Verdict Laboratory — The AI Reckoning

Probability-weighted futures

Four scenarios, normalised to 100%. The leading scenario is flagged below.

Transformation Moderate growth Stagnation Bubble burst

– Probability-weighted 2035 value index — realised economic value per dollar of 2026 AI spend most sensitive to –

–

Transformation

AI becomes general-purpose infrastructure. Revenue compounds, capability keeps climbing, the capex stack is absorbed and then some — the electricity-grid analog wins.

–

Moderate growth

Real, durable value — but slower and narrower than the bull narrative. Big winners, big write-downs, net positive. The railway-mania outcome: financiers lose, the world gains.

–

Stagnation

Capability plateaus near today's frontier. Adoption stays in the gap. AI settles into a large-but-ordinary software category — valuable, not epochal, and never worth the buildout.

–

Bubble burst

Revenue misses the financing structure. Stranded capex, debt distress in the circular loop, a 2000-style repricing. The infrastructure survives; the balance sheets don't.

Trajectory fan, 2026 → 2035

Each curve is a scenario's path of realised value (log scale, 2026 spend = 1). Line weight = current probability. The heavy black line is the probability-weighted path.

How this model works — full disclosure

This simulator is an explicit, inspectable toy model. It is not a forecast and its outputs are not predictions. Each scenario receives an "affinity score" — a weighted sum of your six normalised assumptions — and the four scores are converted to probabilities with a softmax (temperature 0.12). The weights below are our editorial judgments about which assumptions matter most for which future. Reasonable people would set different weights; the point of the instrument is to make your assumptions visible and arguable, including to yourself.

Assumption →	Revenue growth	Model improvement	Energy tailwind	Adoption speed	Capex useful life	Regulatory lightness
Transformation	0.30	0.25	0.10	0.20	0.10	0.05
Moderate growth	Bell curve: peaks when the composite assumption sits near the middle (0.5), decays toward either extreme — "neither miracle nor disaster."
Stagnation	0.20 (low)	0.35 (low)	0.10	0.25 (low)	0.10	—
Bubble burst	0.25 (low)	0.10 (low)	0.20 (low)	—	0.30 (short)	0.15 (strict)

Scenario value multipliers for the headline index: Transformation ×8, Moderate ×3, Stagnation ×1.3, Burst ×0.4 (2035 realised value per 2026 dollar). The trajectory curves are illustrative shapes, not calibrated projections. Everything here is reproducible from the page source — the model is about 80 lines of JavaScript.

Why no predetermined conclusion? Because the investigation's own evidence is genuinely split: the bear case's structural claims (circularity, debt, energy) are better documented, while the bull case's micro claims (task productivity, scientific discovery) are equally solid. The disagreement is about extrapolation — and extrapolation is where your assumptions, not our reporting, do the work. Re-read the strongest cases, then come back and see what you actually believe.