Endosmosis (in French language) is not a perishable concept, but a permanent phenomenon which was discovered by Henri Dultrochet.
Your statement is contradicted by the following quote from the linked article: “La densité de l’hydrogène pur étant quatorze fois et demie moindre que celle de l’air, on comprend qu’il devra passer environ quatre fois plus d’hydrogène dans l’air que d’air dans l’hydrogène”. The translation in English language is: “The density of pure hydrogen being fourteen and a half times less than that of air, we understand that it will have to pass about four times more hydrogen in air than air in hydrogen.”
As a result endosmosis is also applicable to air entering the balloon, but in a lesser proportion due to the respective densities of air and hydrogen. It is an affair of density as I repeat.
That said variables such as “pressure difference, permeability and thickness of material” can occur. The long permanence high altitude airships: the opportunity of hydrogen paper you linked mentions the calculations of the leaks according to the altitude, but not the air entering the balloon that makes a high risk.
It is quite right. By using hydrogen we obtain also soon or late an explosive mixture after some time dependent of variables such as " pressure difference, permeability and thickness of material". So unless some means (such as a valve I evoked previously) are found, the hydrogen use for long time AWE operation seems to not be appropriate. It is not for nothing that almost all manufacturers use helium.
The comparison with devices using “pressurized hydrogen to cool generators” does not stand because endosmosis cannot occur.