variety to absorb the disturbances it introduces."
FOUR SCENARIOS
Concentrated Dominance
State tension increasing on frontier compute.
Multilateral Fragmentation
Competing national compute blocs.
Governed Transition
International coordination succeeds.
Diffuse Proliferation
Democratized compute; distributed power.
FEATURED · WORKING PAPER · JULY 2026
Via Negativa
The AI economy by elimination. Thirty questions on compute, data, models, labor, and the physical world, each forecast not by projection but by testing every candidate answer against six hard constraints and keeping only what survives.
READ THE WORKING PAPER →"Superintelligence may or may not arrive on schedule; the electricity bill, the memory, and the trust will arrive regardless."
SIX CONSTRAINTS, ORDERED BY HOW FAST EACH RELAXES
Thermodynamics
Physical law. The only constraint that bars outright. Never relaxes.
Talent and Absorption
Generational time. Organisations learn at human speed. 10 to 40 years.
Data
Accumulated stock, not replenished. The training corpus is a fixed endowment.
Matter
Industrial production. Transformers, fabs, grid equipment. 2 to 7 years.
Law and Legitimacy
Collective decision. Regulation, liability, public trust. 3 to 24 months.
Capital
A price. The most mutable of the six, and therefore the one least able to bind.
WORKING PAPER
JULY 2026
Via Negativa: The AI Economy by Elimination
Sidney Scott
This memorandum forecasts the AI economy by elimination rather than projection. For each question it enumerates the candidate answers, computes for each the time its binding constraint would need to deliver the required expansion, and compares that against the horizon remaining. Six constraints are derived from a question that never mentions artificial intelligence: what does it take to obtain one more unit of a limiting input? Only the first, physical law, licenses deductive elimination. The other five reprice, which is why every verdict carries a probability rather than a proof. The engine decides eight of the thirty questions here, and that ratio is reported as a result: it is a test for which questions have physically determinate answers, and most do not. Turned on the field's own forecasts, the method produces its most consequential result. Every published estimate of AI's contribution to output is a bet on an unmeasured elasticity between deployed compute and attributable output. Its observed values span a range that reverses all of them, including this memorandum's own GDP band, which the engine eliminates at the low end. The aggregate question is therefore underdetermined on current public data. Two further findings cost the paper more than they gave it. The Residual Ratio, this framework's coined unit, is withdrawn: built from primary sources, none of its four complement components is constructible, so the ratio never had a numerator. And a pre-registered backtest records a maximum-confidence miss, on the one constraint tier the framework had already named as most exposed. What survives is smaller and sourced: the price of a fixed capability falls four to fifty-six times faster than the price of the best available model, a binding constraint attracts no supply response where a cycle does, and one dated call is staked with its adjudication rules specified in full.
READ →COMPUTE 2030
JUNE 2026
Compute 2030: Four Scenarios for the Compute Transition
TAI Research Staff
Four structural scenarios for the compute transition through 2030, analyzed through the lens of Ashby's Law of Requisite Variety. Each scenario represents a distinct configuration of compute power, regulatory capacity, and institutional variety — and each carries different implications for governance, equity, and democratic accountability.
READ →WORKING PAPER
FORTHCOMING 2026
Variety Deficits in AI Governance: A Structural Analysis
TAI Research Staff
An application of Ashby's Law to current AI governance frameworks, identifying structural variety deficits in existing regulatory architectures. We develop a formal measurement methodology for regulatory variety and apply it to six major AI governance frameworks, finding systematic deficits across all six.
READ →POLICY BRIEF
FORTHCOMING 2026
Compute Export Controls and the Good Regulator Theorem
TAI Research Staff
Applies the Good Regulator Theorem to the design of compute export control regimes, arguing that effective controls require regulatory bodies to model the full variety of compute applications. Current export control frameworks exhibit significant variety deficits that undermine their effectiveness.
READ →GRT LECTURE SERIES
FORTHCOMING 2026
The Good Regulator Theorem: Mathematical Foundations and Governance Implications
Inaugural GRT Lecture
A rigorous exposition of Conant and Ashby's 1970 theorem and its implications for the design of regulatory institutions in the age of AI.
FORTHCOMINGCOMPUTE EQUITY INDEX
FORTHCOMING Q4 2026
Compute Equity Index 2026: Inaugural Edition
TAI Research Staff
The inaugural edition of TAI's annual index measuring the distributional effects of the compute transition across geographies, sectors, and demographic groups.
FORTHCOMINGGOVERNANCE ANNUAL
FORTHCOMING 2027
Compute Governance Annual 2026
TAI Research Staff
Annual review of developments in compute governance, analyzing regulatory actions, international agreements, and institutional developments through the lens of Ashby's Law.
FORTHCOMINGOPEN ACCESS POLICY
No Paywalls.
No Embargoes.
All TAI publications are released under open access terms. Research that is publicly relevant should be publicly available. TAI does not accept publication fees, embargo agreements, or any arrangement that restricts access to its research outputs.
PUBLICATION SERIES
Compute 2030
Annual scenario report on the compute transition
Compute Governance Annual
Annual review of compute governance developments
GRT Lecture Series
Transcripts and papers from the lecture series
Compute Equity Index
Annual distributional analysis
Working Papers
Research in progress and preliminary findings
Policy Briefs
Concise policy-relevant analysis