NM AI Research

Independent researcher analysing the technology, energy and artificial-intelligence build-out — how it is financed and operated, and how its products are engineered to shape what people spend and choose. Contested figures rebuilt from primary public data; forecasts, valuations and stated targets scored against what actually happens. Every figure traceable; every method reproducible where the question is quantitative.

Interactive tools explore the data behind the papers

Self-contained, dependency-free front-ends over the frozen, citable datasets — change an input and watch the figure move. Each is reproducible and archived with its own DOI.

CEO Pay-vs-Delivery Scorecard

What the S&P-100's highest-paid CEOs took versus what they delivered. Toggle granted vs realized pay and watch the league table reorder; or read the board's own payout-vs-target beside peer-relative return.

Open the tool → paper DOI · tool DOI · code & data
Python stdlib · SEC EDGAR (Pay-vs-Performance) · CSV · build.py

AI Energy-Demand Forecast Scorecard

How the field forecasts data-centre electricity demand. Measure dispersion only within comparable slices (units and scopes are not interchangeable), and trace the transparency funnel: verified → confirmable → reproducible.

Open the tool → paper DOI · tool DOI · code & data
Python stdlib · primary forecast documents · CSV · build.py

Contingent vs Robust AI Power Demand

How much announced generation actually gets built. A deflation calculator runs any announced capacity through three measured PJM-queue attrition stages — about one announced MW in five is delivered.

Open the tool → paper DOI · tool DOI · code & data
Python stdlib · PJM interconnection queue · CSV · build.py

Papers open-access, reproducible, primary-sourced

Working papers on the AI build-out's finances, energy and risk, and on how digital and retail systems are engineered to shape spending and choice — built from primary public data and the published evidence base, with reproducible scripts wherever the question is quantitative. Published open-access on Zenodo under CC BY 4.0.

2026

The Consumer Incidence of the AI Buildout

A reproducible ledger of how the AI data-centre build-out reaches a single household — device prices, AI-laden subscriptions, electricity, public subsidy and retirement exposure — with each channel's AI-attributable share bounded, the contested energy row given both sides, and water held outside the ledger as a cost that resists an honest number. Two archetype households; lead with the low bound.

10.5281/zenodo.20941004
Python stdlib · primary public data (EIA, PJM IMM, BLS, Good Jobs First, Counterpoint) · scripts in the DOI bundle
2026

Consumer Reliance Architecture

How individual AI dependency is built, priced, and made hard to leave: a reproducible profile of the commercial machinery (pricing tiers, caps, metering, portability) across US, EU, Chinese and companion providers, with a revealed-elasticity test of which squeezes actually retain users. The consumer leg of the reliance trilogy.

10.5281/zenodo.20807279
Python stdlib · public pricing pages · Google Trends · scripts in the DOI bundle
2026

Who Does Your Shopping Agent Work For?

An analysis of agent-mediated commerce: where the hype stands against what is delivered, why the "neutral agent" is a false assumption, who controls the resulting chokepoint, and which defences are actually enforceable.

10.5281/zenodo.20790897
2026

Designed, Not Accidental

An evidence review of how physical and online retail environments are engineered to increase spending, why individual vigilance is an inadequate defence, and what works in its place.

10.5281/zenodo.20788845
2026

AI-Literacy Load by Occupation

A reproducible map of where AI-fluency becomes part of the job — occupation-level AI-usage penetration crossed with the augment/automate mix, from the open Anthropic Economic Index.

10.5281/zenodo.20778969
Python stdlib · Anthropic Economic Index · O*NET · scripts in the DOI bundle
2026

AI Reliance: how deep, and how reversible

A reproducible measure of systematic AI dependency — reliability, downstream attribution, penetration and reversibility — built from primary status-page and government data.

10.5281/zenodo.20763481
Python stdlib · provider status-page JSON · primary government data · scripts in the DOI bundle
2026

The CEO Pay-vs-Delivery Scorecard interactive ↗

A reproducible, descriptive audit setting the S&P-100's highest-paid CEOs' realized pay beside what they delivered.

10.5281/zenodo.20709603 · code & data
Python stdlib · SEC EDGAR (Pay-vs-Performance) · build.py
2026

The AI Energy-Demand Forecast Scorecard interactive ↗

A reproducible audit of how the field forecasts data-centre electricity demand — dispersion, revisions, and transparency.

10.5281/zenodo.20708978 · code & data
Python stdlib · primary forecast documents · build.py
2026

Contingent vs Robust AI Power Demand interactive ↗

How much announced generation gets built — a realized-completion anchor from the PJM interconnection queue.

10.5281/zenodo.20706509 · code & data
Python stdlib · PJM interconnection queue · build.py
2026

The AI Capital-Infrastructure Barbell

A supply-and-demand risk assessment of the AI build-out: a likelihood-by-severity register of where the system is fragile.

10.5281/zenodo.20708803
2026

The AI Private-Valuation Marks-vs-Reality Ledger

A pre-registered, reproducible record of how private AI valuations survive contact with a real public price.

10.5281/zenodo.20709174
Python stdlib · N-PORT / S-1 / Form 4 (SEC) · scripts in the DOI bundle
2026

AI Cost Watch

Tracking the unit economics and cost re-pricing of AI — the gap between cheaper tokens and cheaper AI.

10.5281/zenodo.20682980
2026

The AI Industry: a PESTLE scan

A structured scan of the political, economic, social, technological, legal and environmental pressures on the AI industry.

10.5281/zenodo.20719282
2026

How Much Electricity Does Bot Web Traffic Actually Use?

A transparent bound on the electricity used by automated web traffic — the proof-of-method for the deflation discipline.

10.5281/zenodo.20707001
Python stdlib · Cloudflare Radar · primary sources · scripts in the DOI bundle
2026

Anomalous Constant Domains and Anomalous Void Transmission

A published negative result in cosmology — kept on the record because reporting what did not work is part of the discipline.

10.5281/zenodo.20512148