if the boring old normal model is correct, your brain is made of quarks, and so your brain will only be able to envision and concretely predict things that can predicted by quarks.

I didn't get the 'and so' above at first, but I think it makes sense for the following reason: you can only ever "construct models made of interacting simple things" (possibly elaborated upon and abstracted to such an extent that they no longer seem simple or physical) in that universe because any model you could possibly make in that universe would be causally determined by and entangled with the quarks in your brain. The verbalization and high-level understanding of the model is just another way of explaining what is going on with the quarks in your brain (it explains nothing additionally), and so whatever the 'irreducibly mental' things in your model are, the chain of causal unpacking and explicating ultimately bottoms out with descriptions of quarks, etc., by hypothesis. When you think "non-reductionist", there is a purely reductionist explanation of what you are thinking. If there is just one level, then the explanation for everything is on that level or can be reduced to that level, so you can't concretely envision, as Eliezer says, something that can't be reduced.

I wish I had time to make this clearer, but I don't have any more time today.