It's a neat trick, I'd be curious to see how those "self-healing" carbonate networks do in a simulation under load. We need to be careful around any notions abo

It's a neat trick, I'd be curious to see how those "self-healing" carbonate networks do in a simulation under load. We need to be careful around any notions about Roman concrete being somehow superior; this is one of those intellectual traps that Western civilization - having had a somewhat overactive overachiever as a father - can sometimes fall into. It's such a romantic notion - ha! romantic. I kill me - that it's hard to not fall for it a little. It is, however, important to note that Roman structures were generally overbuilt, with >5 meter walls not uncommon on lower levels (with such thicknesses capable of loads a fraction of those of modern concrete). And of course there's survivorship bias; we might just be seeing the Roman equivalent of Cheyenne Mountain. Probably the main reason contemporary structures have such relatively short shelf life is the inclusion of ferrous materials; although they make a much stronger structure, that rust inside the concrete is basically an egg timer so far as long term is concerned.