Learning a little about steel has given me some new cool physical and conceptual metaphors:

1. Complicated phase diagrams: In basic chemistry I learned about the typical "solid/liquid/gas" phase diagrams, but in metallurgy there are much more interesting phase diagrams describing differing crystal structures and so on. If I understand correctly (which I totally might not) these phase diagrams can even be pretty weird in the sense that the phase can be path dependent (think Ice-9 from Cat's Cradle)
2. Carbides: Steel alloys typically have some elements "in solution", i.e. dissolved in the iron, and others that form tiny islands of carbide and nitride compounds (collectively these islands are called "carbides"), which are actually the biggest factor influencing the material properties of the steel, since they're typically very "hard" (i.e. crystals that really want to stay in their current configuration, and so are very resistant to wear, scratching, and deformation) but not very "tough" (i.e. once you do get them out of their configuration they crack/shatter/chip and cause a cascade effect of shattering carbides that will cause the steel to crack on the macro-level as well)
3. "Properties" as sets of correlated measurements: There are many different ways to measure a piece of steel. Some of these ways of measuring are highly correlated, and we label these clusters as "properties". e.g. "Hardness" is not really one thing: When you're reporting a "hardness" you have to say which kind of measurement you mean, because while the different ways of measuring "hardness" are correlated with each other, the correlation isn't perfect. If a thing is very difficult to indent with a diamond point, it is probably also very difficult to scratch with corundum, but you can't say exactly how difficult.