On the flip side, his geographical determinism is, to put it simply, jaundiced and wrong. His conclusions that might theoretically apply to the interactions between Europeans and indigenous Americans, even if they were based on correct facts (and frequently they were not) do not apply when juxtaposing Europeans and the inhabitants of Africa, or the Middle East, or East and South Asia. On the whole, Diamond, as most anthropologists and macrohistorians are by the very nature of their fields, is guilty of gross overgeneralization, inventing a forest regardless of whether there are actually any trees around. There are also the philosophical objections to an argument based so strongly on "inevitable" longue durée factors as opposed to contingent ones that can be physically linked via minor things like causation.
These are the very broad strokes, of course, but that's the general thrust of why GGS and Collapse suck.