And try as I might, it still looks horrible 😔

If we were starting from scratch, iron is about as common as it gets, geologically, not to mention how much “leftover” iron and steel there would be, post collapse. Barium/strontium for sintering ferromagnets is trickier, but ores like witherite and baryte aren’t uncommon, though they are quite location specific, globally speaking. Not that you don’t actually need full blown permanent magnets, just ferromagnetic material.

However, the easier (though less efficient) option would be to use dynamos with field coils instead of ferromagnets. That way the only resource you need is wire.

I’m working on something similar, coming from a process engineering/agent simulation approach, and looking for collaborators. How should we organise?