Elementary formal models based on the asymptotic limit theorems of information and control theories challenge current culturally and ideologically-constrained Western ‘Tech Bro’ conceptions of Artificial General Intelligence. Confronted by Darwinian selection, cognition has always been paired with real-time regulation under an embodiment that interacts with, but is part of, an often actively adversarial embedding environment. Absent real-time embodiment/regulation, there can be no intelligence as-we-know-it. In particular, Western Tech Bro atomistic cognitive machines will not function well confronting the ‘Clausewitz landscapes’ of fog, friction, and adversarial intent that characterize much of real-world experience. As history has shown, Darwinian selection challenges the culture-bound ideologies of even the most powerful empires. The modeling approach, which is surprisingly general, can be used to construct specialized statistical tools for the analysis of observational and experimental data regarding cognition across a broad range of scales and levels of organization.