That’s a fair point. Technically, with a strong AI stack, you can perform deep analyses across a much broader range of data.

The constraint in my framework is not computing power, but human judgment and decision-making capacity. The funnel is designed to preserve breadth initially, and then narrow down to the point where human conviction and portfolio construction actually occur. Beyond a certain point, more depth doesn't increase alpha; it increases noise and decision fatigue.

Therefore, it's more about where human attention creates the highest marginal value than what AI can process. This trade-off is ultimately an operating model choice, not a technical limitation.