Kim Stanley Robinson is an author with a message, and after reading this book you will understand why terraforming is impossible and why humanity will never succeed with interstellar colonization. Aurora is the counterargument to his earlier Mars trilogy (Red Mars, Green Mars, Blue Mars), which, if you were to read them, would make you understand why terraforming is a practical method for taking humanity to the stars.
The story follows Freya, a passenger of a generation starship, and is narrated by the ship's artificial intelligence. At the start of the book Freya's mom asks the AI to create a narrative of their journey, and we get to follow the AI as it first struggles to learn the basic skills of writing and telling stories; and, in doing so, slowly becomes conscious.
The partial stream-of-consciousness and fourth-wall breaking of the AI is an interesting technique. Whether by design or accident, Aurora doesn't really have any interesting characters, and by having a semi-aware AI tell the story we are conveniently spared any ham-fisted attempts by the author of character exploration and can quickly get to the whole point of the book: being lectured by the author on why humanity will never survive away from Earth and why it is stupid to even think so.
I'll just note that I think the AI is hilarious and would love to read more books written by it. It really has the gift of words.
But Aurora isn't a story as much as it is an argument against interstellar colonization and a statement about how ecologies work, so no matter how many little zingers we are treated to, and no matter how good or bad the plot of the book is (and certainly no matter how well the characters are realized) - what ultimately matters is how well Robinson argues for his position.
The core of Robinson's argument is that any colonization effort will only bring with it such a small portion of the Earth's ecosystem that it will never survive. Humanity is but one part of a superorganism that is made up of all life on Earth, and you can't just cut off a little bit of that superorganism and think that it'll survive in isolation - never mind thrive.
This has proven to be a controversial standpoint in SF circles for two reasons: emotional and scientific. The emotional reason is that if the standpoint is correct, it obliterates everyone's dreams about a galactic humanity. We're stuck on this rock. Forever. And we've made such a mess of it with environmental destruction that we're probably not even going to survive here. This is somewhat hard to stomach for any optimist.
The scientific reason is that, well, Robinson essentially makes the same argument over and over again: small biospheres don't survive, colonization efforts always start with small biospheres, therefore they will fail, therefore colonization efforts are pointless. Besides, any habitable planet we find will be populated by killer germs or so inhospitable that not even killer germs survive.
The argument is not a bad one - witness the difficulties encountered by Biosphere 2[a] - and the corollary that we should focus on taking care of Earth, the only planet we know we can live on, is reasonable. But Robinson has practically elevated his argument to a law and an assumption that we can call "the iron law of island biogeography": island ecologies de-evolve, and any generation ship will end up crewed by a wretched handful of surviving idiot midgets - never mind the colony they are supposed to create. The whole story is then built with the iron law as foundation, and so we (and the crew) end up exactly where you'd think. Some exceptions are made to keep the starship from turning into a rotting charnel house before the book begins, and to keep at least some colonists alive until the end of the book, but it's quite clear that Robinson thought his law was a nice hammer to use when shaping the story and now every plot difficulty looks like a nail for it.
Robinson also brings up a second problem with small ecologies: that we humans often end up fighting each other to death. Stable governments with populations in millions that live in freedom are practically unheard of in the history of the human race. Here, however, he ends up where other left-leaning authors like Iain M. Banks end up: the belief that humans must be ruled by a benevolent God-machine. Now I love the Culture and its Minds, and they sure let us solve interpersonal conflicts and get on with the story, but let's call it the "vanguard of the revolution" or "politburo" for short and it's a lot less convincing as being the optimal way. (Peter Watts, despite his constant misanthropy, had a much clearer view of the consequence of letting a benevolent AI rule in The Freeze-Frame Revolution.)
Both of these arguments ignore the facts that Biosphere 2 actually did manage to reach self-sufficiency, and that while the crew did factionalize, they also did cooperate and pull through. That said - Mars has been shown to be covered in toxic perchlorates carried on fine-grained dust that promises to wreak havoc with both machines and human lungs - so the results from Biosphere 2 don't generalize as much as we would perhaps want them to.
The truth then, seems to be that interplanetary colonization is a lot harder than most SF would tell you, but that there is no iron law forbidding it. The superorganism of Life on Earth could, in theory at least, spread to other planets. Maybe island biogeography is a problem - but Robinson seemingly throws up his hands and declares it impossible to solve, which doesn't follow.
Toward the end the starship - our symbol of humanity among the stars - is destroyed when it, like Icarus in his hubris, flies too close to the sun. We see what the author did there.
The book ends with a proponent for interstellar colonization getting the snot beaten out of them, and it's difficult to not see this as ultima ratio scriptor. If you've read this far and still don't accept the stated thesis, then this is what the author really would want to do instead of presenting one more argument for his case.
Still, Aurora is a good read. I picked it up for $3, and I certainly think I got my money's worth. Read not as a story but as an exploration of the author's ideas it provides a well needed counterpoint to the utopianism so common in SF; even if it certainly isn't the final word.