There is a theory that aggregation does not materially reduce the security afforded by risk sharing because miners and the economy will disperse as necessary, similar to the scattering cockroaches disturbed by a light. The theory irrationally implies that security actually exists because it could exist. This is essentially a rejection of the Threat Level Paradox, which implies that security evolves over time under a persistent threat.
The theory relies on grinders switching miner allegiance. This is based on the Balance of Power Fallacy, which incorrectly models miners as the threat. A shift of hash power from one mine to another does not reduce pooling or the risk associated with it. The risk is that states co-opt large amounts of hash power, substantially reducing the cost of attack. It is an error to assume that states do not collaborate in defense of seigniorage.
The International Monetary Fund (IMF) is an organization of 189 countries, working to foster global monetary cooperation... imf.org
As such one cannot assume that any large mine can exist outside of state control. A reduction in pooling requires an increase in the number of miners, specifically those who are willing and able to operate covertly. This requires that grinders suffer the increased cost associated with reduced pooling.
Yet people cannot be expected to work against their own financial interest. In order for risk sharing to increase, the financial pressures against it must be reversed. An assumption to the contrary is economically irrational.
The theory also ignores economic centralization and delegation. It is an error to assume the economy can rapidly decentralize, and de-delegation would most likely be infeasible in the case of state attacks as currency controls commonly restrict transfer.