Here are the four drivers cited in the 2020 Foreign Affairs article “How a Great Power Falls Apart: Decline Is Invisible From the Inside”.

Amalrik identified four drivers of this process. One was the “moral weariness” engendered by an expansionist, interventionist foreign policy and the never-ending warfare that ensued. Another was the economic hardship that a prolonged military conflict—in Amalrik’s imagination, a coming Soviet-Chinese war—would produce. A third was the fact that the government would grow increasingly intolerant of public expressions of discontent and violently suppress “sporadic eruptions of popular dissatisfaction, or local riots.” These crackdowns were likely to be especially brutal, he argued, when the suppressors—police or internal security troops—were “of a nationality other than that of the population that is rioting,” which would in turn “sharpen enmities among the nationalities.”

It was a fourth tendency, however, that would spell the real end of the Soviet Union: the calculation, by some significant portion of the political elite, that it could best guarantee its own future by jettisoning its relationship to the national capital. Amalrik supposed that this might occur among Soviet ethnic minorities, “first in the Baltic area, the Caucasus and the Ukraine, then in Central Asia and along the Volga”—a sequence that turned out to be exactly correct. His more general point was that in times of severe crisis, institutional elites face a decision point. Do they cling to the system that gives them power or recast themselves as visionaries who understand that the ship is sinking?

It is hard not to think of this in context of my new book, where I propose a bifurcation of all information systems security into either “easy, routine, minimal judgment” (ERM) or “identify, store, evaluate, adapt” (ISEA).

The TL;DR of the article is really this:

A better way to think about political cleavages was to observe which portions of society are most threatened by change and which ones seek to hasten it—and then to imagine how states might manage the differences between the two.

And so the struggle is what to do with such cleavage as it naturally occurs within big data systems.