Reportedly, they're in the process of procuring lower-profile flatware fleetwide which would allow 20 meals to fit in the neo ovens.

It's not impossible that limiting preorder on the 321neo to the number of hot meals actually would require a complete redesign of the meal preorder system in order to actually be an improvement. In the current form, for the flights for which preordering is possible, the preorders are reflected in the catering order. This means that it doesn't need to be strongly consistent at the level of a flight (only, at most, at the level of a PNR); there are quite a few reasonable designs that would exploit this to make a simpler, more available, and more scalable system, and if any of those designs are the ones in use, then there's a non-trivial chance of taking orders for more hot items than could be served, which isn't really an improvement. Getting told that you can't have the meal you preordered (given that the choice of whose ox gets gored likely comes down to 360/DM/PM upgrade vs. paid F) when the passenger next to you gets the dish you ordered is not a great experience: at least FEBO seems a lot less arbitrary (and implicitly favors paid F, or at least paid F that takes FEBO into account for seat selection, over upgrades).

Conversely, UA's preordering system is, from what I read on FT, more explicitly about reserving finite resources: they stop taking preorders after an option sells out, and there seem to be more complaints over there about not getting a preordered item (which perhaps says something about the implementation choices made by UA even in the presence of this difference in requirements... airlines are a business where strong global consistency is generally not the default choice in system design).