Given the odds of these people inventing a huge solution to a hard problem in a publishing vacuum (I am not aware of papers coming out that fundamentally solve

Given the odds of these people inventing a huge solution to a hard problem in a publishing vacuum (I am not aware of papers coming out that fundamentally solve this, for example) I think it is pretty safe to assume it's subject to the same limitations all such systems are subject to. Loss of power leads to loss of data. Commonly mitigated as you already state.