Monthly ArchiveJanuary 2006



Non-fiction 22 Jan 2006 16:40

Guns, Germs and Steel

coverGuns, Germs and Steel
Jared Diamond

You can summarize this book in a short sentence: the reason for the the difference in affluence of today’s societies comes from differences in geography influencing the development of human civilizations during the last 10,000 years or so. It is a bit unnerving that the author spends more than 500 pages discussing this, but I will grant that he does not want to present just his conclusions: he presents the data that led him to it.

The book is, as a whole, very interesting, but several passages are somewhat boring because they are just going over things that where already discussed: some pieces of information are presented again and again with different historical data to support the conclusion. It does strenghten the argument, but also makes for a book that is longer than it should be.

As I said, the book is full of historical data. It can be thought of as a short history of the last 10,000 years of human civilization: from the beginings of agriculture in the Middle East to the colonization of Oceania by Europeans in the last two or three centuries.

The title of the book refers to the three elements that gave an advantage to Europeans over other civilizations when those started navigating around the world and contacting other societies. It is Diamond’s assertion that all three depend on the initial geographic advantages the Eurasia have over all other areas of the globe, and he exposes (in detail) why he thinks that, including examples of other “clashes of civilizations” that did not include the Europeans but, nonetheless, ended up with the geographically-disadvantaged society destroyed or vanished.

Very good book, and probably a decent companion to his other work on civilizations, “Collapse”. I’m looking forward to reading that one.

Buy from Amazon.com

Non-fiction 11 Jan 2006 14:10

The Making of the Movie Trilogy (The Lord of the Rings)

coverThe Making of the Movie Trilogy (The Lord of the Rings)
Brian Sibley

Making The Lord of the Rings was an incredible achievement by Peter Jackson and his whole (enormous) crew. Written after the second movie was done but before the third one, this book has an amazing amount of details about the production and the work involved.

The books covers both technical and artistic achievements, and one of the most impressive things is the sheer size of everything. It’s basically a 10-hour-long movie filmed over almost two years, with such stunts as using the whole audience of an official cricket game as “sound extras” for the orc army. All numbers impress: up to seven units filming at once, thousands of extras, 1,600 pairs of “hobbit feets” worn by just the four main hobbits… there was nothing small in the making of the movie, and no detail was left behind. Props were detailed to the point of exhaustion: even things that were never intended to be shown on screen (such as the inside of armour plates, or swords that would never be drawn) was richly crafted with all the details that could be expected.

More than the text, the illustrations bring the creation of the movie alive. The book is richly illustrated, with pictures, storyboards and paintings of all aspects of production. Truly an amazing book, deserves to be read by any fan of the series.

Buy from Amazon.com

Non-fiction 06 Jan 2006 15:44

On Intelligence

coverOn Intelligence
Jeff Hawkins and Sandra Blakeslee

In this book, Jeff Hawkins, one of the founders of Palm and, later, Handspring, presents his ideas about how our brain works and how this knowledge can change the way we create intelligent computers.

The main focus of the book is that all ideas we have about how to implement artificial intelligence in computers are wrong. We are trying to replicate intelligent behaviour, but this can’t really be done with the computer architectures of today: we should be trying, instead, to replicate architecture.

According to him, the way the cerebral neocortex (the area mainly responsible for intelligent behaviour) works is, basically, by memory: the brain does not compute responses, it remembers them. What we have, then, is not a gigantic, massively parallel CPU: instead, we have an hierarchical associative memory system that remembers invariant representations of the world. An associative memory allows the brain to remember the whole of an object (or situation, or sound etc.) when confronted with just a part of it, and invariant representations allow it to recognize a face in any orientation or light conditions, a song in any key and so on. The hierarchy is important to filter information and to allow for “automatic” learned behaviours, among other things.

Hawkins tries to show how the brain architecture might work to generate this; almost everything is theoretical, but he does make predictions about observations that should prove or disprove (in part or in the whole) what he proposes.

The theory is very interesting, and it does make sense. It seems to match the way the brain seems to work, even though many details are missing (some because they are not known, some because this is not supposed to be a densely technical book). If correct, this may indeed bring (within a few decades) machines with behaviour that can truly be considered intelligent, even if for small domains; something that the current AI is very far from delivering. Jeff Hawkins is someone to keep an eye on.

Buy from Amazon.com