..a bit flabby and unfocused. Almost as if his editor was overawed by the author's reputation.
The structure is odd; anecdotal opening chapter moving on to ancient examples for which there is scant detail (they can't even agree on populations), only when half-way through do we get to recorded history and then, in modern times, we are asked to set aside many tenets because of (understandably) increased inter-connectedness. It's only in later chapters (especially 14) that he gets down to trying to substantiate his argument.
The book is very good on collapses but less so on the reasons for them. It reads like a disaster movie as civilisations crash around the reader. Grimly fascinating and a bit like 'Towering Inferno' - next time you go up in a tall building, you might be next!
There is much 'surface detail', e.g. about dating and archaeological techniques, where I feel Diamond hoping for borrowed hard-science certainty to support to his theories. His writing style is discursive and often lazy, for example, he introduces technical terms, leaves them with you without explanation for a couple of pages, and only gets around to defining them in his own sweet time. Again, a stronger editor should have addressed this.
I'm very supportive of the premises and I think that the big conclusions, about world population, deforestation, extraction and pollution are right - it's just that this isn't as good a book as it should have been.