I can see exactly what Robopocalypse, by Daniel H. Wilson, was supposed to be. It was supposed to be a World War Z-style chronicle, an epic story of lone individuals contributing to the greater whole of the destruction of a foe that threatened not just human society, but human existence itself. It was supposed to be an imaginative look into the potential threats of an increasingly mechanised existence, and a commentary on the importance of individual efforts in an assembly line world, and an archetypal tale of good versus evil.
It was supposed to be all those things. It's not.
Robopocalypse is not a very good book, in no small part because it's trying to actually be several different books at the same time. As I said, the construction of the book, a narrator recounting seminal events from a global battle against a non-human threat to our species, is very World War Z. Unfortunately, where WWZ had real scope, with tales from America and Cuba and South Africa and Russia and China and more, Robopocalypse is hamstrung by frankly inexcusable provincialism. There's a story in Tokyo and one in London, and an American in Afghanistan, and then there're six stories set in and around the northern/north-western United States. Heck, at one point the characters travel from the contiguous United States to Alaska, with not so much as a mention of passing through Canada. South America and Africa might as well not exist, and as for Eurasia, well, that actually leads into the next problem. You see, where WWZ's framing story allowed for a variety of tales to be told, from failure to surrender to doomed last stands to triumphs to ultimate victory, Robopocalypse's framing story explicitly labels these tales with a robot-generated 'heroes' tag, and the content of the stories makes it clear that they're supposed to be human heroes. And sometimes that works; Cormac and his fellow soldiers, Lonnie Wayne Blanton, Lark Iron Cloud, Takeo Nomura and Lurker could all be classified as 'heroes' of this conflict, given their accomplishments, and I'd even be willing to extend that to Mathilda Perez and her mother, and Spc. Paul Blanton. None of them are really capital-h Heroes, but they play important roles in terms of advancing humanity's ends. But the scientist who invents the AI and is killed by it? A second scientist, mad in an asylum, whom the AI kidnaps because it believes he can build more of its kind and who is never heard from again? A team of drillers who unknowingly establish the AI's Alaskan redoubt and die there, without ever telling anyone about it? Even worse than those included, though, are those left out, and here we come back to Eurasia. You see, before the protagonist's army marches on Alaska another army, or perhaps several other armies, make the same attempt, without success. But do they get any space in this 'heroes' archive? Not so much as a word. Their corpses are seen and largely ignored, and the narrator offers up a sentence or two of explanation in the between-story excerpts, but that's the closest the 'heroes' of several billion people and an entire continent come to getting recognition in this recounting of what is supposedly the first truly global conflict in which all men and women joined together.
But lest you think this review is just the complaints of a reader who went in expecting WWZ and got something different, just because like WWZ this book was followed by the author's how-to on surviving the exact same kind of apocalypse, don't worry; this book has a great deal of problems that are entirely its own. The most fundamental problem is the structure of the book. The short time frame of the individual stories rarely allow for meaningful character development, and given that most of them are third-person limited, it's difficult to get inside these people's heads. Mostly they're ciphers, with only the vaguest of motivations and little to really differentiate them from each other. The narration between the stories tries to flesh them out, but it's the classic 'show, don't tell' issue at work. And even the narration falls apart a few times, as it includes incidents and actions that won't happen until well after the conclusion of the framing story from which the narrator is speaking. The atmosphere of the book is likewise a flop, with the action set in an Indian reservation that's never really laid out, a NYC that's apparently been cleaned of every last corpse between the time of the robot uprising and the time the NYC characters venture outside, and the most bland locations in London and Tokyo possible, a houseboat on the docks, a television station, a retirement home and a factory. For a book about a brutal war there's almost no actual evidence of violence in the world, and just on a basic level this world is a blank slate for the reader. In terms of characters, a few humanoid robots join the story later, and they're also problematic, for technical reasons. The newly 'Awakened' humanoid robot that serves as the viewpoint of that particular story talks of never allowing itself to be 'enslaved' again, but these machines certainly didn't seem autonomously sentient before, and it comes across as silly, like claiming that humanity is enslaving its toasters and electric toothbrushes. It also raises the question of why an AI that can corrupt and overwhelm computer systems the world over would bother suppressing these humanoid robot's original programming, rather than just writing whole new operating systems that would happily work with the AI.
And oh, the AI. Once more, we come to the problem of scope. The AI, Archos, simply has none. It's a cartoonishly two-dimensional villain, evil simply because it is evil and performing acts seemingly simply because they're evil acts. Why does it want to kill all the humans? Why, if it wants to kill all the humans, would it use human forced labour camps to construct its forces, rather than just retool existing robotic assembly lines to produce whatever it needs? Why does it make tiny mobile land mines and huge quadrupedal walker-scouts, but not mass-produce the humanoid robots that would so easily meet and defeat the surviving humans on their own terms? Why does it experiment on humans, producing grotesque cyborgs that improve on neither humanity nor the machines, for the most part? Why does it have a single, central hub which, if its destroyed, kills it? Why does it seem to eschew guns and bombs in favour of those aforementioned, and resource- and combat-inefficient, walking mines? Where are the tanks and mobile gun platforms? Why doesn't it have a single plane or helicopter or UAV at its disposal? Where are the nuclear weapons? Don't expect any answers from Archos, who is either schizophrenic or a compulsive liar and who seems to have no motivation beyond 'be cartoonishly evil'. Skynet was a more developed character, and a more believable threat, than this thing.
The problem of scope comes up again and again in this critique, for very good reason. Robopocalypse feels much too small for a chronicle of an apparently global battle against Archos the AI. And it's not just the total absence of South America, Continental Europe, Africa and Australia, and the near-total absence of Eurasia and two-thirds of North America. There's no scope on a personal level, either. The NYC 'resistance' seems to be made up of two people and a handful of unnamed background characters. Gray Horse Army, the best fighting force humanity has and one that manages to go from somewhere in the north-western United States to Alaska, appears to consist of Lark Iron Cloud, Lonnie Wayne Blanton, and Cormac's 'Bright Boy Squad' of six. The 'freeborn' humanoid robots, despite apparently having so many of them that they're forming a city at the end, are represented by a single viewpoint character, two associated characters who have perhaps a half-dozen lines between them, and Nomura's love-doll who spends most of the book offline. This is a tiny, empty world, which would work great if it was supposed to confront the reader with the immensity of the challenge of defeating a world conquered by technology, but the AI's arsenal seems to be tiny and empty too, as tiny and empty as its imagination and, ultimately, the threat it poses.
The bottom line is that this book fails to present either a particularly realistic look at the after-effects of a robot rebellion (nobody for a second seems interested in turning away from technology Amish-style?) or an enjoyably bombastic tale of plucky humans versus coldly calculating machines. It's weak, it's disjointed, it has flat characters and a villain who is as evil as it is dumb, and utterly lacks any sense of scope or scale to sell the idea of a world in the grip of total war. This thing might pass muster as the script for an 80s half-hour animated commercial, but as a serious work of literary fiction it's fit only for the recycling bin.