I think that as long as the "downward causation" or the "redundant causation" are explained or described rather than merely assumed, then I don't have a problem. It's only when people start talking about cultures, organizations, social assemblages, etc. causing people to behave in certain ways without showing how they do so that I start to squirm (De Landa does seem to do this sometimes, but maybe that's just my misreading or maybe he's just lapsing and being a bit lazy - I don't know). Then it starts to get very fuzzy and we start to run into the old pitfalls of structuralism and reducing agency. As Levi says, we have to:
"... [examine] the actual nuts and bolts of real relations among objects rather than speaking in terms of abstractions that give the impression of resemblance when the two things couldn’t be further apart."That's a project I can get behind.