AI Claim Stress-Tester

The source assigns both trajectories an index of ten in twenty twenty-four. Capability compounds by one plus the chosen growth rate each year. Institutional readiness increases linearly by one hundred two divided by lead time plus half the policy lag. At the starting forty-five percent growth, eight year lead and three year lag, the capability curve eventually overtakes readiness. This diagram uses seventy-five pixels per year and point four five pixels per index unit, with both series on the same scale. The critical window is the first sampled year with a positive difference. These are chosen equations, not observed capability measurements, calibrated institutional forecasts or evidence that a claim is true. The variable named total gap area is a sum of positive differences at annual samples. It is not a continuous integral or independently measured disruption. The disruption score divides that sum by twelve, adds thirty-five times the growth fraction, then clips to a range from five through one hundred. The displayed gap in years is another assigned conversion: maximum index difference divided by twenty-five, plus point three times the lead time, clipped at ten. The starting values hit both upper bounds. Once clipping is active, more severe inputs can leave the displayed number unchanged. A critical window value of twenty thirty-five can also be the default sentinel when no positive gap appears; read the underlying trajectory rather than treating that date as a prediction. The counterfactual grid subsidy checkbox does not alter either trajectory or the overall disruption score. It multiplies only the power stress formula by point seven before rounding and clipping. At disruption one hundred, power falls from one hundred to eighty-one, while governance remains at one hundred with a three year policy lag. Labor is ninety-five and capital eighty. The highest sector is selected with a strict greater-than comparison, so ties retain the earlier sector. This changes the displayed bottleneck through ranking, not through a modeled expansion of electrical capacity. The JSON export records these inputs and assigned outputs. No backend claim checking, evidence retrieval or real policy simulation occurs.