I've never been entirely convinced that geography is really a science, but if there was a book that was likely to do so, it's Rivers of Power . What's more, Laurence Smith manages to bring alive the importance of rivers to the Earth, but more particularly to humanity, with some excellent storytelling. The book starts with a nilometer, an ancient structure for measuring the height of the Nile - and the role the Nile has played in Egyptian culture. From here we open out to a whole host of rivers around the world. Rather than focus chapter by chapter on particular locales, Smith leaps from place to place, covering the roles of rivers in, say, wars or trade or climate change. In doing so, he manages to communicate his enthusiasm and a feeling of engagement that makes the book both approachable and enjoyable. There's always something new and different turning up - no one, surely, would expect, for example, a chapter to begin with a discussion of the superhero movie Black

Before opening this book I had the distinctly unnerving wish that I would find it full of blank pages - because this is the 'essential knowledge' series, and our knowledge of extraterrestrials is, well, non-existent (sorry Mr Mulder). To be fair though, inside we get the next best thing. Wade Roush gives a readable, compact history of SETI (the Search for ExtraTerrestrial Intelligence), plus attempts to communicate outwards and a touch of information on potentially inhabitable exoplanets. We also, inevitably, get that magnificent piece of speculation raised to the power of n, the Drake equation which supposedly gives us a feel for the potential range of numbers of planets in the galaxy inhabited by intelligent life (at least we know the range starts at 1) and much puzzling over the old Fermi paradox of why there is no evidence of intelligent life out there. It's this last part that gets just a touch tedious - because it is all navel gazing speculation, it really can n

A part of the increasingly interesting MIT Press Essential Knowledge series, the topic of this book may at first glance be puzzling - but it's about something many of us use every day - information technology that makes use of spatial information, such as the GPS sat nav system. One oddity of this series is that it is very inconsistent in the level the books are pitched at. Some are way too technical for the general reader. This one, though, is a straightforward descriptive text with very limited technical detail. There's nothing here that is likely to baffle someone from outside the field and lots of information on where the various technologies have come from (including John Snow's famous map-based identification of the source of a Victorian London cholera outbreak), the basics of how they work and where they are likely to go from here. If you actually have a need to know this stuff, it's an ideal primer. My only  real concern about the book is that I'm not

Some brief guides miss the point and in reality go on for rather a long time, but this one very much does what it says on the tin. Readable in an hour, it gives us the basics of what modern artificial intelligence is and how it works. To achieve such brevity, James Stone has cut away much of the support mechanism of popular science - so we get very little background and historical context, limited storytelling and not much about applications. Instead, the focus is laser-sharp on delivering the basics of how neural networks work, where modern AI is successful and where it has a long way to go. Broadly, Stone identifies two areas where AI triumphs - image recognition and playing games. If I'm honest, perhaps he is a little generous on the first of these - as he points out himself, tiny elements undetectable to the human eye can be sufficient to totally change what an image is recognised to be, and most of the techniques (he takes us through different types of machine learning

Some science fiction novels feel like more of the same. They can provide an enjoyable new twist on a story, but we've seen their like before. However, Beatiful Intelligence  gives us a whole new take on the development of artificial intelligence. Two teams have broken off from a Japanese AI lab to go there own way and are on the run - they are as likely to be killed as sued. All this takes place in the context of the nexus, a next-generation internet that is far more immersive and intrusive, tracking and noting every activity of everyone. Although this is fiction, Stephen Palmer does treat the reader to a fair amount of discussion of the nature of AI, how it might be achieved, and whether it's possible to distinguish true artificial intelligence and consciousness from a simulation of it. That is interesting enough in its own right, but the attempts of the two teams to keep concealed despite the nexus, and their very different approaches to creating AI are both engaging

This month sees the 30th anniversary of the launch of the Hubble space telescope, so this is a timely book. I was really looking forward to reading it, but I was disappointed when I did. Not everyone will share that disappointment – someone who primarily associates Hubble with ‘pretty pictures’, and has little interest in the science and technology behind them, may well love what Jim Bell has done. It’s the archetypal coffee-table book – lavishly illustrated, with stunning colour photographs filling every other page, alongside text that is high on poetic adjectives and low on technical facts and figures. I’m sure Bell has done a good job of producing the kind of book that he (and/or his publishers) wanted, but from the popular science point of view – which is what this review site is about – it doesn’t do justice to the Hubble telescope or the scientists who work on it. By adopting a picture-driven format, there’s an in-built bias towards photogenic sights like nebulae and gala

There's a certain class of book that can be described as 'it should have been an article'. This is where there are only a few significant points to be made, which would make an interesting magazine article, but the whole thing becomes intensely tedious when dragged out to book length. (A lot of business books fit into this category.) I was distinctly worried that this would apply to Exercise is Medicine - yet despite, in a way, it being true, Judy Foreman manages to make the book one that's packed full of information and an interesting read - even to someone who hates sport and doesn't like medical books. Let's get that main point out of the way - exercise is really good for you. Even a relatively small amount - say 150 minutes per week of brisk walking - will have a significant impact on your health and potentially increase your lifespan. It helps all round from blood pressure to mental state. That's the article part. But what Foreman does is take