The best non-fiction books that turn their topics upside down

As a coder and a lifelong SF reader, I am fascinated by AI. I've even written a LLM chat client named Shelley. Fascination, though, is not the same as uncritical fanboyism.

It is tempting to treat AI as a natural or magical apparition. Crawford's book turns this illusion on its head and explores AI literally from the ground up, beginning with its vast hunger for natural resources. She describes a similar hunger for training data as well as the implicit (or disguised) biases underlying the systems of classification that drive an AI's "understanding" of the world. 

This is not a book about the future of AI so much as a particular map of the state of the project – a look into the wizard's booth. It offers an essential first step in considering what is to come and how we might negotiate it.

I must have read a hundred books about story structure over the years. Somehow, perhaps because of some story-related blind spot on my part, none of them ever seemed to stick.

My problem was always the middle. Middles can sag. It seems that a story's interior becomes little more than the wasteland a protagonist must traverse to get from the mystery of a beginning to the ultimate challenge of an ending.

Yorke's Into The Woods celebrates the middle. He reminds us that the essential crux of a story lies in its midpoint. Right at the heart of an effective story, he argues, lies a fundamental transformation, a change so great that the protagonist emerges with new powers into a new world.

Furthermore, by analysing stories in five acts rather than three, he allows for an elegant symmetry in which the first and fifth, second, and fourth acts mirror one another.

It is not so much that Yorke invents anything new here, but his analysis is so clean, the shapes he distills so compelling, it is hard, after reading this book, to approach any story without some reference to its ideas.

I have always tended to see power as something that is imposed upon us. It is done to us. And, of course, you only need to glance at the Internet or a TV screen to see the truth of that.

But that can't be the extent of it. Power requires consent, which means it must also be at play even when it is not actively exercised.

Discipline and Punish is a work of history and philosophy which charts the evolution of punishment from a spectacle of cruelty in the middle ages, to the impersonal bureaucratic systems of the modern era.

Most interestingly, Foucault examines the ultimate privatisation of control – the way it outsources its management to us as individuals – so that we judge ourselves and watch ourselves, even punish ourselves on behalf of a set of quasi-scientific rules and standards.

In between other projects, I have been conducting research for a non-fiction book about the particular kinds of parallel universe that every story world creates. And that's how I came across Reality+.

The question as to whether we are living in a simulation is beloved of moviegoers, stoners, and undergraduates (quite a large intersection in that Venn diagram). It's also worthy of serious philosophical and scientific inquiry. David J Chalmers is a philosopher, and his book treats the question with entertaining rigour.

The book goes beyond the simulation hypothesis, though, to examine the philosophical implications of our inevitable colonisation of the virtual realm. At the heart of Chalmers' argument is a rejection of the opposition between the real and the simulated.

Lived experience, Chalmers claims, is real experience, as freighted with value and possibility in the virtual world as it is out here in the probably real, possibly simulated universe.

As a serial freelancer with an uneasy relationship to authority, one of my small obsessions is freedom (and how I can achieve it whilst still earning money). Because freedom is a theme in my life it's also a theme I seek out in the stories we tell ourselves.

Here's a story. Once upon, a time we subsisted in small roving bands and all lived in a kind of anarchy. Either that was a good thing (Rousseau) or a very unpleasant one (Hobbes). And then, with an agricultural revolution, came a change. More settled people and, later, more people still in cities, meant governments, laws, and money. Hierarchies were sadly (Rousseau) or blessedly (Hobbes) inevitable. 

The right and the left tend to fall into one or other of these camps. If they disagree about the desirability of the outcome, though, they usually just accept the progression as an iron law.

In The Dawn of Everything, David Graeber and David Wengrow challenge this consensus at its root. They suggest that there is nothing inevitable or even historically accurate about this fall from equality and that only a selective view of history provides evidence for it as a fixed process.

And if that's true, and where we find ourselves is not inevitable, then that's a new kind of freedom. Perhaps we'll get to choose where we go next.